ORIGINAL ARTICLE. S Park 1, M-Y Kim 2, SH Baik 3, J-T Woo 4, YJ Kwon 1, JW Daily 1, Y-M Park 5, J-H Yang 2 and S-H Kim 6

Transcription

1 European Journal of Clinical Nutrition (2013) 67, & 2013 Macmillan Publishers Limited All rights reserved /13 ORIGINAL ARTICLE Gestational diabetes is associated with high energy and saturated fat intakes and with low plasma visfatin and adiponectin levels independent of prepregnancy BMI S Park 1, M-Y Kim 2, SH Baik 3, J-T Woo 4, YJ Kwon 1, JW Daily 1, Y-M Park 5, J-H Yang 2 and S-H Kim 6 BACKGROUND/OBJECTIVES: Gestational diabetes mellitus (GDM) risk factors are well established for Caucasians, but not for Asians. We hypothesized that nutrient intakes, plasma adipokines and/or gestational hormones might be linked to GDM development among pregnant Korean women. This study sought to identify new risk factors for GDM and adverse pregnancy outcomes according to body weight at prepregnancy. SUBJECTS/METHODS: All subjects were pregnant women visiting the Cheil General Hospital and Women s Healthcare Center between June 2006 and March Non-GDM (n ¼ 531) and GDM (n ¼ 215) participants were divided into normal-weight and overweight groups according to prepregnancy body mass index (BMI) above or below 23 kg/m 2 at 24 28th week of gestation. At that time, glucose tolerance, insulin resistance as homeostatic model assessment for insulin resistance, insulin secretory capacity as homeostatic model assessment for b-cell function, anthropometric measurement, nutrient intakes, and plasma levels of adipokines and gestational hormones were determined. RESULTS: GDM women gained more weight in early pregnancy than non-gdm among normal-weight women. GDM was mainly associated with increased insulin resistance in overweight women and decreased insulin secretory capacity in normal-weight women. Plasma visfatin and adiponectin were lower and progesterone levels higher in GDM than non-gdm independent of BMI while plasma resistin levels were higher in non-gdm, but not GDM, overweight women. Energy and saturated fat intakes were higher in GDM independent of body weight, whereas taurine intakes were lower in GDM than non-gdm only in normal-weight women. CONCLUSIONS: Low visfatin and adiponectin and high progesterone levels in the circulation and high energy and saturated fat intakes were common risk factors for GDM and pregnancy outcome such as large for gestational age. Daily reference intakes for energy and fat during pregnancy need to be re-evaluated according to prepregnancy BMI. European Journal of Clinical Nutrition (2013) 67, ; doi: /ejcn Keywords: gestational diabetes; weight gain; BMI; adipokines; prolactin; progesterone INTRODUCTION Well-characterized risk factors for gestational diabetes mellitus (GDM) include overweight and obesity, older maternal age, previous GDM and family history of diabetes; however, these known risk factors are not reliable predictors of the occurrence of GDM in individuals or even in some populations. 1 GDM screening using traditional risk factors cannot detect almost half of GDM women. 2 Ethnicity is also known to affect the occurrence and severity of GDM and type 2 diabetes; a possible consequence of low insulin secretory capacity of pancreatic b-cells among high-risk populations, including Asians. 3,4 Specific nutritional risk factors have not been clearly identified for GDM development. Adipokines and inflammatory cytokines are also known to impact the development of both type 2 diabetes and GDM. 5 Inflammation can result in the destruction of pancreatic b-cells and may also contribute to the progression of insulin resistance. 6 Visfatin, a newly identified adipokine, is essential for normal insulin action but at higher levels, as seen with obesity, it is associated with impaired insulin secretion and insulin resistance. 7 Adiponectin, which decreases with increased adipose tissue deposits although it is secreted by adipocytes, is a wellcharacterized insulin sensitizer; low levels of high molecular weight adiponectin have been associated with increased incidence of type 2 diabetes in Japanese men. 8 Furthermore, changes in gestational hormonal status during pregnancy may contribute to dysregulation of glucose homeostasis. 9 Changes in insulin resistance and insulin secretory capacity during pregnancy are concomitant with increases in circulating prolactin, progesterone and estrogen levels Department of Food and Nutrition, Hoseo University, Asan, Korea; 2 Department of Obstetrics and Gynecology, Cheil General Hospital and Women s Healthcare Center, Kwandong University College of Medicine, Seoul, Korea; 3 Department of Internal Medicine, Korea University College of Medicine, Seoul, Korea; 4 Department of Endocrinology and Metabolism, Kyung Hee University School of Medicine, Seoul, Korea; 5 Department of Preventive Medicine, The Catholic University of Korea College of Medicine, Seoul, Korea and 6 Department of Endocrinology and Metabolism, Cheil General Hospital & Women s Healthcare Center, Kwandong University College of Medicine, Seoul, Korea. Correspondence: Dr S-H Kim, Department of Endocrinology and Metabolism, Cheil General Hospital and Women s Healthcare Center, Kwandong University, College of Medicine, 1-19 Mook-Dong Jung-Gu, Seoul, Korea. hoonie.kim@cgh.co.kr Received 23 March 2012; revised 18 November 2012; accepted 19 November 2012

2 Although obesity is the most consistent risk factor for type 2 diabetes and GDM, nearly half of Korean GDM is not associated with obesity. We hypothesized that the development of GDM in normal-weight and overweight women who followed weight gain guidelines during pregnancy, but still developed GDM, might be explained by secretion profiles of adipokine/cytokines and gestational hormones, and by diet. Therefore, our overall objective was to identify major risk factors that would help identify women at risk for GDM and facilitate the establishment of interventions to prevent the onset of GDM and improve pregnancy outcome. SUBJECTS AND METHODS Subjects and screening for GDM All pregnant women visiting the Cheil General Hospital and Women s Healthcare Center (Seoul, Korea) between June 2006 and March 2009 for their first visit at 8 10 weeks of pregnancy took a short survey recording prepregnancy body weights, medication use, and measurements of body weight and blood glucose levels, and received a general education session for managing pregnancy including food intake. The pregnant women were asked to meet weight gain guidelines according to their body mass index (BMI) by regulating food intake during the remaining pregnancy to reduce obesity-related complications. GDM screening at the 24 28th week of gestation used a universal twostep GDM screening program with a 50 g glucose challenge test. Women with a negative result were deemed non-gdm and those testing positive were administered a 100 g, 3-h oral glucose tolerance test and GDM was identified according to criteria outlined by Carpenter and Coustan. 10 Of the pregnant women screened, 554 out of (3.8%) normal-weight or underweight women and 532 of the 3643 (14.6%) overweight or obese women were diagnosed with GDM; 531 non-gdm and 215 GDM women enrolled in the study. Blood collections and questionnaire surveys were obtained at 25.1±0.95 weeks of gestation and gestational age was not significantly different among the groups. All GDM women attended at least three nutritional and lifestyle education sessions on maintaining body weight and glucose control, and all GDM women maintained normal blood glucose levels using diet, exercise and/or insulin injection and non-gdm women also maintained normoglycemia. The study protocol was approved by the Ethical Committee of the Institutional Review Board of Cheil General Hospital and all subjects provided written informed consent on the GDM screening day. Subjects then completed an interviewer-administered questionnaire and were assessed for blood pressure, height and body weight. Blood was collected for assessing lipid and adipokine profiles and gestational hormones. In addition, food intakes were determined by a skilled dietician using the 24-h recall method. Questionnaires were used to obtain the following information: age, prepregnancy weight, obstetric history including gestational age and parity, and family history of diabetes among firstdegree relatives. Laboratory measurements, physiologic indexes and nutrient intakes at 24 28th week of pregnancy Plasma glucose concentrations were measured by the glucose oxidase method using a YSI 2300 STAT (YSI; Yellow Springs, OH, USA). Plasma insulin, prolactin and progesterone concentrations were measured using a human-specific radioimmunoassay kit (Linco Research, St Charles, MO, USA). Homeostatic model assessments for insulin resistance (HOMA-IR) and b-cell function (HOMA-B) were calculated using the equations developed by Matthews et al. 11 Hemoglobin A 1c in GDM subjects was measured using a Variant II HbA 1c dual kit from Bio-Rad Laboratories (Hercules, CA, USA). Circulating adipokines (adipsin, adiponectin, resistin, leptin and visfatin) were measured on a multiplex suspension array system (Bio-Plex, Bio-Rad Laboratories). Plasma lipid profiles (total, high-density lipoprotein and lowdensity lipoprotein cholesterol, and triglyceride) and C-reactive protein were determined by enzymatic colorimetry. Plasma low-density lipoprotein cholesterol was calculated using the Friedewald equation. Nutrient intakes were analyzed from the food intakes using CAN-PRO (version 3; Korean Nutrient Society, Seoul, Korea) and some were calculated as the percentage of the Korean Dietary Reference Intake for pregnant women of similar age. 12 Statistical analysis Data are expressed as mean±s.d. for normal distribution or median (interquatile range) for skewed distribution. Statistical analyses were performed using SAS Version 9.1. (SAS Institute Inc., Cary, NC, USA). Subjects were divided into two groups according to prepregnancy BMI with a cutoff value of 23 for overweight according to the WHO expert consultation because Asians generally have a higher percentage of body fat than Caucasians of the same age, sex and BMI and a substantial risk of type 2 diabetes and cardiovascular disease among normal-weight individuals when using a BMI cutoff point of 25 kg/m 2. 13,14 In order to determine the effects of GDM and prepregnancy BMI, we conducted twoway analysis of variance according to GDM and prepregnancy BMI for continuous variables and a w 2 test or Fisher s exact test for categorical variables. Adjustments were made for the following variables: maternal age, gestational age, family history and parity. In addition, if a difference among the four groups was found in a continuous variable, we utilized post hoc Tukey s testing to determine which groups were significantly different while Cochran Mantel Haenszel test for a categorical variable was performed. Kruskal Wallis test was used for the parameters with a skewed distribution. Backward stepwise logistic regression analysis was used to determine which factors in prepregnancy and during pregnancy were independently associated with normal-weight and overweight GDM women. In addition, factors independently affecting large for gestational age (LGA) in normal-weight and overweight GDM women were assessed by backward regression analysis. A P-valueo0.05 was considered statistically significant. RESULTS Characteristics of pregnant women Mean prepregnancy BMIs of non-gdm and GDM subjects were 19.8±1.5 and 24.7±3.2 kg/m 2, respectively, and GDM was positively correlated with BMI. GDM women were older if overweight and were shorter than non-gdm regardless of weight. Prepregnancy BMI of normal-weight women was not different between non-gdm and GDM women, but was higher at the 24 28th week in GDM women (Supplementary Table 1). In overweight women, prepregnancy and 24 28th week BMIs were greater in the GDM group than the non-gdm group (Supplementary Table 1). However, BMI at delivery was not significantly different between the non-gdm and GDM women. Weight gain during pregnancy was associated with prepregnancy BMI and GDM (Supplementary Table 1). Women with a higher parity were more overweight regardless of GDM and women with a family history of diabetes exhibited a higher prevalence of GDM. Glucose and lipid metabolism at 24- to 28-week pregnancy Plasma glucose levels while fasting and at 1 h after a 50 g glucose challenge were not affected by body weight (Table 1). Fasting plasma insulin levels were higher in overweight women, and were only elevated in GDM women who were overweight. Area under the curve of glucose was not affected by body weight, but area under the curve of insulin was significantly higher in overweight than normal-weight women (Table 1). HOMA-IR, an index of insulin resistance, was higher in GDM than non-gdm but the increase was minimal in the normal-weight GDM group, but was much higher in the overweight group. By contrast, HOMA-B, an index of insulin secretion capacity, was decreased in all GDM women regardless of body weight, was more important in normalweight GDM women considering the minimal changes in insulin resistance. Thus, GDM development in normal-weight women was mostly related to insulin secretory dysfunction. Although blood pressure remained in a normal range for all groups, it was higher in overweight and GDM women (Supplementary Table 2). Plasma total cholesterol levels were not changed by body weight and GDM states while plasma high-density lipoprotein concentrations were lower in overweight women than normal-weight women, but GDM women had higher concentrations of high-density lipoprotein than non-gdm women (Supplementary 197 & 2013 Macmillan Publishers Limited European Journal of Clinical Nutrition (2013)

3 198 Table 1. Overnight-fasted plasma glucose and insulin levels at 24 28th weeks of pregnancy according to prepregnancy BMI and GDM Normal weight (BMIo23) Overweight (BMIX23) Non-GDM (n ¼ 395) GDM (n ¼ 98) Non-GDM (n ¼ 136) GDM (n ¼ 117) Fasting glucose (mmol/l) 4.4±0.3 a 4.7±0.5 b 4.5±0.3 a,b 5.2±0.7 c,d,e Glucose at 1 h after 50 g glucose challenge (mmol/l) 6.2±0.8 b 9.1±1.0 c 6.3±0.8 b 9.4±1.5 c,d,e Fasting insulin (pmol/l) 80.6±38.9 a 77.1±25.7 a 88.9±40.3 b 109.0±43.1 c,d,e Area under the curve of glucose at 100 g OGTT 26.1± ±5.4 Area under the curve of insulin at 100 g OGTT 1368± ±3120 e HbA 1C (%) 5.3±0.3 b 5.8±0.4 c 5.4±0.4 b 5.8±0.5 c,d HOMA-IR 2.1±1.0 f 2.3±0.9 a 2.5±1.2 b 3.6±1.6 c,d,e HOMA-B 293.7±229.2 c 197.7±86.6 b 288.7±153.5 c 219.1±129.9 b,d Adiponectin (mg/l) 5.93 (6.04) 4.40 (6.13) 5.94 (4.96) 4.18 (3.19) d Adipsin (mg/l) 0.43 (0.33) 0.39 (0.33) 0.53 (0.24) 0.58 (0.26) e Leptin (mg/l) 5.77 (4.87) 6.19 (6.73) (7.94) (9.57) e Resistin (mg/l) 6.83 (6.46) 6.58 (5.06) (10.97) 5.44 (4.43) d Visfatin (mg/l) 1.38 (0.743) 0.86 (0.59) 1.17 (1.78) 0.63 (0.53) d,,e C-reactive protein (nmol/l) 97.1 (139.1) (121.0) (210.5) (298.1) d,,e Prolactin (nmol/l) 6.01±3.75 a 6.36±2.26 a 5.64±2.47 b 5.61±1.96 b,e Progesterone (nmol/l) 568.9±178.4 b 707.2±227.4 c 442.7±146.3 c 642.0±226.4 c,d,e Abbreviations: BMI, body mass index; GDM, gestational diabetes mellitus; HbA1 c, hemoglobin A1 c, HOMA-B, homeostatic model assessment for b-cell function; HOMA-IR, homeostatic model assessment for insulin resistance; OGTT, oral glucose tolerance test. Values are mean±s.d. or median (interquatile range). a,b,c,f Means in the same row with different superscripts were significantly different among the groups by Tukey s test at Po0.05. d Significantly different by GDM at Po0.05. e Significantly different by BMI at Po0.05. Table 2. Nutrient intake at weeks of pregnancy according to prepregnancy BMI and GDM Normal weight (BMIo23) Overweight (BMIX23) Non-GDM (n ¼ 395) GDM (n ¼ 98) Non-GDM (n ¼ 136) GDM (n ¼ 117) Energy intake (kcal/day) ±410.8 a ±543.5 b ±371.9 c ±595.3 b,d Energy (% EER) 84.1±17.1 a 91.4±22.2 b 76.3±15.2 c 85.8±22.7 a,d,e Carbohydrate (g/day) 285.4±59.2 a 317.1±90.9 b 271.3±50.6 a 308.8±85.2 b,d Fiber (g/day) 21.83±5.9 a 26.28±9.2 b 20.4±5.4 a 24.35±8.9 b,d Protein (g/day) 80.4±20.7 a 87.5±23.9 b 78.5±22.9 a 90.8±33.2 b,d Protein (% RI) 115.5±12.8 a 125.2±23.7 b 112.2±22.6 a 129.4±31.9 b,d Plant protein (g/day) 35.42±9.0 c 40.08±12.6 a 36.85±18.2 c 42.97±19.6 b,d,e Animal protein (g/day) 45.42± ± ± ±25.5 Fat (g/day) 64.5±21.7 a 68.7±29.7 b 60.2±19.5 c 67.4±26.2 b,d Fat (% of energy) 28.5±5.8 b 28.1±7.1 b 27.3±5.1 a 26.9±7.1 a,e Taurine (mg/day) 21.2±5.2 b 7.5±2.6 c 14.1±3.1 a 16.6±5.4 a,d Abbreviations: BMI, body mass index; EER, estimated energy requirement; GDM, gestational diabetes mellitus; RI, recommended intake. Values are mean±s.d. a,b,c Means in the same row with different superscripts were significantly different among the groups by Tukey s test at Po0.05. d Significantly different by GDM at Po0.05. e Significantly different by BMI at Po0.05. Table 1). Plasma triglyceride levels were higher in overweight and GDM women than normal-weight and non-gdm women. Plasma adipokines and gestational hormones at 24- to 28-week pregnancy GDM, but not body weight, was associated with lower plasma adiponectin concentrations and overweight non-gdm women had greatly elevated resistin levels (Table 1). Plasma adipsin and leptin concentrations were higher in overweight than normalweight women but were not affected by GDM. Plasma visfatin concentrations were lower in GDM and overweight women while plasma C-reactive protein concentrations increased in women with GDM and overweight. Plasma prolactin concentrations were lower in overweight women regardless of GDM while plasma progesterone concentrations were increased in GDM women but decreased by overweight (Table 1). Nutrient intakes at 24- to 28-week pregnancy GDM women had higher energy, carbohydrate, fiber and total fat intakes than non-gdm in both normal-weight and overweight subjects although their intakes, except for total fat, were within the recommended ranges (Table 2). When total fat intake was categorized by different fatty acids, the consumption of saturated and monounsaturated fatty acids was higher in subjects with GDM than non-gdm subjects regardless of BMI (Table 2). However, the intakes of polyunsaturated fatty acids and cholesterol were not significantly different among the groups (Supplementary Table 3). As GDM women had higher intakes of energy than non-gdm, they consumed higher amounts of most nutrients. However, normal-weight GDM women had lower intakes of taurine, which is related to b-cell survival and insulin secretion (Table 2). Parameters that influenced GDM in normal-weight and overweight pregnant women In normal-weight women, family history, plasma progesterone and triglyceride levels, and saturated fatty acid intakes exhibited positive associations with GDM whereas HOMA-B and plasma visfatin levels were negatively associated (Table 3). These parameters explained 33.8% of the variation in GDM development in normal-weight pregnant women. However, in overweight pregnant women, diastolic blood pressure, HOMA-IR, plasma progesterone and triglyceride levels and caloric intake were European Journal of Clinical Nutrition (2013) & 2013 Macmillan Publishers Limited

4 Table 3. Parameters explaining the variation of GDM development in the normal and overweight women by backward stepwise logistic regression analysis 95% Confidence interval Odds ratio P-value 1, GDM normal weight R 2 ¼ Family history HOMA-B Plasma progesterone (nmol/l) Plasma triglyceride (mmol/l) Plasma visfatin (mg/l) Saturated fatty acid intake (% RI) 1, GDM overweight R 2 ¼ DBP (mm Hg) HOMA-IR Plasma progesterone o (nmol/l) Plasma triglyceride (mmol/l) Plasma adiponectin (mg/l) Plasma resistin (mg/l) Plasma CRP (nmol/l) Calorie intake (% EER) Abbreviations: CRP, C-reactive protein; DBP, diastolic blood pressure; EER, estimated energy requirement; GDM, gestational diabetes mellitus; HOMA-B, homeostatic model assessment for b-cell function; HOMA-IR, homeostatic model assessment for insulin resistance; RI, recommended intake. positively associated with GDM while plasma adiponectin and resistin levels were negatively associated with GDM (Table 3). These parameters explained 59.1% of the variation in overweight pregnant women. Therefore, insulin secretion capacity was a more important factor for inducing GDM in normal-weight pregnant women whereas insulin resistance was for overweight pregnant women. Plasma triglyceride and progesterone levels were common factors associated with the development of GDM. Pregnancy outcomes After diagnosis of GDM, dietary consumption was managed resulting in blood glucose levels that were well maintained. This is not uncommon as a diagnosis of GDM motivates both the patients and health-care providers to manage the GDM. 15 As a result, gestational age at delivery, delivery methods and incidence of a low Apgar score (an indicator of neonatal morbidity) were not affected by prepregnancy BMI and GDM (Table 4), indicating that prepregnancy BMI and GDM does not affect neonatal morbidity in well-controlled GDM. In addition, macrosomia and small for gestational age did not differ according to GDM and prepregnancy BMI. However, LGA frequency was significantly higher only in normal-weight GDM women compared with normal-weight non-gdm women. The birth weight of babies and the preterm incidence were higher in overweight groups at prepregnancy regardless of GDM (Table 4). Thus, these results suggested that GDM and prepregnancy overweight resulted in minimal adverse pregnancy outcomes when blood glucose levels and weight gain during the last trimester pregnancy are well controlled. Parameters that influenced LGA in normal-weight and overweight pregnant GDM women In normal-weight GDM women, Na intake was significantly positively associated with LGA whereas taurine intake was negatively associated with LGA (Table 5). However, in overweight GDM women, diastolic blood pressure and caloric intake were positively associated with LGA whereas plasma visfatin levels and potassium and Zn intake were negatively associated with LGA (Table 5). These parameters explained 29.5% of the variation in LGA in overweight GDM. DISCUSSION Like type 2 diabetes, GDM is most commonly associated with both insulin resistance and inadequate compensatory insulin secretion. However, unlike type 2 diabetes, GDM only occurs during pregnancy and spontaneously resolves after childbirth. During normal pregnancy, progesterone induces insulin resistance that does not result in diabetes because of increased compensatory insulin secretion because of prolactin-induced b-cell proliferation. 16,17 However, in some pregnancies GDM develops as a consequence of either unusually high insulin resistance, perhaps because of the contribution of pre-existing insulin resistance in overweight women, or because of inadequate b-cell expansion and concomitant insulin insufficiency. In this study, HOMA-IR and HOMA-B analyses revealed that in overweight women GDM was mostly due to insulin resistance, which is similar to most GDM in Western countries; however, the 50% of GDM women who were lean were no more insulin resistant than overweight non-gdm women and low insulin secretory capacity was the primary defect causing the GDM, although prolactin secretion was highest in that group. Therefore, GDM in normal-weight women was not due to either excessive insulin resistance or to low prolactin secretion, but due either to inherent low insulin secretory capacity or to other endogenous or exogenous factors that impaired insulin secretion. A prominent factor that led to GDM in both weight groups was a higher energy consumption that was accompanied by higher carbohydrate consumption. It is possible that the higher carbohydrate consumption led to diabetes because of a carbohydrate load that neither group of GDM women could accommodate and that simply lowering their dietary carbohydrate content would have prevented much of the GDM. On the other hand, there were some protective factors against developing GDM. In overweight women, elevated resistin levels and higher taurine intakes appeared to provide significant protection against GDM, and resistin appeared to protect against infants being born LGA. Higher levels of both adiponectin and visfatin were protective against GDM regardless of body weight. Visfatin is the same molecule as nicotinamide phosphoribosyltransferase, the rate-limiting enzyme in nicotinamide adenine dinucleotide biosynthesis. 7 Nicotinamide adenine dinucleotide has a regulatory role in modulating the activity of the nicotinamide adenine dinucleotide-dependent deacetylase, sirtuin-1, which improves insulin sensitivity in liver, adipose tissue and skeletal muscle entirely or in part by upregulating the biosynthesis of adiponectin Furthermore, sirtuin-1 also potentiates glucose-stimulated insulin secretion by pancreatic b-cells 21 and low visfatin/nicotinamide phosphoribosyltransferase impairs glucose-stimulated insulin secretion. 22 Low visfatin levels would be expected to increase insulin resistance and decrease insulin secretory capacity; and would thus exacerbate impaired whole body glucose regulation in both lean and overweight Korean GDM women. The other important observations of this study were the effects of nutrient intakes and dietary habits. This study revealed that both normal-weight and overweight GDM women had higher energy and saturated fat intakes than non-gdm women although energy intakes of all groups were lower than the recommended energy requirementsforpregnantwomen,buttheyhadproperweightgain until the 24 28th week of the pregnancy. Fat intake was 27 28% of energy intake in all groups, but the GDM groups consumed more saturated fat than non-gdm. Women who had a habit of consuming 199 & 2013 Macmillan Publishers Limited European Journal of Clinical Nutrition (2013)

5 200 Table 4. Pregnancy outcomes according to prepregnancy BMI and GDM Normal weight (BMIo23) Overweight (BMIZ23) Non-GDM (n ¼ 395) GDM (n ¼ 98) Non-GDM (n ¼ 136) GDM (n ¼ 117) Gestational age at delivery (weeks) 39.1± ± ± ±1.89 Delivery methods (%) Normal labor Cesarean section Preterm (%) c Baby sex (male %) d Baby weight (g) 3186±475 a 3138±471 a 3300±458 b 3281±537 b,d Large for gestational age (%) d Small for gestational age (%) Macrosomia (44000 g, %) Abbreviations: BMI, body mass index; GDM, gestational diabetes mellitus. Values are mean±s.d. or percentage. a,b Means in the same row with different superscripts were significantly different among the groups by Tukey s test at Po0.05. c Significantly different by GDM at Po0.05. d Significantly different by BMI at Po0.05. Table 5. Parameters explaining the variation of large for gestational age in the normal and overweight GDM women by backward stepwise regression analysis Estimate of coefficient 95% Confidence interval Normal weight R 2 ¼ Sodium intake (g/day) to Vitamin C intake (mg/day) to Taurine intake (mg/day) to Overweight R 2 ¼ DBP (mm Hg) to Plasma visfatin (mg/l) to Plasma leptin (mg/l) to Calorie intake (% EER) to Potassium intake (mg/day) to Zinc intake (mg/day) to Abbreviations: DBP, diastolic blood pressure; EER, estimated energy requirement; GDM, gestational diabetes mellitus. less fatty foods and snacks and more milk products were at lower risk of developing GDM (data not shown). Taurine consumption was much lower in normal-weight GDM women than non-gdm. Therefore, high energy intakes with higher saturated fats and lower taurine intake may contribute to the development of GDM. Owing to its antioxidant and anti-inflammatory action, taurine protects against apoptosis of b-cells and enhances their regeneration. 23 Thus, a low intake of taurine might block b-cell expansion during pregnancy contributing to GDM and might be related to LGA development because the increase in b-cell mass is necessary to maintain glucose homeostasis. This study had some limitations. First, as a cross-sectional study, it did not determine cause-and-effect. Second, the division of BMI may not be reflective of weight status at all-time points since prepregnancy BMI was used for dividing the two groups such as normal-weight and overweight. Third, the biochemical levels at the beginning of pregnancy were not determined to compare with later trimesters. Finally, the nutrient intakes determined by 24-h recall might introduce some bias in determining usual intake since 24-h recall is based on the memory, and may not be fully recalled especially by obese women, and there are differences in intakes during weekdays and weekends. Many studies have revealed that the 3-day food record is considered as the gold standard for determining usual intakes. 24 However, considering expenses and accuracy of the measurement, 24-h recall can be appropriate methods to determine actual and usual intake. In addition, our preliminary study showed that 24-h recall was close to 3-day record in their actual intakes. In conclusion, this study clearly demonstrated that GDM development in normal-weight Korean women is primarily related to poor insulin secretory capacity, whereas in overweight Korean GDM women it is associated with both insulin resistance and to inadequate insulin secretion. The data also suggest that current recommended energy intakes during pregnancy needs to be re-evaluated. This study also found that in addition to traditional risk factors such as BMI, age, parity and family history of diabetes; HOMA-B assessment of insulin secretory capacity might be a valuable tool for early assessment of GDM risk. Other useful predictors of GDM may be high triglyceride, low visfatin and adiponectin, and high progesterone levels in the circulation. Dietary interventions for preventing the onset of GDM and improving pregnancy outcome should include careful management of energy intake, especially saturated fat intake, and possibly increasing the intake of taurine-rich foods. CONFLICT OF INTEREST The authors declare no conflict of interest. ACKNOWLEDGEMENTS This work was supported by grants from the Korean Research Foundation in Korea (R ) and the Korea Health 21 R&D Project, Ministry of Health and Welfare, Republic of Korea (Grant no. A ). AUTHOR CONTRIBUTIONS The author s responsibilities were as follows: SP contributed to designing the study, analyzing data and preparing the manuscript; M-YK, J-TW and J-HY analyzed data and reviewed the manuscript; YJK performed data analysis; JWD and Y-MP joined to prepare the manuscript; SHB and S-HK collected samples and data and prepared the manuscript. REFERENCES 1 Hedderson MM, Williams MA, Holt VL, Weiss NS, Ferrara A. Body mass index and weight gain prior to pregnancy and risk of gestational diabetes mellitus. Am J Obstet Gynecol 2008; 198: 409 e1-e7. 2 Coustan DR, Nelson C, Carpenter MW, Carr SR, Rotondo L, Widness JA. Maternal age and screening for gestational diabetes: a population based study. Obstet Gynecol 1989; 73: Park S, Park JE, Daily JW, Kim SH. Low gestational weight gain improves infant and maternal pregnancy outcomes in overweight and obese Korean women with gestational diabetes. Gynecol Endocrinol 2011; 27: European Journal of Clinical Nutrition (2013) & 2013 Macmillan Publishers Limited

6 4 Chan JC, Malik V, Jia W, Kadowaki T, Yajnik CS, Yoon KH et al. Diabetes in Asia: epidemiology, risk factors, and pathophysiology. JAMA 2009; 301: Volpe L, Di Cianni G, Lencioni C, Cuccuru I, Benzi L, Del Prato S. Gestational diabetes, inflammation, and late vascular disease. J Endocriol Invest 2007; 30: Donath MY, Böni-Schnetzler M, Ellingsgaard H, Ehses JA. Islet inflammation impairs the pancreatic beta-cell in type 2 diabetes. Physiology 2009; 24: Saddi-Rosa P, Oliveira DSV, Giuffrida FMA, Reis AF. Visfatin, glucose metabolism and vascular disease: a review of the evidence. Diabetol Metab Syndr 2010; 2: Hirose H, Yamamoto Y, Seino-Yoshihara Y, Kawabe H, Saito I. Serum high-molecularweight Adiponectin as a marker for the evaluation and care of subjects with metabolic syndrome and related disorders. J Atheroscler Thromb 2010; 17: Carpenter MW, Coustan DR. Criteria for screening tests for gestational diabetes. Am J Obstet Gynecol 1982; 144: Brănis teanu DD, Mathieu C. Progesterone in gestational diabetes mellitus: guilty or not guilty? Trends Endocrinol Metab 2003; 14: Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985; 28: Committee of Dietary reference intake in Korean Nutrition Society. Dietary reference intake for Koreans. Korean Nutrition Society: Seoul, WHO Expert Consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet 2004; 363: Lee S, Choi S, Kim HJ, Chung YS, Lee KW, Lee HC et al. Cutoff values of surrogate measures of insulin resistance for metabolic syndrome in Korean non-diabetic adults. J Korean Med Sci 2006; 21: Salmenhaara M, Uusitalo L, Uusitalo U, Kronberg-kippilä C, Sinkko H, Ahonen S et al. Diet and weight gain characteristics of pregnant women with gestational diabetes. Eur J Clin Nutr 2010; 64: Arumugam R, Horowitz E, Lu J, Collier J, Ronnebaum S, Fleenor D et al. The interplay of prolactin and the glucocorticoids in the regulation of b-cell gene expression, fatty acid oxidation, and glucose-stimulated insulin secretion: implications for carbohydrate metabolism in pregnancy. Endocrinology 2008; 149: Park S, Kim DS, Daily JW, Kim SH. Serum prolactin concentrations determine whether they improve or impair b-cell function and insulin sensitivity in diabetic rats. Diabetes Metab Res Rev 2011; 27: de Kreutzenberg SV, Ceolotto G, Papparella I, Bortoluzzi A, Semplicini A, Dalla MC et al. Downregulation of the longevity-associated protein sirtuin 1 in insulin resistance and metabolic syndrome: potential biochemical mechanisms. Diabetes 2010; 59: Milne JC, Lambert PD, Schenk S, Carney DP, Smith JJ, Gagne DJ et al. Small molecule activators of SIRT1 as therapeutics for the treatment of type 2 diabetes. Nature 2007; 450: Qiao L, Shao J. SIRT1 regulates adiponectin gene expression through Foxo1-c/ enhancer-binding protein a transcriptional complex. J Biol Chem 2006; 281: Ramachandran D, Roy U, Garg S, Ghosh S, Pathak S, Kolthur-Seetharam U. Sirt1 and mir-9 expression is regulated during glucose-stimulated insulin secretion in pancreatic b-islets. FEBS J 2011; 278: Revollo JR, Körner A, Mills KF, Satoh A, Wang T, Garten A et al. Nampt/PBEF/ Visfatin regulates insulin secretion in beta cells as a systemic NAD biosynthetic enzyme. Cell Metab 2007; 6: Song Y, Manson JE, Tinker L, Howard BV, Kuller LH, Nathan L et al. Insulin sensitivity and insulin secretion determined by homeostasis model assessment and risk of diabetes in a multiethnic cohort of women: the Women s Health Initiative Observational Study. Diabetes Care 2007; 30: Thompson FE, Subar AF. Dietary assessment methodology. In: Coulston AM, Boushey CJ (eds). Nutrition in the Prevention and Treatment of Disease. 2nd edn. Academic Press: San Diego, CA, USA, Supplementary Information accompanies the paper on European Journal of Clinical Nutrition website ( & 2013 Macmillan Publishers Limited European Journal of Clinical Nutrition (2013)