Adrenal Function in Adolescence is Related to Intrauterine and Postnatal Growth

Background and objectives: Intrauterine growth restriction is thought to be implicated in long-term programming of hypothalamic–pituitary–adrenal axis activity. We investigated adrenal function in adolescents born small for gestational age (SGA) in relation to their postnatal growth and cardiovascular parameters. Materials and Methods: Anthropometric parameters, blood pressure, heart rate, dehydroepiandrosterone sulfate (DHEAS), and cortisol levels were assessed in 102 adolescents aged 11–14 years followed from birth (47 SGA and 55 born appropriate for gestational age (AGA)). Results: Mean DHEAS levels were higher in SGA adolescents with catch-up growth (SGACU+), compared with AGA. Second-year height velocity and body mass index (BMI) gain during preschool years were positively related to DHEAS levels. Morning cortisol levels and systolic and diastolic blood pressure were higher in SGA adolescents without catch-up growth (SGACU−) compared with AGA. Second-year BMI gain was inversely, and 2–12 years increase in subscapular skinfold thickness was directly associated with cortisol levels. Size at birth and postnatal growth explained 47.8% and 38.2% of variation in DHEAS and cortisol levels, respectively. Conclusion: Adrenal function in adolescence is affected by prenatal and postnatal growth: small size at birth with postnatal catch-up growth is related to higher DHEAS secretion, whereas increased cortisol levels and blood pressure are higher in short SGA adolescents.


Introduction
Growth in utero may have implications for growth and development throughout life. Children born small for gestational age (SGA) have been shown to be at higher risk of metabolic consequences later in life, such as obesity, insulin resistance, dyslipidemia, and hypertension [1,2]. It has been suggested that these effects may arise owing to fetal programming of the hypothalamic-pituitary-adrenal (HPA) axis [1]. Stressful conditions in utero may lead to small birth size, alterations in regulation of the HPA axis and increased cortisol levels in both mothers and infants [3,4]. Cortisol plays a central role in the regulation of human metabolism [5], and disturbed secretion of cortisol may in turn be related to increased risk of metabolic syndrome [2,6,7]. In addition, activation of the HPA axis may be demonstrated by exaggerated adrenarche and increased dehydroepiandrosterone sulfate (DHEAS) secretion [1]. Exaggerated adrenarche and small size at birth are known to be related to polycystic ovary syndrome in girls, conveying an additional increase in the risk of metabolic syndrome [8,9].
Evidence suggests that postnatal growth pattern may also influence later metabolic and hormonal profile in children born SGA [1]. Actually, approximately 90% of infants born SGA catch up rapidly with their peers in terms of both weight and length during the first few months after birth. Thus, at 2 years of age, only 10% of children born SGA remain short; these are mainly the infants who were short at birth [10,11]. It has been postulated that the absence of catch-up growth early in infancy can be linked to endocrine disturbances [12]. During adolescence, the human body undergoes dramatic changes: increases in growth hormone (GH) and sex steroid secretion lead to increased insulin resistance and accelerated somatic growth, as well as to changes in body composition. All these processes may unmask underlying disturbances in metabolic function and hormonal profile. Although previous studies of adrenal function have looked at hormone changes in children born SGA, most have focused on children who remain short relative to their peers [13][14][15][16][17][18][19] or on SGA children undergoing investigations for precocious pubarche [20].
In this study, we aimed to investigate the relationship of adrenal function with postnatal growth and cardiovascular parameters in adolescents born SGA both with and without catch-up growth in comparison to their peers born appropriate for gestational age (AGA).

Statement of Ethics
The study was approved by the regional biomedical research ethics committee at the Lithuanian University of Health Sciences in Kaunas (Nr. BE-2-42, approved on 2011.06.14), and informed written consent was obtained from all parents prior to the study start.

Study Population
The study cohort included 102 Caucasian newborns (47 born SGA (24 boys/23 girls) and 55 born AGA [23 boys/32 girls]) born in Kaunas between 1998 and 2000 in whom growth was followed prospectively from birth at our center [21]. All children in the present study were born between 32 and 42 weeks of gestation. In children born SGA, birth weight and/or length was below 2 standard deviation scores (SDS) of the mean according to sex and gestational age, relative to Swedish birth weight and length references based on data from 800,000 healthy newborns born between 1990 and 1999 at a gestational age of 24-43 weeks [22]. Weight and length at birth in AGA newborns was between -2 and +2 SDS [22]. Children were examined at 2, 5, 12, 18 and 24 months after birth, and once during later childhood (mean age, 6.3 ± 0.07 years) [22,23]. At the time of investigation, children were between 11 and 14 years of age (mean age, 12.5 ± 0.1 years), and mean pubertal stages were: pubic hair development, 2 (interquartile range, 2-3), gonadarche (stages according to Tanner for breast development in girls and genital stages in boys), 2 (interquartile range, 2-3), testicular volume in boys, 5 mL (interquartile range, 2-9 mL). 40% of girls (N = 22) were post-menarche. As 7 of the 47 children born SGA did not experience catch-up growth in height up to 6 years of age (the last investigation before adolescence), the SGA group was analyzed as a whole, as well as according to whether they experienced catch-up growth in height or not.

Study Design
Anthropometric measurements: Height was measured to the nearest 0.1 cm using a Harpenden wall-mounted stadiometer (Holtain, Ltd., Crosswell, UK) and weight was measured to the nearest 0.1 kg using Soehnle electric scales (Soehnle, Backnang, Germany). Height and weight measurements were converted to SDS according to an algorithm for children from birth to 2 years [22], combined with an algorithm for total height and weight up to 20 years of age [24]. Body mass index (BMI) was calculated as a ratio of weight (kg) to height (m) and converted to SDS [25]. Pubertal stages were defined according to criteria of Tanner for breast development in girls, genital stages in boys and pubic hair development in both sexes. Testicular volume was measured using orchidometer.
Definition of catch-up growth: As height in adolescence is also influenced by onset of puberty, children born SGA were assigned to groups based on height before puberty: SGA children with a height above -2 SDS of the mean of the reference population according to sex and current age at 6 years of age were considered to be with catch-up growth (SGA CU+ ) [22,24]; SGA children with a height below -2 SDS of the mean at 6 years of age were considered to be without catch-up growth (SGA CU− ). Height and weight growth during childhood are presented in Figure 1. defined according to criteria of Tanner for breast development in girls, genital stages in boys and pubic hair development in both sexes. Testicular volume was measured using orchidometer. Definition of catch-up growth: As height in adolescence is also influenced by onset of puberty, children born SGA were assigned to groups based on height before puberty: SGA children with a height above -2 SDS of the mean of the reference population according to sex and current age at 6 years of age were considered to be with catch-up growth (SGACU+) [22,24]; SGA children with a height below -2 SDS of the mean at 6 years of age were considered to be without catch-up growth (SGACU−). Height and weight growth during childhood are presented in Figure 1. Skinfold thickness was measured to the nearest 0.1 cm at the subscapular, upper arm (triceps) and thigh area on the left-hand side using the Harpenden skinfold caliper. Two measurements were made at each site and the mean value used for analysis. Limb skinfold thickness was calculated as the sum of skinfold thicknesses in the upper arm and thigh area.
Cardiovascular parameters: Systolic and diastolic blood pressure and heart rate were measured using an automatic device with a cuff size appropriate for arm circumference. Two measurements were made of each variable and the mean value used for analysis.

Hormonal Measurements
Blood samples for hormone measurements were taken once between 08:00 and 09:00 h in the morning after overnight fasting, centrifuged immediately and stored at -20 °C until hormone concentrations were determined.
DHEAS concentration was determined using the commercially available DHEA-SO4 RIA reagents kit (Institute of Izotopes, Co., Ltd., Budapest, Hungary) which has a detection limit of 0.064 μmol/L, an intra-assay coefficient of variation (CV) of 4.6 and an inter-assay CV of 5.8%. Skinfold thickness was measured to the nearest 0.1 cm at the subscapular, upper arm (triceps) and thigh area on the left-hand side using the Harpenden skinfold caliper. Two measurements were made at each site and the mean value used for analysis. Limb skinfold thickness was calculated as the sum of skinfold thicknesses in the upper arm and thigh area.
Cardiovascular parameters: Systolic and diastolic blood pressure and heart rate were measured using an automatic device with a cuff size appropriate for arm circumference. Two measurements were made of each variable and the mean value used for analysis.

Hormonal Measurements
Blood samples for hormone measurements were taken once between 08:00 and 09:00 h in the morning after overnight fasting, centrifuged immediately and stored at −20 • C until hormone concentrations were determined. DHEAS concentration was determined using the commercially available DHEA-SO 4 RIA reagents kit (Institute of Izotopes, Co., Ltd., Budapest, Hungary) which has a detection limit of 0.064 µmol/L, an intra-assay coefficient of variation (CV) of 4.6 and an inter-assay CV of 5.8%.
Cortisol concentration was determined using the Cortisol RIA reagent kit (DIAsource ImmunoAssays SA, Louvain, Belgium) with a detection limit of 2.5 nmol/L, an intra-assay CV of 7.7% and an inter-assay CV of 15.1%.

Statistical Analyses
The distribution of quantitative variables was tested for normality using the Kolmogorov-Smirnov test. Skewed parameters were log-transformed before analyses to ensure Gaussian distribution. Data for normally distributed variables are presented as mean and standard error of the mean (SEM); rank variables are presented as median and interquartile range. Study characteristics were compared using independent-sample t-tests for continuous variables and the χ 2 test for binary and rank variables. Between-group comparisons for continuous variables were made using univariate general linear models with least square difference adjustment; all values were adjusted for current age, pubertal stage, BMI SDS and sex; values for blood pressure analyses were additionally adjusted for current height. Partial correlation tests and Pearson correlation coefficients were used to analyze the relationship between DHEAS or cortisol levels and size at birth, early growth (in every measured interval) and cardiovascular parameters (blood pressure and heart rate).
Hierarchical multiple regression models using standardized coefficients were used to explain the variation in both DHEAS and cortisol levels, and to evaluate associations between DHEAS and cortisol levels during adolescence and perinatal and postnatal factors, after controlling for sex, age, pubertal stage, and BMI SDS . A p value of <0.05 was considered statistically significant. Statistical analyses were performed using SPSS statistical software (version 20.0, IBM ® , New York, NY, USA).

Study Population
At the current investigation, adolescents born SGA were significantly shorter and lighter than adolescents born AGA but had thicker subscapular skinfolds and higher waist to height ratio. There were no differences in current age and pubertal stages between total SGA and AGA groups. Study population characteristics are presented in Table 1. Birth weight SDS was lower in boys born SGA compared to girls born SGA (p = 0.040). In adolescence, girls born AGA had thicker limb and subscapular skinfolds than boys born AGA (p = 0.011 and p = 0.016, respectively). There was no other difference in anthropometric measurement between boys and girls born either SGA or AGA.
In total, 7 of the 47 children born SGA (14.9%; 4 girls, 3 boys) did not experience catch-up growth and (their height was below −2 SDS at 6 years of age); the remaining 40 children born SGA experienced spontaneous catch-up growth. At the time of the study, SGA CU− children were significantly shorter and lighter than SGA CU+ children and children born AGA (all p < 0.001). BMI SDS was also lower in SGA CU− children compared with SGA CU+ and AGA children (p = 0.003 and p=0.001, respectively); however, waist to height ratio was higher in SGA CU− adolescents compared with SGA CU+ and AGA adolescents (p = 0.003 and p < 0.001, respectively). There were no differences in limbs or subscapular skinfold thicknesses between SGA CU− and SGA CU+ adolescents (p = 0.456 and p = 0.719, respectively). At the time of study, SGA CU+ children were also shorter than children born AGA (p = 0.038), but there was no significant difference in weight SDS , BMI SDS and waist to height ratio between SGA CU+ and AGA adolescents.  Pubic hair development stage did not differ between SGA and AGA groups in boys or girls (p = 0.820 and p = 0.964, Figures 2 and 3, respectively).
Pubic hair development stage did not differ between SGA and AGA groups in boys or girls (p = 0.820 and p = 0.964, Figures 2 and 3, respectively).  As study children were in different pubertal stages, we compared pubertal maturation, expressed by testicular volume in boys and breast development stages in girls. When separated by sex, there was no difference in maturation between SGA and AGA children (all SGA vs. AGA, boys: p = 0.120; girls: p = 0.694). Also, there was no difference in testicular size between SGACU− and SGACU+ boys (p = 0.481, Figure 2). The stages of breast development in SGACU+ and AGA girls were similar (p = 0.179), with most girls being in the third stage for breast development according to Tanner ( Figure  3). However, 2 out of 4 SGACU− girls were prepubertal (in the first stage for breast and pubic hair development) (p = 0.013) compared with SGACU+ girls (0 out of 19); p = 0.024 compared with AGA girls (4 out of 31). Pubic hair development stage did not differ between SGA and AGA groups in boys or girls (p = 0.820 and p = 0.964, Figures 2 and 3, respectively).  As study children were in different pubertal stages, we compared pubertal maturation, expressed by testicular volume in boys and breast development stages in girls. When separated by sex, there was no difference in maturation between SGA and AGA children (all SGA vs. AGA, boys: p = 0.120; girls: p = 0.694). Also, there was no difference in testicular size between SGACU− and SGACU+ boys (p = 0.481, Figure 2). The stages of breast development in SGACU+ and AGA girls were similar (p = 0.179), with most girls being in the third stage for breast development according to Tanner ( Figure  3). However, 2 out of 4 SGACU− girls were prepubertal (in the first stage for breast and pubic hair development) (p = 0.013) compared with SGACU+ girls (0 out of 19); p = 0.024 compared with AGA girls (4 out of 31). As study children were in different pubertal stages, we compared pubertal maturation, expressed by testicular volume in boys and breast development stages in girls. When separated by sex, there was no difference in maturation between SGA and AGA children (all SGA vs. AGA, boys: p = 0.120; girls: p = 0.694). Also, there was no difference in testicular size between SGA CU− and SGA CU+ boys (p = 0.481, Figure 2). The stages of breast development in SGA CU+ and AGA girls were similar (p = 0.179), with most girls being in the third stage for breast development according to Tanner (Figure 3). However, 2 out of 4 SGA CU− girls were prepubertal (in the first stage for breast and pubic hair development) (p = 0.013) compared with SGA CU+ girls (0 out of 19); p = 0.024 compared with AGA girls (4 out of 31).

Relationship to Size at Birth and Presence/Absence of Catch-Up Growth
Overall, DHEAS levels were significantly higher in children born SGA than in children born AGA (Table 2). Separate analysis of SGA CU− and SGA CU+ groups showed that DHEAS levels were only significantly higher in SGA CU+ children compared with AGA children (Table 2). There was no significant difference between the two SGA groups (p = 0.963).

Relationship to Sex
When SGA and AGA children were separated into groups according to sex, there were no significant difference in DHEAS levels between boys born SGA compared with boys born AGA and no difference between the respective groups in girls (Table 2). Also, there was no difference in DHEAS levels between boys and girls born SGA or AGA (p = 0.557 and p = 0.412, respectively).

Relationship to Birth Characteristics
Overall, in the two SGA groups and the AGA group combined, DHEAS levels in adolescence were inversely associated with gestational age (r = −0.211, p = 0.034), birth weight (r = −0.306, p = 0.002), birth weight SDS (r = −0.345, p < 0.001), birth length (r = −0.283, p = 0.004), birth length SDS (r = −0.279, p = 0.005), BMI at birth (r = −0.296, p = 0.003) and ponderal index at birth (r = −0.229, p = 0.022) (analysis was adjusted for sex, current age, pubertal stage, and BMI SDS ). Analyzing separately by sex, only in boys birth weight, length and BMI were associated with DHEAS levels in adolescence (r = −0.315, p = 0.035; r = −0.325, p = 0.029 and r = −0.299, p = 0.046, respectively) (Supplementary Materials Table S1).  Table S1). A multiple regression model was constructed to assess the relationship between variations in DHEAS levels and variables that were significantly related to DHEAS in the univariate analyses, but without significant inter-correlations, controlling for sex, age, pubertal stage, and current BMI SDS . In this model, measures of postnatal growth, but not birth size, were significantly related to DHEAS levels in adolescence ( Table 3). The total variance explained by the model as a whole was 60.6%, p = 0.013. Postnatal growth explained 47.8% of the variance in DHEAS levels (p = 0.006).

Cortisol
Relationship to Size at Birth and Catch-Up Growth Overall, serum cortisol levels did not differ between the SGA and AGA groups. However, SGA CU− children had significantly higher serum cortisol levels than both AGA (Table 4) and SGA CU+ children (p = 0.023 and p = 0.015, respectively).

Relationship to Sex
Overall, there were no sex differences in cortisol levels between the SGA and AGA groups (Table 4). Also, no differences were found between boys vs. girls in SGA or AGA groups (p = 0.985 and p = 0.779, respectively).

Relationship to Birth Characteristics
For all children studied, cortisol levels during adolescence did not correlate with size at birth or gestational age (Supplementary Materials Table S2).

Relationship to Postnatal Growth
Cortisol levels in adolescence correlated inversely with BMI gain during the second year of life and positively with increase in subscapular skinfold thickness between 2 years of age and adolescence (r = −0.306, p = 0.041 and r = 0.477, p = 0.002, respectively). In the SGA group, cortisol levels were positively associated with length growth velocity during the first 2 months of life (r = 0.470, p = 0.020), but inversely associated with length growth velocity between 2 and 5 months of life and with current weight SDS (r = −0.511, p = 0.021 and r = −0.304, p = 0.048, respectively). Analyzing by sex, increase in subscapular skinfold thickness between 2 years of age was associated with cortisol levels only in adolescent girls (r = 0.584, p = 0.007) (Supplementary Materials Table S2).
Factors significantly related to cortisol levels at between 11 and 14 years of age and without significant inter-correlation were assessed in a multiple regression analysis. In this model, the most important factors related to cortisol level in adolescence were BMI gain during the second year of life and increase in subscapular skinfold thickness from 2 years to adolescence ( Table 5). The total model explained 40.3% of the variance in cortisol levels at the time of investigation, p = 0.024. Postnatal growth (length growth velocity during the first 2 months of life, BMI gain during the second year of life and increase in subscapular skinfold thickness between 2 years of age and adolescence) explained 38.2% of the variance in cortisol levels (p = 0.002). Regression coefficients in the final model are presented in Table 4.

Cardiovascular Parameters
Overall, there was no significant difference in systolic and diastolic blood pressure between the SGA and AGA groups. However, systolic blood pressure was higher in SGA CU− children compared with SGA CU+ and AGA children (SGA CU− , 125.1 ± 5.4 mmHg vs. SGA CU+ , 111.2 ± 1.9, p = 0.015; and vs. AGA, 109.0 ± 1.7 mmHg, p = 0.009). There was no difference in systolic blood pressure between SGA CU+ and AGA children (p = 0.426).
In the total cohort, systolic blood pressure was inversely related to weight gain between 1.5 and 6 years of life (r = −0.535, p = 0.003). Moreover, a positive relationship was established between cortisol levels and diastolic blood pressure (r = 0.203, p = 0.040).

Principle Findings
In this study, we were able to demonstrate that both size at birth and early postnatal growth are important predictors of adrenal function in adolescence.

DHEAS Association with Postnatal Growth
Previous studies have looked at DHEAS levels in preschool-and school-aged children who had been born SGA and most of them found higher levels in SGA children who had attained normal stature compared with those born AGA, but no difference between those who did not experience catch-up growth and AGA [13][14][15][16]. In combination with our data, these findings suggest that rapid postnatal growth-in particular during the second year of life-is an important predictor of adrenal androgen secretion in adolescence and it is more important than size at birth. Furthermore, together with the findings from Growth and Obesity Chilean Cohort Study in prepubertal children [26], our study suggests that not only length/height growth but also weight or BMI gain appears to be associated with adrenal function in later life.
A relationship between high BMI and hyper-responsiveness of the HPA axis has been shown in other recent studies [27]. In our study, however, a relationship between the increase in BMI rather than the actual BMI and DHEAS concentration in adolescence have been demonstrated: increase in BMI from birth to 6 years of life was found to be an independent predictor of the elevation of DHEAS concentration in adolescence, even after adjustment for current height, pubertal stage, and BMI SDS .
BMI can be seen as a surrogate marker of adiposity, although it does not reflect body adipose tissue compartmentalization. Indeed, the SGA-born children in our study were even leaner as adolescents than those born AGA, but had significantly higher waist to height ratios, indicating a more central distribution of adipose tissue. Furthermore, previous studies have found lean and total fat mass to be comparable in children born SGA or AGA [28,29], but with a tendency for increased visceral fat distribution in SGA children [28,[30][31][32][33][34][35][36]. Additionally, DHEAS concentration in our study was inversely related to the increase in limb skinfold thickness, which reflects growth of peripheral adipose tissue. Taken together, it may be suggested that the specific postnatal growth pattern in SGA children, with rapid linear growth during the first 2 years of life, a greater gain in BMI from birth to 6 years of age, and a more central adipose tissue distribution in adolescence are associated with exaggerated adrenal androgen secretion.

Secondary Sexual Characteristics and DHEAS Secretion
In the present study, the stages of pubic hair and gonadal development were similar in boys born SGA and those born AGA. Also, similarly to study by Beck Jensen [7], there were no differences in serum DHEAS levels between SGA and AGA boys. In previous studies higher DHEAS levels were described in girls with low birth weight and with a history of precocious pubarche [8]. In our study, there were no differences between the SGA and AGA groups in terms of DHEAS levels or stage of pubic hair and breast development. However, similarly to previous studies [37,38], girls born SGA were younger at onset of menarche than girls born AGA, suggesting that they experienced a faster transition through puberty. Interestingly, in regression analysis, current pubertal stage was not important predictor for DHEAS secretion, suggesting that early growth is a strongest predictor of variation of DHEAS secretion in adolescence.

Cortisol Secretion and Cardiovascular Function
In our study, adolescents who were born SGA and did not experience catch-up growth had higher cortisol levels compared with adolescents born SGA with subsequent catch-up growth and AGA. Moreover, in children born SGA as a group, cortisol levels were significantly inversely associated with current weight SDS . Studies in adults have shown that cortisol secretion has a U-shaped relationship with BMI: higher cortisol secretion was linked not only to obesity, but also to being underweight [39]. These data are in line with our findings: SGA adolescents without catch-up growth had the highest cortisol levels and lowest BMI SDS . Furthermore, cortisol levels in adolescence were directly associated with the increase in subscapular skinfold thickness from 2 years of age to adolescence. The increase in subscapular skinfold thickness may reflect the tendency for a more centralized distribution of adipose tissue, which in turn may be related to a worse metabolic profile later in life.
Cortisol levels in our study were not related to pubertal stage, confirming findings from the study by Knutsson et al. which found a significant inter-individual variability in endogenous secretion pattern estimated as mean diurnal cortisol levels in a large group of healthy children. However, there was little within-individual variability in diurnal cortisol rhythm when the same individuals were followed longitudinally throughout puberty [40].
In a recent study by Kouda et al., adolescents with relatively low body fat but with a more centralized fat distribution than their peers, tended to have higher levels of blood pressure in later life [41]. In line with these data, in our study, SGA children without catch-up growth were leaner, had a greater waist to height ratio and higher systolic and diastolic blood pressure compared with children born SGA with subsequent catch-up growth or AGA. Previous studies have highlighted the increased risk of metabolic syndrome in children born SGA, particularly in those who experience catch-up growth [30,35,36]. In contrast, our data suggest that the risk of metabolic syndrome in later life may be higher in children born SGA who did not experienced catch-up growth than in those with catch-up growth. However, owing to the small number of SGA adolescents without catch-up growth in our study, these associations require confirmation in studies with larger sample sizes.

Study Strengths and Limitations
The main strength of our study is the detailed longitudinal follow-up from birth to adolescence that was conducted within a single center, using reproducible and reliable measurement techniques. A further strength is the use of a single algorithm for height and weight from birth corrected for gestational age for all measurements.
Limitations of the study include the small sample size and different stages of pubertal development in study adolescents, although statistical analyses were adjusted for pubertal stages.
Assessment of adrenal function was also limited owing to the fact that samples were taken at a discrete time point, despite the well-known diurnal rhythm of adrenal hormones secretion; it is likely that assessment of the cortisol diurnal rhythm would be a more sensitive marker of adrenal function than a single measurement.

Conclusions
In conclusion, small size at birth is a predictor of higher DHEAS levels in adolescence in children born SGA who experienced catch-up growth. Children who were born SGA and did not experience catch-up growth are at risk of increased cortisol secretion and higher systolic and diastolic blood pressure in adolescence. Thus, the potential risk of metabolic consequences later in life in all individuals born SGA warrants further investigation.

Supplementary Materials:
The following are available online at http://www.mdpi.com/1010-660X/55/5/167/s1, Table S1. Relationship between dehydroepiandrosterone sulfate (DHEAS) levels and size at birth, and early growth in the total group of adolescents, adjusted for current age, pubertal stage, and BMI SDS ; Table S2. Relationship between cortisol levels and size at birth, and early growth in the total group of adolescents, adjusted for current age, pubertal stage, and BMI SDS .
Author Contributions: R.V. together with K.A.W. designed the study. The study was coordinated, and data were collected by R.V., M.V. and I.P. Statistical analysis was performed by I.P. I.P. drafted the manuscript, and it was finalized together with R.V. and K.A.W. All authors read and approved the final manuscript.
Funding: This study was funded by the Lithuanian Research Council (grant No. MIP-103-2011) and the Swedish federal governmental LUA/ALF grant to Sahlgrenska University Hospital.