Insulin Sensitivity and Secretory Response in Adults Born Preterm: The Helsinki Study of Very Low Birth Weight Adults

Transcription

1 ORIGINAL Endocrine ARTICLE Research Insulin Sensitivity and Secretory Response in Adults Born Preterm: The Helsinki Study of Very Low Birth Weight Adults Eero Kajantie, Sonja Strang-Karlsson, Petteri Hovi, Karoliina Wehkalampi, Jari Lahti, Nina Kaseva, Anna-Liisa Järvenpää, Katri Räikkönen, Johan G. Eriksson, and Sture Andersson National Institute for Health and Welfare (E.K., P.H., K.W., N.K., J.G.E.), Diabetes Prevention Unit, Helsinki, Finland; Children s Hospital (E.K., S. S-K., P.H., K.W., N.K., A-L.J., S.A.), Helsinki University Central Hospital and University of Helsinki, HUS, Helsinki, Finland; Department of Obstetrics and Gynecology (E.K.), MRC Oulu, Oulu University Hospital and University of Oulu, OYS, Oulu, Finland; Department of Behavioral Sciences (J.L., K.R.), University of Helsinki, Helsinki, Finland; Folkhälsan Research Centre (J.L., J.G.E.,), Helsinki, Finland; Department of General Practice and Primary Health Care (J.G.E.), University of Helsinki, Helsinki, Finland; Unit of General Practice (J.G.E.), Helsinki University Central Hospital, HUS, Helsinki, Finland; and Vasa Central Hospital (J.G.E.), Vasa, Finland Context: Preterm birth is associated with an increased risk of type 2 diabetes in adult life. The mechanisms are poorly known. Objective: We studied insulin sensitivity and secretion in adults born preterm at very low birth weight (VLBW; 1500 g). Design: Longitudinal Birth Cohort Study (Helsinki Study of Very Low Birth Weight Adults). Setting: The study was conducted at Uusimaa, Finland. Participants: One hundred seven adults born at VLBW and 100 controls born at term not small for gestational age (SGA), group-matched for sex, age, and birth hospital. The mean age was 25.0 years. Main Outcome Measures: We performed a 14-sample intravenous glucose tolerance test and calculated insulin sensitivity (Si), insulin secretory response (AIR), and disposition index, by Minimal Model (Minmod Millennium ). Results: Compared with controls, VLBW adults had lower Si (mean difference 11.9%, 95% CI 22.1 to 0.4%, adjusted for sex, age, and body mass index) and higher AIR (19.9%; %). The association with Si attenuated when further adjusted for height, parental diabetes, parental education, smoking, maternal smoking, hormonal contraception, and physical activity, but the association with AIR remained. Disposition index was similar. There was no difference between the 40 VLBW adults born SGA and the remaining VLBW adults. Conclusions: Adults born preterm at VLBW have lower insulin sensitivity than their term-born peers with a similar body size. In young adulthood, this remains compensated by higher insulin secretion. We suggest that this represents an early stage in the pathway leading to type 2 diabetes. Our results underline the importance of a healthy lifestyle and prompt vigilance in the screening of type 2 diabetes and impaired glucose tolerance in adults born preterm. (J Clin Endocrinol Metab 100: , 2015) ISSN Print X ISSN Online Printed in U.S.A. Copyright 2015 by the Endocrine Society Received August 11, Accepted October 2, First Published Online October 10, 2014 Abbreviations: BMI, body mass index; BP, blood pressure; DI, disposition index; OGTT, oral glucose tolerance test; Si, insulin sensitivity. 244 jcem.endojournals.org J Clin Endocrinol Metab, January 2015, 100(1): doi: /jc

2 doi: /jc jcem.endojournals.org 245 People born at low birth weight (LBW) have, as adults, an increased risk of type 2 diabetes (1, 2). This risk has been attributed to slow fetal growth, rather than short duration of gestation. However, recent studies have suggested that also people born preterm (before 37 weeks of gestation) are at an increased risk of type 2 diabetes in adult life, independent of their fetal growth (3 6). This is potentially important, because each year close to 15 million infants worldwide, 11 out of every 100, are born preterm (7, 8). The mechanisms that link preterm birth with type 2 diabetes are not known. Studies in young adults born preterm at very low birth weight (VLBW) (VLBW; 1500 g) or very low gestational age (VLGA) (VLGA; 32 weeks) suggest decreased insulin sensitivity as assessed by fasting insulin or HOMA-IR index (9), or by hyperinsulinemiceuglycemic clamp in a smaller study (10). Yet, few of the young adults born at VLBW fulfill the criteria for type 2 diabetes or impaired glucose tolerance (IGT) (9), suggesting that the reduced insulin sensitivity might be compensated by increased insulin secretory response. If so, these individuals would be more vulnerable to type 2 diabetes with cell reserves decreasing with age. Compensatory insulin response was suggested by a study in prepubertal children (11), but has not been studied in adults. With this background, we hypothesized that adults born at VLBW have lower insulin sensitivity than controls, compensated by increased insulin secretory response. We tested this hypothesis by performing intravenous (IV) glucose tolerance test (IVGTT) in the Helsinki Study of Very Low Birth Weight Adults. Materials and Methods Participants The participants come from the Helsinki Study of Very Low Birth Weight Adults, a follow-up cohort of all children born preterm at VLBW between 1978 and 1985 and discharged alive from the neonatal intensive care unit (NICU) of Children s Hospital at Helsinki University Central Hospital, the only tertiary neonatal center in the Uusimaa province of Finland. As control subjects, we selected for each VLBW infant, the next available infant of the same sex born at term that was not small for gestational age (SGA) (12) in the same hospital. The control group was recruited in conjunction with the first clinical examination at mean age of 22 years (9). That examination included an oral glucose tolerance test (OGTT) and measurement of body composition by dual x-ray absorptiometry that have been used as outcomes in previous studies (9, 13) but are also used for nonparticipant analyses and as covariates in the present paper. The present study is based on the second clinical visit at mean age of 25.0 years, which has been described in detail (14, 15). Briefly, of the 338 participants (166 VLBW and 172 control) of the first clinical examination, 25 were excluded from this visit because of having refused future contact (n 4), being abroad (n 11), untraced (n 2) or having a medical reason precluding the studies (n 8; pregnancy, medication, type 1 diabetes, developmental delay). We thus invited 159 VLBW and 154 control individuals of whom 113 (71.1%) and 105 (68.2%) participated. Of them, IVGTT was not performed on 4 VLBW and 3 control participants because of cannulation difficulties, on 1 VLBW and 1 control who had high fasting plasma glucose ( 6.9 mmol/l), and for 1 VLBW and 1 control who used systemic glucocorticoids. Thus, 107 VLBW and 100 term participants underwent IVGTT. The study was approved by the Ethics Committee of Children s and Adolescents Diseases and Psychiatry at Helsinki and Uusimaa Hospital District. Each subject signed an informed consent. Sharing of materials may require a new ethics approval and/or new consent. Comparison of participants with nonparticipants To facilitate comparison of the present study with our previous results on OGTT (9), we compared the 107 VLBW and 100 control participants of the present study with the 59 VLBW and 72 control individuals who participated in the previous OGTT study but not in the present IVGTT study. This comparison is shown in Supplemental Table 1. The participants of the present study were slightly younger than the nonparticipants. VLBW participants were less likely to smoke than VLBW nonparticipants. Control participants had higher fasting insulin than control nonparticipants; this difference was not observed among the VLBW. Fasting and 2-h glucose and 2-h insulin concentrations were similar as were all peri- and neonatal characteristics and measurements of body size and composition and physical activity. Intravenous glucose tolerance test The standard 14-sample IV glucose tolerance test was started between 7 and 12 h (for one participant at 13:30 h) after an overnight fast. The participants were instructed not to smoke during the fast. During the test, the participant was lying in a semisupine position. A fasting blood sample was drawn after an IV cannula had been inserted. The participant was given 0.66 g/kg bolus of IV glucose, and blood samples were drawn at baseline, 2, 4, 8, 19, 22, 25, 30, 40, 50, 70, 100, 120, and 180 min after glucose. At 20 min, a 0.03 IU/kg bolus of IV insulin was given. Plasma insulin concentrations were determined by Immunotech Insulin(e) IRMA kit (A Beckman coulter company) and glucose concentrations by the enzymatic hexokinase method. We used MinMod Millennium 6.02 (MINMOD Inc.) to calculate indices of glucose metabolism. As main outcome variables, we used fasting insulin, insulin sensitivity (Si, quantifying the capacity of insulin to promote glucose disposal), insulin secretory response [AIR, calculated between baseline and 10 min as the insulin area under the curve (AUC) over baseline], and disposition index (DI, calculated as AIR*Si, an estimate of beta cell function). In addition, we assessed glucose effectiveness (Sg, capacity of glucose to mediate its own disposal). Measurement of covariates Prenatal and neonatal data came from medical records. Preeclampsia was defined by standard criteria (16), and chorioamnionitis, based on maternal fever, leukocytosis, and elevated C-

3 246 Kajantie et al Insulin Sensitivity and Secretion in VLBW Adults J Clin Endocrinol Metab, January 2015, 100(1): reactive protein (CRP), and bronchopulmonary dysplasia (BPD), based on Northway criteria (17), had been diagnosed by one neonatologist (A.L.J.). Height and weight were measured and body mass index (BMI) was calculated as kg/m 2. As a measure of childhood socio-economic status we used educational attainment of the higher-educated parent, obtained by questionnaire together with data on parental history of diabetes, current smoking of the participant, and use of hormonal contraception. As a measure of physical activity, we used energy expenditure from total physical activity as determined by the 30-item 12-month modified Kuopio Ischemic Heart Disease Risk Factor Study Questionnaire (15). Data analysis Results of the IVGTT and plasma insulin concentrations were log-transformed to attain normality. For comparison of descriptive data between VLBW participants and controls, we used the t-test and 2 test. We used multiple linear regression to adjust these comparisons for covariates. Because the associations between VLBW birth and main outcomes were similar in women and men (p values for interaction sex*vlbw birth 0.4), we report pooled analyses that include women and men. Results Clinical characteristics Clinical characteristics of the participants are shown in Table 1. These include the results of the OGTT and body composition at the previous examination at a mean age of 22.4 years in the sample of this study; for the whole source cohort, they have been described (9). Two-hour glucose and insulin concentrations were higher in VLBW participants, while there was no difference in fasting glucose and insulin. This result was similar in multiple regression models that adjusted for age, sex, and BMI. The differences in 2-h glucose and insulin were attenuated when further adjusted for other covariates (height, parental educational attainment, parental diabetes, maternal smoking in pregnancy, current smoking of the participant, use of hormonal contraception, and physical activity). Insulin sensitivity and secretory response We then assessed the results of the IVGTT. Geometric means of the fasting insulin concentrations were 4.77 (geometric SD 1.79) mmol/l for the VLBW and 4.72 (2.42) mmol/l for the control group (P.9). Si was 4.39 (1.64) 10 4 /min 1 (mu/l) and 4.70 (1.67) 10 4 / min 1 (mu/l) (P.3), Sg (1.79) and (1.88) (P.3), AIR 427 (1.74) and 381 (1.78) mu/l (P.1), and disposition index 1878 (1.69) and 1791 (1.74) (P.5). Table 2 shows correlation coefficients between different measures of glucose metabolism in the IVGTT and also a comparison with the OGTT measurements previously reported on (9). Associations of covariates with the outcomes are shown in Supplemental Table 2. We assessed mean differences of the main outcomes between VLBW and term groups, adjusting for covariates by linear regression (Figure 1 and Supplemental Table 3). When adjusted for BMI, VLBW adults had lower insulin sensitivity and higher insulin secretion, which together resulted in a similar disposition index. The association with insulin sensitivity was attenuated when adjusted for all covariates in the full model, whereas the difference in insulin secretion was little changed after this adjustment. There was no difference in fasting insulin (P.26 in all models) or Sg (P.18). We also assessed mean differences adjusting for lean body mass and body fat percentage instead of BMI. They were measured on average 3 years earlier and were available for 102 VLBW and 87 control participants. The results were similar: mean difference in insulin sensitivity was 12.1% ( %) and insulin secretion 20.4% ( %). When we excluded the 25 VLBW and 20 control participants whose BMI had changed more than 2 kg/m 2, these mean differences were 11.9% ( %) and 15.8% ( %). We repeated the analyses after excluding 7 VLBW and 3 control participants with a chronic disorder (detailed in Table 1) that could affect physical fitness and thus glucose metabolism. The results were little changed: mean differences adjusted for sex, age, and BMI (analogous to Model 2 in Supplemental Table 3) were 11.3% (95% CI %) for insulin sensitivity, 18.1% ( %) for insulin secretion, and 6.2% ( %) for disposition index. Perinatal and neonatal characteristics within the VLBW group We compared the 40 participants born VLBW SGA with the 67 participants born VLBW appropriate for gestational age (AGA). There was no difference in any of the markers of glucose metabolism (P.3 in all models). We then assessed the continuous associations of length of gestation and birth weight SD score with the outcomes by including both in a simultaneous regression with covariates of Model 2. Birth weight SD score was not related to any outcome, but a one week higher length of gestation was associated with 8.6% ( %) higher fasting insulin and 6.5% ( %) higher insulin secretion. When this analysis was restricted to those 32 born AGA, or those 97 born before 32 weeks of gestation, the associations attenuated and were no longer statistically significant. When analyzing the effect of body size attained at term, a one SD higher weight at 40 postmenstrual weeks (available for 79 participants) was associated with 13.4% ( %) higher AIR; this association remained similar when further adjusted for birth weight SD score. There

4 doi: /jc jcem.endojournals.org 247 Table 1. Characteristics of the Study Participants VLBW n 107 Term n 100 Men, n (%) 50 (46.7%) 42 (42.0%).5 Prenatal and postnatal Maternal Smoking in Pregnancy 17 (15.9%) 14 (14.0%).7 Maternal Pre-Eclampsia, n (%) 26 (24.3%) 9 (9.0%).003 Maternal Chorioamnionitis, n (%) 9 (8.4%) 0 Length of Gestation (weeks) 29.3 (2.3) 40.1 (1.1).0001 Birth Weight (g) 1128 (219) 3601 (484).0001 Birth Weight SD Score 1.3 (1.5) 0.0 (1.0).0001 Weight SD Score at Term a 2.5 (1.2) N/A Small for Gestational Age, n (%) 40 (37.4%) 0 Bronchopulmonary Dysplasia, n (%) 23 (21.5%) 0 Current Age (y) 25.0 (2.1) 25.0 (2.2).9 Height (cm) Women (7.6) (6.1).003 Men (7.7) (6.2).0001 Body Mass Index (kg/m 2 ) Women 22.5 (3.9) 23.4 (5.0).3 Men 23.0 (3.6) 24.4 (3.3).06 Current Smoking, n (%) 24 (22.4%) 40 (40.0%).006 Total Physical Activity (Energy Expenditure; kcal/week) b (15) 1059 (2.8) 1459 (2.4).02 Oral Contraceptives, n (% c ) 26 (45.6%) 27 (46.6%).9 Inhaled Glucocorticoids, n (%) 11 (10.3%) 4 (4.0%).2 Neurosensory Impairment or History of Severe Disease, n (%) d 7 (6.5%) 3 (3.0%).2 Parental Education.3 Basic or Less 10 (9.3%) 6 (6.0%) Secondary 21 (19.6%) 18 (18.0%) Lower-Level Tertiary 43 (40.2%) 43 (43.0%) Upper-Level Tertiary 33 (30.8%) 33 (33.0%) Maternal Diabetes, n (%) 4 (3.7%) 4 (4.0%).9 Paternal Diabetes, n (%) 8 (7.5%) 5 (5.0%).5 Measurements during previous clinical examination e Age (y) 22.1 (2.1) 22.2 (2.1) Lean body mass f Women 39.0 (5.1) 42.7 (6.3).001 Men 53.7 (7.4) 61.6 (8.0).0001 Body Fat Percentage f Women 31.0 (6.0) 32.2 (6.3).3 Men 20.6 (6.6) 20.1 (6.0).7 Oral Glucose Tolerance Test Fasting Plasma Glucose (mmol/l) 4.73 (0.37) 4.68 (0.39).3 2-h Glucose 5.55 (1.30) 5.09 (1.25).01 Fasting Plasma Insulin (mmol/l) g 5.55 (1.68) 5.38 (1.64).7 2-h Insulin g (1.93) (2.36).02 a Available for 79 VLBW participants. b Available for 96 VLBW and 97 term participants. c Of women. d Four VLBW subjects had cerebral palsy, 3 had mental retardation (these numbers include one subject with both cerebral palsy and mental retardation), and 1 was blind. One of the term-born had mental retardation, 1 a history of Hodgkin s disease, and 1 a history of brain tumor. e These data have been previously reported on and are shown here for comparison. p f Available for 102 VLBW and 87 term participants. g Geometric means and SDs. was no relationship between any of the outcomes and maternal pre-eclampsia (P.2), maternal smoking in pregnancy (P.13), or chorioamnionitis (P.11). The 23 VLBW participants with a history of BPD had 19.3% ( %) lower fasting insulin 27.2% ( %) lower insulin secretion and 23.8% ( %) lower disposition index than the 84 VLBW participants with no such history; there was no difference in insulin sensitivity (P.4) or Sg (P.5). Associations with insulin secretion remained similar when adjusted for length of gestation or weight SD score at 40 postmenstrual weeks; other associations attenuated and were not statistically significant.

5 248 Kajantie et al Insulin Sensitivity and Secretion in VLBW Adults J Clin Endocrinol Metab, January 2015, 100(1): Table 2. Correlation Coefficients Between Measurements of Glucose Metabolism IVGTT at 25.0 y OGTT at 22.1 y Ins 0 Si Sg AIR DI IVGTT at 25.0 y Ins 0 Si 0.40 a Sg a AIR 0.56 a 0.50 a 0.03 DI 0.22 b 0.41 a 0.36 a 0.59 a OGTT at 22.1 y Gluc Ins a 0.38 a a a Gluc a 0.29 a a 0.17 c 0.27 a Ins a 0.42 a 0.22 b 0.26 a a 0.66 a IVGTT, intravenous glucose tolerance test; OGTT, oral glucose tolerance test; Si, insulin sensitivity; AIR, insulin secretory response; DI, disposition index (Si*AIR); Gluc 0 and Gluc 2, Fasting and 2-h glucose; Ins 0 and Ins 2, Fasting and 2-h insulin; Sg,glucose effectiveness. a P b P c P Discussion Mean difference VLBW-term (%) Insulin sensitivity (SI) Acute insulin response (AIR) Disposition index (DI) 1 Adjusted for: 1. sex, age, BMI 2. 1+height, smoking, 1 parental education, parental diabetes, maternal smoking in pregnancy, current smoking, hormonal contraception 3. 2+physical activity Figure 1. Mean difference in glucose metabolism between VLBW and control adults, adjusted for covariates. The bars stand for percent difference (with 95% confidence interval) of VLBW as compared with the control group. Corresponding numbers are shown in Supplemental Table 3 in the online supplement. Gluc 0 Ins 0 Gluc 2 Ins 2 We found that, as compared with their term-born peers with a similar body size, young adults born at VLBW have lower insulin sensitivity compensated by higher insulin secretory response, resulting in a similar disposition index. The differences in insulin sensitivity were, however, relatively modest: 12%, corresponding to 0.25 SD, when accounting for the lower BMI of VLBW adults (9). They attenuated to nonsignificance when adjusted for socioeconomic position, parental history of diabetes and lifestyle. Differences in insulin secretion were more pronounced, around 20% or 0.31 SD between groups. Our study is consistent with previous studies showing poorer insulin sensitivity in children and young adults born at VLBW/VLGA and extends them by showing compensatory insulin release present in early adulthood. Lower insulin sensitivity was found in 57 young adults born at VLBW/VLGA and 30 controls by hyperinsulinemic euglycemic clamp (10). Consistent with these studies, in an OGTT study in the source cohort of the present study, we previously showed higher 2-h insulin and fasting insulin concentrations and HOMA-IR index among VLBW adults. However, few individuals fulfilled the criteria for type 2 diabetes or IGT (9), which lead us to hypothesize that the reduced insulin sensitivity is compensated by increased insulin secretion. This hypothesis was also supported by another study of 26 VLBW adults and 24 controls who underwent a 5-sample OGTT (18). More accurate estimates of insulin secretion are obtained by IVGTT, which was used in a study of children aged 4 10 years. In that study, the 50 VLGA children had reduced insulin sensitivity and compensatory insulin release than did 22 controls born at term (11). Our study shows for the first time that the compensatory insulin release, concealing the reduced insulin sensitivity, extends at least to young adult life. There is one previous study that used IVGTT in young adults born preterm. That study found no association between length of gestation and insulin sensitivity among 305 participants. The result is, however, difficult to compare with ours as the study had a heterogeneous control group that included participants recruited for being born SGA (19).

6 doi: /jc jcem.endojournals.org 249 Our findings were remarkably similar for VLBW adults regardless of whether they were born SGA or AGA. This phenomenon has been seen in other studies assessing glucose regulation (9 11) or other cardiometabolic risk factors including body composition (9) and blood pressure (BP) (9, 20, 21). Prenatal environment of VLBW subjects born SGA is frequently characterized by placental dysfunction resulting in lack of nutrients and oxygen and thus is very different from that of those born AGA, while postnatal challenges are, to a great extent, similar. We have previously proposed that the similar cardiometabolic risk factors in SGA and AGA groups suggest that differences in these risk factors may be a result of postnatal events or prematurity itself, rather than prenatal conditions (9). Our present findings in relation to postnatal conditions reinforce this suggestion: despite similar insulin sensitivity, insulin secretory response was lower in VLBW participants with a more adverse postnatal history as indicated by BPD or slow growth between birth and term. Several mechanisms can potentially underlie our findings. Muscle is a key target tissue of insulin, and reduced insulin sensitivity could be due to lower muscle mass. Lean body mass, which consists mostly of muscle, was substantially lower in VLBW adults, but allowing for this did not explain the present difference in insulin sensitivity or our previous findings of impaired glucose regulation in OGTT (9). VLBW adults have higher resting energy expenditure than expected based on their lean body mass (22). This suggests the presence of metabolically more active tissue, which could counteract the reduction in insulin sensitivity. The difference in fat distribution remains a possibility. A study comparing 23 young adults born at 33 weeks or less with 25 controls found increased visceral, intramyocellular, and intrahepatic fat, which may all contribute to reduced insulin sensitivity (23). Conditions associated with VLBW birth might cause epigenetic changes that persist to adult life (24) and affect glucose regulation. For example, there are differences in different methylation of the IGF2 locus between VLBW adults and controls (25). This locus is involved in glucose-insulin metabolism, but the difference has thus far been shown only in leukocytes and awaits confirmation in tissues involved in glucose regulation. Finally, differences in lifestyle could be one explanation. Consistent with this, our results were slightly attenuated when adjusted for physical activity. In addition, we have previously reported that VLBW adults have a less healthy diet including lower intakes of fruit, vegetables, and milk products (26). However, the present study included no dietary assessment and we are thus unable to assess dietary factors as a potential mechanism. The low insulin sensitivity in VLBW adults was compensated by increased insulin secretory response. This limits the increase in plasma glucose concentrations and as young adults most individuals remain normoglycemic. While existing follow-up studies suggest that in high-risk populations, low rather than high secretory response predicts IGT among normoglycemic individuals. The follow-up in these studies is limited to approximately 10 years, and individuals converting to IGT or type 2 diabetes may initially represent a later stage of the development of these conditions. In these studies, low insulin sensitivity and low secretory response are both independent predictors of IGT and type 2 diabetes (27). Therefore, we believe that the combination of low insulin sensitivity and compensatory insulin response represents an earlier stage on the pathway leading to IGT and type 2 diabetes (28). Consistent with this, in a randomized study of intensive lifestyle intervention of individuals with IGT, the intervention was followed by improved insulin sensitivity and reduction in first-phase insulin-secretion, as assessed by OGTT (29). This suggests that lifestyle interventions are important in reversing the early pathophysiological changes and preserving -cell function and underlines the importance of healthy lifestyle in adults born at VLBW. The main strength of the study is the use of IVGTT which, while labor-intensive, gives an accurate estimate of insulin sensitivity and secretion, and the well-characterized population of VLBW adults and carefully groupmatched controls. As to limitations, selection bias remains a possibility. Although 70% of those invited participated, controls who participated in the present study had higher fasting insulin in the previous clinical examination than controls who did not participate, indicating lower insulin sensitivity. This may have attenuated the difference in insulin sensitivity, which was smaller than we expected, corresponding to 0.25 SD; a previous study using hyperinsulinemic euglycemic clamp found a difference of approximately 0.8 SD (10), although the groups in that study were smaller and confidence intervals relatively wide. Moreover, a detailed nonparticipation analysis (9) of the previous clinical examination, the source population of the present study, showed that VLBW adults with cerebral palsy were less likely to participate. This may have further attenuated the difference in insulin sensitivity. However, the nonparticipation analyses of the previous and the present studies raised little concern for any other type of participation bias. As to confounding, our data allowed us to adjust for a range of factors including family socio-economic position, family history of diabetes, and maternal smoking in pregnancy, and other covariates, such as smoking and physical activity. These factors do explain a part of the difference in insulin sensitivity, whereas the differences in insulin secretion remain remarkably similar. Moreover, data on body composition

7 250 Kajantie et al Insulin Sensitivity and Secretion in VLBW Adults J Clin Endocrinol Metab, January 2015, 100(1): were based on a measurement 3 years before and should thus be interpreted with caution. We conclude that adults born preterm at VLBW have lower insulin sensitivity than their term-born peers with a similar adult body size, although in our study this difference was relatively small and in part explained by confounding factors. In young adulthood, this remains compensated by higher insulin secretion. The combination of low insulin sensitivity and compensatory insulin secretion may be an early phase in the pathway leading to type 2 diabetes and its complications. Our results underline the importance of promotion of healthy lifestyle and prompt vigilance in the screening of type 2 diabetes and impaired glucose tolerance in adults born preterm. Acknowledgments Address all correspondence and requests for reprints to: Eero Kajantie, National Institute for Health and Welfare, PL 30, Helsinki, Finland. eero.kajantie@helsinki.fi. This work was supported by Academy of Finland, Emil Aaltonen Foundation, Finnish Foundation for Pediatric Research, Finnish Special Governmental Subsidiary for Health Sciences (evo), Jalmari and Rauha Ahokas Foundation, Juho Vainio Foundation, Medical Societies of Finland (Duodecim and Finska Läkaresällskapet), National Graduate School of Clinical Investigation, Novo Nordisk Foundation, Otto A. Malm Donation Fund, Perklén Foundation, Päivikki and Sakari Sohlberg Foundation, Research Foundation for the Orion Corporation, Signe and Ane Gyllenberg Foundation, Sigrid Jusélius Foundation, Waldemar von Frenckell Foundation, Wilhelm and Else Stockmann Foundation, Victoriastiftelsen, and Yrjö Jahnsson Foundation. Disclosure Summary: The authors have nothing to disclose. References 1. Eriksson JG, Osmond C, Kajantie E, Forsén TJ, Barker DJ. Patterns of growth among children who later develop type 2 diabetes or its risk factors. Diabetologia. 2006;49: Whincup PH, Kaye SJ, Owen CG, et al. Birth weight and risk of type 2 diabetes: A systematic review. JAMA. 2008;300: Lawlor DA, Davey Smith G, Clark H, Leon DA. The associations of birthweight, gestational age and childhood BMI with type 2 diabetes: findings from the Aberdeen Children of the 1950s cohort. Diabetologia. 2006;49: Kaijser M, Bonamy AK, Akre O, et al. Perinatal risk factors for diabetes in later life. Diabetes. 2009;58: Kajantie E, Osmond C, Barker DJ, Eriksson JG. Preterm birth a risk factor for type 2 diabetes? The Helsinki birth cohort study. Diabetes Care. 2010;33: Pilgaard K, Færch K, Carstensen B, et al. Low birthweight and premature birth are both associated with type 2 diabetes in a random sample of middle-aged danes. Diabetologia. 2010;53: Blencowe H, Cousens S, Oestergaard MZ, et al. National, regional, and worldwide estimates of preterm birth rates in the year 2010 with time trends since 1990 for selected countries: A systematic analysis and implications. Lancet. 2012;379: NCHS. National center for health statistics, final natality data Retreived feb 6th, 2012 from Hovi P, Andersson S, Eriksson JG, et al. Glucose regulation in young adults with very low birth weight. N Engl J Med. 2007;356: Rotteveel J, van Weissenbruch MM, Twisk JW, Delemarre-Van de Waal HA. Infant and childhood growth patterns, insulin sensitivity, and blood pressure in prematurely born young adults. Pediatrics. 2008;122: Hofman PL, Regan F, Jackson WE, et al. Premature birth and later insulin resistance. N Engl J Med. 2004;351: Pihkala J, Hakala T, Voutilainen P, Raivio K. Uudet suomalaiset sikiön kasvukäyrät [Characteristic of recent fetal growth curves in Finland]. Duodecim. 1989;105: Hovi P, Andersson S, Järvenpää AL, et al. Decreased bone mineral density in adults born with very low birth weight. PLoS Med. 2009; 6:e Pyhälä R, Räikkönen K, Pesonen AK, et al. Parental bonding after preterm birth: child and parent perspectives in the Helsinki study of very low birth weight adults. J Pediatr. 2011;158:251 6.e Kaseva N, Wehkalampi K, Strang-Karlsson S, et al. Lower conditioning leisure-time physical activity in young adults born preterm at very low birth weight PLoS One. 2012;7:e Report of the national high blood pressure education program working group on high blood pressure in pregnancy. Am J Obstet Gynecol 2000;183:S1 S Northway WH Jr, Rosan RC, Porter DY. Pulmonary disease following respirator therapy of hyaline-membrane disease. Bronchopulmonary dysplasia. N Engl J Med. 1967;276: Smith CM, Wright NP, Wales JK, et al. Very low birth weight survivors have reduced peak bone mass and reduced insulin sensitivity. Clin Endocrinol (Oxf). 2011;75: Willemsen RH, Leunissen RW, Stijnen T, Hokken-Koelega AC. Prematurity is not associated with reduced insulin sensitivity in adulthood. J Clin Endocrinol Metab. 2009;94: Doyle LW, Faber B, Callanan C, Morley R. Blood pressure in late adolescence and very low birth weight. Pediatrics. 2003;111: Hovi P, Andersson S, Räikkönen K, et al. Ambulatory blood pressure in young adults with very low birth weight. J Pediatr. 2010; 156:54 59.e Sipola-Leppänen M, Hovi P, Andersson S, et al. Resting energy expenditure in adults born preterm - the Helsinki study of very low birth weight adults. PLoS One. 2011;6:e Thomas EL, Parkinson JR, Hyde MJ, et al. Aberrant adiposity and ectopic lipid deposition characterize the adult phenotype of the preterm infant. Pediatr Res. 2011;70: Cruickshank MN, Oshlack A, Theda C, et al. Analysis of epigenetic changes in survivors of preterm birth reveals the effect of gestational age and evidence for a long term legacy. Genome Med. 2013;5: Wehkalampi K, Muurinen M, Bruce Wirta S, et al. Methylation of igf2 locus is altered 20 years after preterm birth at very low birth weight. PLoS One. 2013;8:e Kaseva N, Wehkalampi K, Hemiö K, et al. Diet and nutrient intake in young adults born preterm at very low birth weight. J Pediatr. 2013;163: Weyer C, Tataranni PA, Bogardus C, Pratley RE. Insulin resistance and insulin secretory dysfunction are independent predictors of worsening of glucose tolerance during each stage of type 2 diabetes development. Diabetes Care. 2001;24: Defronzo RA. Banting lecture. From the triumvirate to the ominous octet: A new paradigm for the treatment of type 2 diabetes mellitus. Diabetes. 2009;58: de Mello VD, Lindström J, Eriksson J, et al. Insulin secretion and its determinants in the progression of impaired glucose tolerance to type 2 diabetes in impaired glucose-tolerant individuals: The Finnish diabetes prevention study. Diabetes Care. 2012;35: