Introduction
Low birth weight (LBW, birth weight < 2,500 g) represents 8.5% of live births and is associated with prematurity and/or restricted intrauterine growth. Reference Silveira, Victora and Horta1 Among maternal LBW-related conditions are nutritional disorders (malnutrition, obesity, and micronutrient deficiency) and diseases (arterial hypertension, diabetes mellitus, and infections). 2,Reference Black, Taylor and Arole3 A similar aspect of these situations is the change in the supply of nutrients or oxygen via the placenta, influencing growth and body composition at different stages of fetal development. Reference Papageorghiou, Kennedy and Salomon4
In the postnatal period, LBW infants can have accelerated weight gain and increased risk of developing diseases such as diabetes mellitus and cardiovascular diseases, especially those born small for gestational age (SGA). Reference Huang, Lin and Wang5,Reference Syddall, Sayer and Simmonds6 Recent studies relate this risk to changes in adipose tissue development during the intrauterine and postnatal phase and the secretion of adipokines and inflammatory factors that would be associated with the development of insulin resistance. Reference Kramer, Martin and Bogdanovich7,Reference Muhlhausler and Smith8 Among more than 100 substances were produced by adipocytes. Leptin and adiponectin are the most studied adipokines with relevance during the first years of life. Reference Kiess, Petzold and Töpfer9
Some relevant leptin’s actions are appetite and energy expenditure control by hypothalamic action. Leptin concentrations correlate with the amount of body fat and are high at the end of pregnancy. The newborn’s levels fall after birth. However, these values increase again as the infant gains weight and accumulates fat mass. Reference Kiess, Petzold and Töpfer9 There is a possibility to observe an increase in leptin values in situations of visceral fat accumulation and insulin resistance. Reference Muhlhausler and Smith8,Reference Kiess, Petzold and Töpfer9 In turn, adiponectin is produced by mature adipose tissue and subcutaneous topography. Adiponectin has anti-atherogenic, anti-diabetic, and anti-inflammatory action. Reference Kiess, Petzold and Töpfer9 Newborns have higher adiponectin concentrations than children and adults. Also, pregnant women with diabetes mellitus and obesity have lower adiponectin concentrations. Reference Kiess, Petzold and Töpfer9 Besides this, breast milk is related to decreased obesity in breastfed infants by several mechanisms, including the presence of leptin, lower protein content, favorable microbiota, and flavors diversity. Reference Logan, Siziba and Koenig10,Reference Victora, Bahl and Barros11
Therefore, as in the pregnant woman’s health and nutritional condition, weight gain and infant feeding interfere in a critical phase and may increase the risk for NCDs. Reference Syddall, Sayer and Simmonds6,Reference Muhlhausler and Smith8 We hypothesize that maternal health during pregnancy and infant feeding may influence the levels of adipokines in the first year of life. This study aimed to describe the concentrations of adipokines (leptin and adiponectin) in LBW infants, and the possible associations between leptin and adiponectin with maternal nutritional status, adequacy for gestational age, nutritional recovery, and current dietary intake in this group of infants.
Methods
Study design
This is a cross-sectional study conducted from May 2016 to November 2017, with LBW infants between 9 and 12 months of age, from the universe of children monitored at the Low Birth Weight Clinic (newborns weighing <2500 g and/or gestational age ≥ 34 weeks), after being born at the University Hospital of the Federal University of São Paulo, São Paulo, Brazil. Children with congenital malformations (heart disease, neural tube closure defects, and hydrocephalus), severe dysphagia, using probes or stomas, genetic syndromes, and food restrictions (food allergies, celiac disease) were excluded. The Research Ethics Committee of the Federal University of São Paulo approved the study (N° 0536/2016). All procedures were carried out after the parents/guardians signed the informed consent form.
During the study period, 86 LBW infants attended the outpatient clinic for the first visit. Of these, 65 met the inclusion criteria. Nine children did not participate in the study, and two children did not complete all stages of the assessment. Therefore, the final sample consisted of 54 children (Fig. 1). The non-included patients had characteristics similar to those evaluated in the study regarding gender, gestational age, birth weight, and pregnancy complications.
Fig. 1. Flow chart of identification and inclusion of infants in the study.
Data were obtained through information from the electronic medical record of previous visits, structured face-to-face interviews with guardians, anthropometric, dietary assessment, and laboratory tests.
Sociodemographic data and background
In the presential interview with the parents/guardians, data were collected on the socioeconomic status, parental education, obstetric history, health, previous maternal nutritional condition, and the child’s birth conditions. The method proposed by the World Health Organization 12 was used for classifying pre-gestational maternal nutritional condition. On this occasion, LBW infant’s dietary history information was also obtained, such as time of exclusive and total breastfeeding, use of infant formula and whole cow’s milk, time of introduction and composition of complementary feeding, and prophylactic iron and vitamin D supplementation.
Birth anthropometric data measured with <12 h of life were recorded from the child’s electronic medical record and used to describe the adequacy of newborns into small (SGA), adequate (AGA), and large for gestational age. Reference Villar, Ismail and Victora13
Anthropometry
Anthropometric measurements of weight, length, and head circumference were obtained as recommended by the World Health Organization. 12 Data were entered into the WHO Anthro 3.2.2 program (WHO, 2010), and the anthropometric indicators were presented in the form of z-score for age: Weight (WAZ), Length (LAZ), body mass index (BMIZ), head circumference (HCZ). The cut-off points adopted were those proposed by the WHO, 2006. Reference De Onis, Onyango and Borghi14 The variation in the values (Δ) of ΔWAZ, ΔLAZ, and ΔHCZ between birth and the moment of the evaluation was calculated to describe infants’ growth and nutritional recovery.
Dietary intake
Those responsible for the children were instructed to inform all the food and beverages consumed in the last 24 h (type of food, brand of processed products, method of preparation, and quantity consumed), based on three reminders for each child, one personally and the other two by telephone, on non-consecutive days, with at most 2 weeks between them, including a weekend day.
Nutritional intake was calculated with the support of the DietWin® program, Plus version, and the values obtained for energy (kcal/d) and proteins (g/kg/d) were compared to the nutritional recommendations proposed by the Dietary Reference Intake, prepared by the Institute of Medicine for the healthy population of Canada and the United States. 15
Also, a Food Frequency Questionnaire was applied, which identified the frequency (daily, weekly, monthly, sporadic, or not consuming) of consumption of ultra-processed foods in the diet according to the NOVA Reference Monteiro, Cannon and Moubarac16 classification.
Laboratory tests
On the day of the interview and anthropometric measures, 8 ml of blood was collected by peripheral venipuncture in the morning. The samples were aliquoted in dry tubes and centrifuged to separate the serum (4°C, 10 min, 3,500 rpm) to determine adipokine concentrations. The material was stored in Eppendorf® tubes and frozen in a −80°C freezer for further analysis. Leptin concentrations were performed by the Milliplex MAP Human Adipokine Magnetic Bead Panel 2-Endocrine Multiplex Assay Kit, developed by Luminex™ xMAP Technology (Millipore, Billerica, MA) with intra-assay <10% CV and inter-assay accuracy <15% CV and standard curve range (assay range) of 0.04 to 600 ng/ml. Adiponectin concentrations were performed by the ELISA method (Human Adiponectin Elisa Kit EZHADP-61K, Sigma Aldrich) with precision for intra-assay <10% CV, inter-assay <15% CV, and standard curve range (assay range) 1.5–100 ng/ml.
Statistical analysis
The data were entered and consolidated in an Excel spreadsheet (Office Microsoft®) and analyzed in the statistical package SPSS 25.0 (IBM®). Categorical variables were presented in absolute and percentage values, compared using the Chi-square test. Continuous variables were assessed for their distribution using the Kolmogorov–Smirnov test, kurtosis values, and distribution using histograms.
The variables with parametric distribution were presented in the form of mean and standard deviation, compared by the Student t test for independent and paired variables, in the variation of anthropometric indicators (WAZ, LAZ, and HCZ). Variables with non-parametric distribution (leptin) were presented as medians (interquartile range) and compared using the Mann–Whitney test.
Pearson’s (variables with parametric distribution) and Spearman’s (variables with non-parametric distribution) correlation coefficients were used to assess the association between leptin, adiponectin values, and leptin/adiponectin ratio with variables related to intake (energy and protein) and anthropometric indicators (WAZ, LAZ, HCZ, and BMIZ). The effect of WAZ variation on leptin levels was assessed on a logarithmic scale using Generalized Linear Models adjusted for pre-gestational BMI, with a 95% confidence interval. The level of significance adopted for all analyses was 5%.
Leptin serum levels described by Savino et al., 2014 Reference Savino, Rossi and Benetti17 in healthy infants between 6 and 12 months were 1.8 ± 1.9 ng/ml, and these values were used to calculate the power sample size. Adopting α-bidirectional = 0.05 and β = 0.2, the samples size of 54 infants has the power to detect a difference of 1.1 ng/ml in leptin serum levels.
Results
This study evaluated 54 LBW infants with a mean chronological and adjusted age, for preterm infants, of 10.0 ± 1.5 and 9.1 ± 1.6 months, respectively. We observed that 32 (59.3%) of the children were female, 36 (66.7%) were preterm, and 25 pregnancies (46.3%) were twin. The mean gestational age was 35.9 ± 1.7 weeks, birth weight was 2,222 ± 231 g, and 23 (42.6%) were born SGA (Table 1). None of the infants was conceived by in vitro fertilization.
Table 1. General characteristics of low birth weight infants
Among the main maternal characteristics (n = 42) were a mean age of 30.1 ± 7.0 years, most women (32, 76.2%) had between 8 and 11 years of schooling, and 28 (66.7%) had a household income of one to three minimum wages. Thirty-four women (81.0%) had complications during pregnancy, mainly urinary tract infection (39.5%), pregnancy-specific hypertensive disease (34.8%), and gestational diabetes (25.5%). Other maternal diseases were reported by 11 mothers (20.3%): chronic renal failure (n = 2), congenital heart disease (n = 2), cardiac arrhythmia (n = 1), systemic lupus erythematosus (n = 1), nasal papilloma (n = 1), central nervous system tumor (n = 1), hypothyroidism (n = 1), hepatitis C (n = 1), and sickle cell anemia (n = 1). Regarding the pre-gestational nutritional condition, 19 women (45.2%) were eutrophic, 8 (19.2%) were malnourished, and 6 (14.3%) and 9 (21.4%) were overweight and obese, respectively.
At the time of the assessment, all infants received complementary foods. Only four (7.4%) were breastfeeding, 24 (44.4%) received infant formula, and 26 (48.1%) received cow’s milk with or without human milk.
Almost all infants (98%) consumed energy above the recommendation, on average, 310.5 ± 92 kcal/d above the estimated need. Protein consumption exceeded the recommendation in all infants, with a mean protein intake of 5.1 ± 1.5 g/kg/d. Also, 87.6% of infants (n = 47) consumed ultra-processed foods.
Most of the infants, 44 (81.5%) were eutrophic, eight (14.8%) were overweight, two were undernutrition (3.7%), and four (7.4%) had short stature. The mean variation (Δ) of the z-score of anthropometric indicators (birth and at the time of this assessment) is shown in Table 2, with a statistically significant increase in Δ WAZ (0.60; 95% CI 0.24, 0.96, p = 0.001) and LAZ (95% CI 0.83, 0.83, 95% CI 0.51, 1.16, p < 0.001).
Table 2. Variation of the z-score (Δ) of anthropometric indicators (at birth and at the time of assessment) of low birth weight infants
AGA, adequate for gestational age; SGA, small for gestational age; WAZ, weight/age; HCZ, head circumference/age; LAZ, length/age score z.
*Level of significance of the paired t test.
As for adipokines, the median of leptin concentrations, the mean adiponectin levels, and the leptin/adiponectin ratio were 2.0 (1.0; 3.4) ng/ml; 28.4 ± 7.0 ng/ml, and 0.09 ± 0.08, respectively.
Table 3 shows the distribution of the main maternal characteristics and infants stratified by adequacy for gestational age. More important variations were observed in anthropometric indicators of SGA infants than AGA, especially for WAZ (1.54; 95% CI 1.17, 1.91; p < 0.001). Also among SGA, we observed a higher percentage of mothers with pre-gestational BMI compatible with undernutrition (7 (30.4%) versus 1 (3.2%); p = 0.03), lower frequency of prematurity (8 (35%) versus 28 (90%); p < 0.01), and lower birth weight (2,128 ± 222 g versus 2,292 ± 215 g; p < 0.01). However, no difference was observed between SGA and AGA concerning the dietary history and current dietary intake (Table 3).
Table 3. Maternal, perinatal, sociodemographic characteristics, dietary history, and dietary intake in low birth weight infants stratified by gestational age
AGA, adequate for gestational age; SGA, small for gestational age.
1 Significance level of the Chi-square test.
2 Significance level of the Student t test.
3 Significance level of the Mann–Whitney test.
It should also be noted that in SGA infants, leptin concentrations (3.0 (1.7; 3.0) ng/ml versus 1.6 (0.9; 2.6); p = 0.032)) and the leptin/adiponectin ratio (0.13 ± 0.08 versus 0.07 ± 0.07; p = 0.018) were higher and those of adiponectin (26.2 ± 7.3 µg/ml versus 30.1 ± 6.4 ng/ml; p = 0.050) tended to be lower when compared with AGA infants. There was no statistically significant difference in leptin, adiponectin, and leptin/adiponectin concentrations between premature and full-term infants.
Pre-gestational BMI was a modifier of the effect of the variation ΔWAZ on leptin levels (p = 0.027) (Fig. 2). The higher the maternal BMI value, the more positive the association between ΔWAZ variation and infants’ leptin levels.
Fig. 2. Association between WAZ variation (ΔWAZ) and leptin levels adjusted for pre-gestational maternal Body Mass Index (BMI).
None of the variables related to energy, protein intake, duration of breastfeeding, use of infant formulas, or whole cow’s milk was associated with leptin, adiponectin, and leptin/adiponectin ratio.
Discussion
In this study, adipokines were taken at 9–12 months and LBW SGA infants had higher leptin concentrations than AGA infants. Also, leptin concentrations correlated with the WAZ variation, mostly when the mothers were pre-gestational overweight. The feeding of these infants was characterized by excess energy and protein intake, frequent consumption of ultra-processed foods, and low breastfeeding rates. Literature data confirm that the dietary pattern in the first months of life influences weight gain and the risk of obesity in the long term. Reference Rolland-Cachera and Michaelsen18–Reference Stettler, Stallings and Troxel20 This scenario may justify that almost 15% of these children born with low birth weight are already overweight with <1 year of age.
The maternal nutritional status was relevant in this study in different aspects, such as maternal malnutrition was associated with a higher likelihood of SGA birth, and pre-gestational maternal overweight potentiating the effect of accelerating the infant’s postnatal WAZ on the leptin concentrations. These findings suggest that the nutritional condition of women of childbearing age, even before pregnancy, should be considered within the strategies for promoting appropriate intrauterine and postnatal growth. Reference Van de Beek, Hoek and Painter21–24 In this study, approximately half of the women and 15% of the infants were overweight, which points to the importance of preventing overweight in the pediatric age group from an adequate nutritional condition of women in the pre-gestational stage.
A meta-analysis with 37 cohort studies in developed countries showed that maternal obesity was more critical than excessive weight gain during pregnancy when evaluating overweight and obesity in children up to 18 years of age. Reference Voerman, Santos and Patro Golab25 Among the various factors that could justify this association, the study authors mention that maternal obesity in the early stages of pregnancy would disrupt the concentrations of adipokines and inflammatory factors during fetal development. These findings are corroborated by another prospective study that evaluated inflammatory markers and adipokines in the three trimesters in healthy pregnant women. The authors observed that overweight/obese women in early pregnancy had higher leptin and insulin levels and lower adiponectin levels. Reference Perichart-Perera, Muñoz-Manrique and Reyes-López26
Among the cited mechanisms involved in adverse outcomes in pregnancy are interactions in placental miRNA expression associated with preeclampsia and restricted intrauterine growth. Reference Zhao, Moley and Gronowski27 In a study that assessed the placental miRNA profile in pregnant women with a pre-gestational history of overweight/obesity or excessive gestational weight gain, four among eight placental miRNAs (miR-100, miR-1285, miR-296, and miR-487) that were differentially expressed in maternal obesity were associated with lower birth weight and increased weight gain for children at 1, 4, and 12 months of age. Reference Carreras-Badosa, Bonmatí and Ortega28 Also, a case-control study with SGA and AGA births showed that the expression of miR-122-5p in childhood was a great predictor of adiponectin levels in young SGA-born adults. Moreover, the presence of miR-122-5p was related to insulin resistance and the risk of developing diabetes. Reference Inzaghi, Kistner and Germani29
The amount and distribution of adipose tissue are among the significant determinants of leptin and adiponectin concentrations throughout life. In physiological situations, leptin has an anorexic function and increases energy expenditure at the hypothalamic level. However, situations with accumulated subcutaneous adipose tissue, such as obesity and hyperinsulinemia, can result in hyperleptinemia and the development of central and peripheral resistance to leptin. Reference Muhlhausler and Smith8,Reference Kiess, Petzold and Töpfer9 On the other hand, lower adiponectin values are observed in situations of obesity and type 2 diabetes mellitus. Reference Kiess, Petzold and Töpfer9
Newborns with low birth weight (preterm and SGA) have less body fat and lower leptin levels than those with adequate birth weight. Reference Muhlhausler and Smith8,Reference Steinbrekera, Colaizy and Vasilakos30 In this study, the median leptin concentration was 2.0 ng/ml. These values are in agreement with a study carried out with 317 healthy infants and younger than 18 months, but who were born AGA; the authors observed a median and 90th percentile of leptin levels between 6 and 12 months of age of 1.44 ng/ml and 5.18 ng/ml, respectively. Reference Savino, Rossi and Benetti17
Although a recent meta-analysis found no difference in leptin concentrations between SGA and AGA Reference Goto31 newborns, in this study, leptin levels and the leptin/adiponectin ratio in SGA infants was higher than in AGA infants, regardless of gestational age. This finding may suggest that the SGA group showed excessive nutritional recovery with an accumulation of fat, which would result in some level of resistance to leptin’s action, which can be considered an unfavorable metabolic profile and risk for NCDs. Reference Muhlhausler and Smith8
Also, our findings are related to recently published experimental studies. In these studies, mice that went through a combination of restricted intrauterine growth and accelerated postnatal growth showed changes in insulin sensibility throughout their lives. Reference Fernandez-Twinn, Hjort and Novakovic32,Reference Berends, Dearden and Tung33 Their results were enhanced by maternal obesity and an obesogenic postnatal diet, Reference Loche, Blackmore and Carpenter34 situations we faced during our study.
In another experimental study, the authors showed in rats that the combination of a high-calorie diet (chocolate and soft drinks) in mothers and their offspring could induce permanent changes in leptin signaling. The regulation of appetite was more sensitive to leptin’s effect than that of energy expenditure, suggesting different programming of sensitivity to leptin in the hypothalamic arcuate nucleus. Hypothalamic astrocytes contain leptin receptors, and it is postulated that astrogliosis (proliferation of astrocytes) associated with a fat-rich diet would impair leptin’s action in the control of energy expenditure in the central nervous system Reference Kjaergaard, Nilsson and Secher35,Reference Pan, Hsuchou and Xu36 in the long term.
This study’s strengths are the assessment of dietary intake, gestational history, and maternal nutritional condition. On the other hand, admitted limitations are the absence of body composition measures and the small number of children included.
This study described that LBW infants suffer early changes in leptin and adiponectin concentrations, influenced by maternal (pre-gestational BMI), intrauterine (adequacy for gestational age – SGA), and postnatal factors (excessive weight gain). This combination of factors may increase the risk of NCDs in this group of children.
Acknowledgements
Multidisciplinary outpatient team of Universidade Federal de São Paulo and the nurse Lusinete Lopes, for the care in collecting laboratory tests from children.
Financial support
Fundação de Amparo à Pesquisa do Estado de São Paulo [FAPESP: 2016/09428-1] and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior [CAPES].
Conflicts of interest
None.
Ethical standards
The study received approval from the Research Ethics Committee of Universidade Federal de São Paulo (number: 1.550.249). Children’s caregivers (mainly the children’s mothers) have signed an informed consent, after an information sheet was explained and delivered to them.
