THE ASSOCIATION between low birth weight and the

Transcription

1 X/00/$03.00/0 Vol. 85, No. 4 The Journal of Clinical Endocrinology & Metabolism Printed in U.S.A. Copyright 2000 by The Endocrine Society Insulin Resistance Early in Adulthood in Subjects Born with Intrauterine Growth Retardation* D. JAQUET, A. GABORIAU, P. CZERNICHOW, AND C. LEVY-MARCHAL U-457, INSERM, Hôpital R. Debré, Paris, France ABSTRACT In a case-control study that investigated the effect of intrauterine growth retardation (IUGR) on glucose homeostasis, 20-yr-old adults born with IUGR were shown to be hyperinsulinemic in an oral glucose tolerance test, suggestive of insulin resistance. The aim of this study was to ascertain the decreased insulin sensitivity in young IUGR-born adults compared to that in controls. We studied 26 IUGR-born subjects and 25 controls, aged 25 yr. Insulin sensitivity was assessed by peripheral glucose uptake and monitoring free fatty acid (FFA) concentrations under euglycemic hyperinsulinemic clamp. The percent body fat was significantly higher in the IUGR group ( % vs %; P 0.02), contrasting with comparable body mass index in both groups. Insulin-stimulated glucose uptake was significantly lower in IUGR-born subjects than in controls ( vs THE ASSOCIATION between low birth weight and the later development of impaired glucose tolerance and type 2 diabetes was first demonstrated in an English cohort of 64-yr-old men (1) and has been confirmed by others (2, 3). In addition, previous studies reported an association between low birth weight and the insulin resistance syndrome observed at 30 at 50 yr of age (4, 5). These studies all investigated the relation between birth weight and glucose homeostasis later in life, but none specifically addressed the question of the consequences of intrauterine growth retardation (IUGR). For this purpose, we constituted a cohort of young subjects, carefully selected by size at birth from a population-based registry (6). In a case-control study, we showed that young adults (20 yr old) born with IUGR had normal glucose tolerance, but higher plasma insulin and proinsulin concentrations during an oral glucose tolerance test (OGTT) than controls matched for age, gender, and gestational age (7). This hyperinsulinemia suggested the development of insulin resistance early in adulthood, which remained to be confirmed. The mechanisms underlying the relation between body size at birth and impaired glucose tolerance or type 2 diabetes are unclear. IUGR is known to severely alter the fetal development of adipose tissue (8, 9), which in adults plays a key role in the development of insulin resistance (10, 11). It was recently reported in 70-yr-old men that the relation between low birth weight and glucose intolerance is mediated Received September 9, Revision received December 15, Accepted December 30, Address all correspondence and requests for reprints to: Dr. D. Jaquet, U-457, INSERM, Hôpital R. Debré, 48 boulevard Sérurier, Paris, France. djacquet@infobiogen.fr. * This work was supported in part by Pharmacia-Upjohn, France. Supported by a fellowship from Guigoz (France). mg/kg fat-free mass min; P 0.05), and the difference remained significant after adjustment for body mass index, total body fat, or waist to hip ratio. In IUGR-born subjects, insulin-stimulated FFA suppression correlated significantly with peripheral glucose uptake (r ; P 0.02). First phase insulin release in the iv glucose tolerance test, adjusted for insulin sensitivity, did not significantly differ between IUGR and control groups ( vs pmol/l; P 0.86). In conclusion, IUGR subjects have decreased insulin-stimulated glucose uptake as early as 25 yr of age without major impairment of insulin secretion. Low glucose uptake is associated with a lesser degree of FFA suppression in adipose tissue, which suggests a role of adipose tissue at an early stage of insulin resistance in these subjects. (J Clin Endocrinol Metab 85: , 2000) through insulin resistance (12). It would therefore be hypothesized that this association could involve a prior step of insulin resistance, as observed in the common form of type 2 diabetes. However, Hales et al. previously postulated that undernutrition in utero could impair insulin secretion later in life and contribute to the risk of type 2 diabetes (1). The aim of the present study was to measure insulin sensitivity in 25-yr-old adults born with IUGR compared with that in adults born under normal conditions. An additional aim was to test for a possible decreased insulin secretion relative to insulin sensitivity in subjects born with IUGR. To achieve these goals, insulin sensitivity was assessed by the measurement of insulin-stimulated glucose uptake and the monitoring of free fatty acids (FFA) under insulin stimulation. Insulin secretion was assessed by first phase insulin release (FPIR) following iv glucose stimulation. Subjects and Methods Study population The study was a case-control study in which all subjects were selected according to their birth data from the cohort of the previous study (12). Briefly, subjects were identified from a population-based registry of the metropolitan area of the city of Haguenau in France. This registry recorded information on all pregnancies, deliveries, and perinatal events in the area from (11). Local standard growth curves by gestational age and gender were derived from all live births registered. Our baseline cohort was of singleton subjects born full term ( 37 weeks) during , with IUGR (weight and/or height below the third percentile for gestational age and gender, according to the local standard growth curve; n 236). Control subjects were the next fullterm singleton birth of same gender in the registry, with weight and height between the 25th and 75th percentiles (n 281). Nondiabetic subjects aged 21 yr or more were selected for the purpose of the present study. IUGR was defined as birth weight below the third percentile. From the baseline cohort, 77 subjects born with IUGR, who were not lost for follow-up, were declared eligible. The eligible controls 1401

2 1402 JAQUET ET AL. JCE&M 2000 Vol 85 No 4 were 80 randomly selected from the group of control subjects. The rates of participation in this study were 34% and 31% in the IUGR and control groups, respectively, and 26 subjects born with IUGR and 25 controls were included in the present study between June 1997 and June In the IUGR and control groups, there were no significant differences between the participants and nonparticipants in terms of parental history of type 2 diabetes, cardiovascular disease and/or hypertension, dyslipidemia, and birth weight (Table 1). The etiology of IUGR was gestational hypertension (50%), smoking (30%), congenital abnormalities (7%), maternal short stature (7%), and unknown reason (6%). Three subjects had more than one factor. The study protocol was reviewed and approved by the Paris/St. Louis ethical committees. Methods All subjects underwent a first medical visit at which clinical data were recorded, and a 75-g OGTT was performed. Plasma glucose and insulin were measured at 0, 30 and 120 min. The percentage of body fat mass was derived from bioelectrical impedance analysis (RJL Systems, Clinton Township, NJ). Blood pressure was measured on the right arm of seated subjects after 5 min of rest, using an automated device (Dinamap, Critikon, Neuilly-Plaisance, France) and a cuff of recommended size for the mid-upper arm circumference. Three measurements were made at a 1-min interval, and the average of the last two measurements was used for calculating the mean arterial blood pressure [(systolic blood pressure 2 diastolic blood pressure)/3]. Subjects were admitted to the Clinical Investigation Unit of the Hôpital R. Debré (Paris, France) to assess insulin sensitivity by peripheral glucose uptake during a euglycemic hyperinsulinemic clamp as described by De Fronzo et al. (13). Sixty minutes after the start of the insulin infusion (40 mu/m 2 min), the steady state plasma glucose concentration was targeted at mg/dl. It has previously been demonstrated that in nondiabetic subjects, hepatic glucose production is suppressed by an insulin infusion rate of 40 mu/m 2 min (14). Glucose uptake was calculated as the average glucose infusion rate during two 20-min periods ( min), corrected for a target glucose of 100 mg/dl and adjusted for fat-free mass (milligrams per kg fat-free mass/min). Intraindividual coefficients of variations of blood glucose during the first and second periods were % and %, respectively. Mean plasma glucose concentrations before correction were similar in IUGR subjects and controls (99 2 and 98 2 mg/dl, respectively). Plasma insulin measured at steady state did not differ significantly between IUGR subjects and controls (plasma insulin vs U/mL). FFA concentrations were measured at baseline and 90 min after the start of the insulin infusion. The relative decrease in FFA concentrations during insulin stimulation was calculated as follows: (baseline FFA insulin stimulated FFA)/baseline FFA. Insulin secretion was assessed by the FPIR, which was measured during an iv glucose stimulation test (0.3 g/kg). FPIR was defined as the differential mean value of insulin at min over baseline [(insulin min/4) baseline insulin]. Analytical methods Plasma glucose concentrations during the clamp were measured by the glucose oxidase method using an on-site analyzer (Beckman Coulter, Inc., instruments, Fullerton, CA). Plasma insulin concentrations were measured using a double antibody RIA (ERIA Diagnostics Pasteur, Paris, France). Cross-reactivity with proinsulin and derived metabolites was less than 1%. Assay sensitivity was 1.2 pmol/l. Glucose, cholesterol, high density lipoprotein cholesterol, FFA, and triglycerides were measured by enzymatic methods. Statistical analysis All data were entered and analyzed using the SAS statistical package (SAS Institute, Inc., Cary, NC). Results are expressed as the mean sd. The differences between the IUGR and control groups were tested by 2 test for qualitative variables and Student s t test for quantitative variables. Plasma insulin, triglycerides, and FFA were log transformed before statistical analyses. The effect of group (IUGR vs. control) on peripheral glucose uptake and fasting insulin, adjusting for body mass index (BMI), percentage of total body fat, or waist to hip ratio, used regression models (general linear models procedure). The independent effect of IUGR on insulin secretion relative to insulin sensitivity was tested using the log-transformed variables in a general linear models procedure (15). Interaction between the group and the log-transformed glucose uptake on FPIR was also tested in this model. Correlations between FFA and triglyceride concentrations and BMI, waist to hip ratio, fasting insulin, and insulinstimulated glucose uptake were tested using linear regression models. P 0.05 was considered significant. Results Clinical characteristics Table 2 shows clinical characteristics at birth and at the time of the study in the two groups. As expected from the inclusion criteria, birth weight and ponderal index were significantly lower in IUGR-born subjects (P for both comparisons). Gestational age and gender distributions were similar in both groups. At the time of the study, mean ages did not significantly differ between the two groups. Body weight, BMI, and waist to hip ratio did not differ significantly between IUGR and control groups. In contrast, the percentage of body fat mass was significantly higher in subjects born with IUGR (P 0.02). As in the previous study, body height was significantly lower in the IUGR group (P 0.02). Mean arterial blood pressure were similar in both groups ( vs mm Hg). Smoking habits did not significantly differ between the IUGR and control groups (7 of 26 vs. 13 of 25; P 0.07). Oral glucose tolerance test data All subjects in the control group had normal glucose tolerance according to both WHO and American Diabetes Association criteria (16, 17). In the IUGR group, one woman had TABLE 1. Parental history of cardiovascular diseases and metabolic disorders and birth weight in the study population IUGR Participation (n 26) Nonparticipation (n 51) P value Parental history of Type 2 diabetes (n) Cardiovascular disease (n) Dyslipidemia (n) Birth wt (g) Control Participation (n 25) Nonparticipation (n 56) P value Parental history of Type 2 diabetes (n) Cardiovascular disease (n) Dyslipidemia (n) Birth wt (g)

3 INSULIN RESISTANCE IN ADULTS BORN WITH IUGR 1403 TABLE 2. Clinical characteristics at birth and at the time of the study in IUGR and control groups IUGR (n 26) Control (n 25) P value At birth Gender (male/female) 12/14 14/ Gestational age (wk) Birth wt (g) Ponderal index (g/cm 3 100) Placental wt (g) Adult Age (yr) BW (kg) Ht (cm) BMI (kg/m 2 ) Waist to hip ratio Body fat mass (%) TABLE 3. Correlation of insulin-stimulated FFA suppression and triglycerides concentrations to anthropometric and metabolic parameters in IUGR subjects (n 26) and controls (n 25) BMI W/H Fasting Insulin Glucose uptake r P r P r P r P Fasting triglycerides IUGR Controls Insulin-stimulated a FFA suppression IUGR Control a Insulin-stimulated FFA suppression (baseline FFA insulin-stimulated FFA)/baseline FFA). impaired glucose tolerance (2 h postload plasma glucose, 187 mg/dl). Plasma glucose at baseline and after the glucose load did not differ significantly between IUGR and controls at any time point (Fig. 1). Fasting plasma insulin was significantly higher in subjects born with IUGR ( vs U/mL; P 0.03). The difference between mean fasting plasma insulin levels remained significant after adjusting for BMI (P 0.03). When the insulin response was analyzed with respect to the glucose response using the ratio of the areas under the curves, this ratio was significantly higher in the IUGR group than in the control group ( vs mu/g; P 0.03). Insulin sensitivity Insulin-stimulated glucose uptake was lower in the IUGR group than in controls ( vs mg/kg fat-free mass min; P 0.05; Fig. 2). Being born with IUGR showed an independent effect on insulin-stimulated glucose uptake after adjusting for BMI in comparison to controls (adjusted means, 6.8 vs. 7.8 mg/kg fat-free mass min; P 0.05), and as expected, BMI had a strong independent effect (P ). A similar result followed after adjustment for the percentage of total body fat (adjusted means, 6.8 vs. 7.7 mg/kg fat-free mass min; P 0.05) or for the waist to hip ratio (adjusted means, 6.6 vs. 8.0 mg/kg fat-free mass min; P 0.04). Eight subjects born with IUGR (31%) showed an insulin-stimulated glucose uptake out of the control distribution and were regarded as insulin resistant (Fig. 2). These insulin-resistant IUGR subjects did not significantly differ from the insulinsensitive IUGR subjects in terms of BMI ( vs kg/m 2 ; P 0.07), body fat mass ( vs %; P 0.54), birth weight ( vs g; P 0.23), and ponderal index ( vs kg/m 3 ). In IUGR subjects, no significant relation was found between insulinstimulated glucose uptake and etiology of IUGR. Acute insulin response Mean values of FPIR were higher in subjects born with IUGR than in controls (76 47 vs U/mL). IUGR had no effect on FPIR (P 0.86) after adjustment for insulinstimulated glucose uptake. As expected, insulin-stimulated glucose uptake had a strong effect on FPIR (P ), but there was no interaction between the group and insulinstimulated glucose uptake (P 0.65). Serum lipids profile At baseline, cholesterol and high density lipoprotein cholesterol did not differ between the IUGR and control groups [ vs mmol/l (P 0.52) and vs mmol/l (P 0.34), respectively]. Neither fasting triglycerides ( vs mmol/l; P 0.08) nor mean FFA values at baseline ( vs mol/l; P 0.13) or the relative decrease in FFA concentrations during insulin stimulation ( vs ; P 0.24) significantly differed between the two groups. In both groups no significant correlation was found among fasting triglycerides, insulin-stimulated FFA suppression, and BMI or waist to hip ratio (Table 3). In the IUGR group, triglyceride concentrations were positively correlated with fasting insulin and inversely correlated with insulin-stimulated glucose uptake. The relative decrease in FFA concen-

4 1404 JAQUET ET AL. JCE&M 2000 Vol 85 No 4 FIG. 1. Plasma glucose and insulin variations during OGTT in IUGR and control groups. F, IUGR group; E, control group. trations during insulin stimulation was significantly correlated with insulin-stimulated glucose uptake and was marginally and inversely correlated with fasting insulin. These correlations were not observed in the control group (Table 3). Discussion The major finding of the present study is that adults born with IUGR have decreased insulin-stimulated glucose uptake as early as 25 yr of age. Being born with IUGR had an independent effect on insulin sensitivity after adjustment for BMI or total body fat mass, which are known to have a strong independent effect on insulin resistance. The higher insulin concentrations at baseline and in response to glucose observed in these IUGR subjects are consistent with peripheral insulin resistance. Interestingly, one third of the subjects born with IUGR had insulin-stimulated glucose uptake out of the reference range and similar to that reported in insulin-resis-

5 INSULIN RESISTANCE IN ADULTS BORN WITH IUGR 1405 FIG. 2. Individual and mean glucose uptake values in IUGR and control groups. Individual values in both groups are represented in the left panel. E, Control subjects; F, IUGR subjects. The right panel represents mean values in the two groups (, control group; f, IUGR group). The P value is given for the mean comparison adjusted for BMI. tant subjects (18). Our reference values were in keeping with values of peripheral glucose uptake usually observed in nondiabetic, nonobese, healthy subjects (18). Although it has been shown in some animal models that perinatal undernutrition could affect hepatic function in adulthood (19), there are no such data in humans. We thus made the assumption that hepatic glucose production would be suppressed by the insulin infusion rate used during the clamps in IUGR subjects with normal glucose tolerance, as reported in normal glucose-tolerant subjects (14). Therefore, we conclude that these nondiabetic subjects born with IUGR are insulin resistant. It is noteworthy that the parental histories of type 2 diabetes, cardiovascular disease, and dyslipidemia were similar in the two groups. Taken together, our data confirm the association between IUGR and insulin resistance, as reported in several older populations (4, 12). Previous studies also observed that the prevalences of syndrome X (4, 5), type 2 diabetes, or glucose intolerance (1 3, 12) were sharply potentiated by obesity in adults with low birth weight. The strong independent effect of BMI on insulin sensitivity similarly observed in our study points to IUGR as a contributing factor to insulin resistance, whereas the effects of other known risk factors for insulin resistance, such as obesity, remain unequivocal in these subjects. At 25 yr of age, our study population already showed decreased insulin sensitivity. This feature seems isolated, as we could not demonstrate any statistically significant difference between IUGR subjects and controls in BMI, blood pressure, triglycerides, or cholesterol, which all contribute to syndrome X. The absence of the complete syndrome could be attributed to the young age of our study population compared with that in previous studies in the literature (4, 5). Therefore, isolated decreased insulin sensitivity in early adulthood would indicate that insulin resistance and hyperinsulinemia are early defects in the development of syndrome X (20, 21). It should be noted that although nonsignificant, IUGR subjects had higher mean BMI and triglyceride concentrations, which strengthens the abovementioned hypothesis. In subjects born with IUGR, the relative decrease in FFA concentrations during insulin stimulation correlated significantly with insulin-stimulated glucose uptake. We regarded this observation as indirect evidence for insulin resistance in adipose tissue associated with low glucose uptake. Insulin resistance in adipose tissue is usually observed at a late stage of insulin resistance along with impaired glucose tolerance and type 2 diabetes (10, 11, 22). However, our data suggest that insulin resistance in adipose tissue appears early in IUGR subjects. These observations would thus argue in favor of a role of adipose tissue in the development of IUGRassociated metabolic disorders. IUGR is known to severely alter adipose tissue development in utero, resulting in a 5- to 6-fold decrease in body fat mass at birth (8, 9). Likewise, we have previously shown that serum leptin concentrations are decreased in fetuses and newborns with IUGR, in keeping with the decreased body fat mass (23). Babies born with IUGR demonstrate a postnatal weight catch-up growth characterized by an increased growth velocity during the first 2 yr of life (24). The increased percentage of body fat mass observed in adults with IUGR could be interpreted as abnormalities persisting in the growth of adipose tissue in adulthood, due to the special time course of adipose tissue development during the fetal and neonatal periods. Therefore, we propose that the adipose tissue of subjects born with IUGR is functionally and/or constitutively altered, with long term metabolic consequences on glucose homeostasis.

6 1406 JAQUET ET AL. JCE&M 2000 Vol 85 No 4 Using parameters derived from the OGTT, Barker s group reported that insulin secretion was decreased in subjects with low birth weight, and they hypothesized that undernutrition during fetal life could impair the development of -cell function (1, 25). In our previous study we did not find any difference in insulin secretion indexes derived from OGTT between IUGR and controls (7). In the present study we hypothesized that in young adults born with IUGR, impaired insulin secretion, if any, would be related to insulin sensitivity. Using iv glucose tolerance testing, we did not observe either a crude or a relative defect in acute insulin secretion. Therefore, impaired -cell function does not appear to be the primary defect leading to abnormal glucose tolerance and type 2 diabetes in humans born with IUGR. In our study population we could not find any relationship between insulin sensitivity and birth data in the IUGR group. However, caution should be used in the interpretation of these results due to the small size of our study group. Furthermore, attention should be paid to the greater variability in all clinical and biological variables studied in the IUGR group compared with that in our control group. This points to the likely heterogeneity of the IUGR group. We propose that this variability could depend on interactions between environmental factors and the genotype of the fetus (26). Such interactions make it difficult to determine the mechanisms underlying the long term metabolic changes associated with intrauterine growth retardation. In summary, we have demonstrated decreased insulinstimulated glucose uptake in young adults born with IUGR, with no evidence of major impairment of -cell function. In addition, our data suggest that insulin resistance in adipose tissue appears early in the course of insulin resistance in these subjects. Considering the key role of adipose tissue and lipid metabolism in the pathophysiology of insulin resistance and type 2 diabetes, it could be hypothesized that the time course of adipose tissue development in IUGR subjects is involved in the long term metabolic changes observed in IUGR adults. However, further exploration is required to investigate the pathophysiological mechanism underlying the development of insulin resistance in this population in the light of the known developmental abnormalities during fetal life, especially those occurring in adipose tissue of small for gestational age fetuses. Acknowledgments We acknowledge Dr. D. Chevenne for his supervision of laboratory analyses; the contributions of C. Traband, Dr. C. Collin, Dr. Boerher, and the laboratory staff at the Hôpital of the City of Haguenau; and the nursing and medical staff of the Clinical Investigation Unit of the Hospital R. Debré for supervising subjects. References 1. Hales CN, Barker DJP, Clark PMS, et al Fetal and infant growth and impaired glucose tolerance at age 64. Br Med J. 303: McCance DR, Pettitt DJ, Hanson RL, et al Birthweight and non insulin dependent diabetes: thrifty genotype, thrifty phenotype, or surviving small baby genotype? Br Med J. 308: Lithell HO, McKeigue PM, Berglund L, Mohsen R, Lithell UB, Leon DA Relation of size at birth to non-insulin dependent diabetes and insulin concentrations in men aged years. Br Med J. 312: Valdez R, Athens MA, Thompson GH, Bradshaw BS, Stern MP Birthweight and adult health outcomes in a biethnic population in the USA. Diabetologia. 37: Barker DJP, Hales CN, Fall C, Osmond C, Phipps K, Clark PMS Type 2 (non-insulin-dependent) diabetes mellitus hypertension and hyperlipidaemia (syndrome X): relation to reduced fetal growth. Diabetologia. 36: Papiernik E, Bouyer J, Dreyfus J, et al Prevention of preterm births: a perinatal study in Haguenau, France. Pediatrics. 76: Léger J, Lévy-Marchal C, Bloch J, et al Reduced final height and indications for insulin resistance in 20 year olds born small for gestational age: regional cohort study. Br Med J. 315: Widdowson EM, Southgate DAT, Hey EN Nutrition and metabolism of the fetus and infant. In: Visser HKA, ed. The Hague: Martinus Nijhoff: Lapillonne A, Braillon P, Chatelain PG, Delmas PD, Salle BD Body composition in appropriate and small for gestational age infants. Acta Paediatr. 86: McGarry JD Disordered metabolism in diabetes: have we underemphasized the fat component. J Cell Biochem. 55S: Reaven GM The fourth musketeer: from Alexandre Dumas to Claude Bernard. Diabetologia. 38: Mc Keigue PM, Lithell HO, Leon DA, et al Glucose tolerance and resistance to insulin-stimulated glucose uptake in men aged 70 years in relation to size at birth. Diabetologia. 41: De Fronzo RA, Tobin JD, Andres J, et al Glucose clamp technique: a method for quantifying insulin secretion and resistance. Am J Physiol. 237:E214 E Groop LC, Bonadonna RC, Del Prato S, et al Glucose and free fatty acid metabolism in non insulin dependent diabetes mellitus. Evidence for multiple sites of insulin resistance. J Clin Invest. 84: Khan SE, Prigeon RL, McCulloch DK, et al Quantification of the relationship between insulin sensitivity and -cell functions in human subjects. Evidence for a hyperbolic function. Diabetes. 42: WHO Study Group on Diabetes Mellitus Second report. Geneva: WHO; Tech Rep Ser Expert Committee on the Diagnosis and Classification of Diabetes Mellitus Report of the expert committee on the diagnosis and classification of diabetes mellitus. Diabetes Care. 20: Bergman RN, Finegood DT, Ader M, et al Assessment of insulin sensitivity in vivo. Endocr Rev. 6: Desai M, byrne CD, Zhang J, Petry CJ, Lucas A, Hales N Programming of the hepatic insulin-sensitive enzymes in offspring of rat dams fed with isocaloric protein restricted diet. Am J Physiol. 272:G1083 G Ferrannini E, Haffner SM, Mitchell BD, Stern MP Hyperinsulinemia, the key feature of cardiovascular and metabolic syndrome. Diabetologia. 34: Foster DW Insulin resistance, a secret killer? N Engl J Med. 320: Girard J Role of free fatty acids in insulin resistance of subjects with non-insulin dependent diabetes. Diabetes Metab. 21: Jaquet D, Léger J, Lévy-Marchal C, Oury JF, Czernichow P Ontogeny of leptin in human fetuses and newborns: effect of intra-uterine growth retardation on serum leptin concentrations. J Clin Endocrinol Metab. 83: Albertsson-Wikland K, Wennergren G, Wenergren M, Vibergsson G, Rosberg S Longitudinal follow-up of growth in children born small for gestational age. Acta Paediatr. 82: Robinson S, Walton RJ, Clark PM, Barker DJP, Hales CN, Osmond C The relation of fetal growth to plasma glucose in young men. Diabetologia. 35: Cambien F, Léger J, Mallet C, Lévy-Marchal C, Collin D, Czernichow P Angiotensin I-converting enzyme gene polymorphism modulates the consequences of in utero growth retardation on plasma insulin in young adult. Diabetes. 47: