The importance of maternal insulin resistance throughout pregnancy on neonatal adiposity

Abstract Background Although previous studies evaluated the association of maternal health parameters with neonatal adiposity, little is known regarding the complexity of the relationships among different maternal health parameters throughout pregnancy and its impact on neonatal adiposity. Objectives To evaluate the direct and indirect associations between maternal insulin resistance during pregnancy, in women with obesity, and neonatal adiposity. In addition, associations between maternal fasting glucose, triglycerides (TG), non‐esterified fatty acids (NEFA), and neonatal adiposity were also assessed. Methods This is a longitudinal, secondary analysis of the DALI study, an international project conducted in nine European countries with pregnant women with obesity. Maternal insulin resistance (HOMA‐IR), fasting glucose, TG, and NEFA were measured three times during pregnancy (<20, 24‐28, and 35‐37 weeks of gestation). Offspring neonatal adiposity was estimated by the sum of four skinfolds. Structural equation modelling was conducted to evaluate the direct and indirect relationships among the variables of interest. Results Data on 657 mother‐infant pairs (50.7% boys) were analysed. Neonatal boys exhibited lower mean sum of skinfolds compared to girls (20.3 mm, 95% CI 19.7, 21.0 vs 21.5 mm, 95% CI 20.8, 22.2). In boys, maternal HOMA‐IR at <20 weeks was directly associated with neonatal adiposity (β = 0.35 mm, 95% CI 0.01, 0.70). In girls, maternal HOMA‐IR at 24‐28 weeks was only indirectly associated with neonatal adiposity, which implies that this association was mediated via maternal HOMA‐IR, glucose, triglycerides, and NEFA during pregnancy (β = 0.26 mm, 95% CI 0.08, 0.44). Conclusions The timing of the role of maternal insulin resistance on neonatal adiposity depends on fetal sex. Although the association was time‐dependent, maternal insulin resistance was associated with neonatal adiposity in both sexes.


| BACKG ROU N D
Maternal obesity is associated with neonatal 1-4 and childhood adiposity, and childhood obesity 2,5 is not only associated with a higher chance of obesity, but also premature death and disability in adulthood. 5 While maternal insulin resistance and fasting glucose have a positive association with neonatal adiposity, 6-13 the role of maternal triglycerides and non-esterified fatty acids on neonatal adiposity remains controversial. [6][7][8]11 It is unclear whether such associations are causal, in the causal pathway or a consequence of the metabolic pathways involved.
Shapiro et al 6 reported that HOMA-IR, glucose, but not triglycerides and non-esterified fatty acids at 24-32 weeks of gestation mediated the association between maternal pre-pregnancy body mass index (BMI) and neonatal adiposity. Although this study 6 provided relevant insights regarding the possible underlying mechanisms in the relationship between maternal metabolic profile and neonatal adiposity, some aspects still require further evaluation.
First, it is crucial that the entire pregnancy period is taken into account. Fetal development depends on the maternal metabolic profile throughout the whole pregnancy period, [6][7][8][9][10][11][12][13] illustrated by the fact that obese women or women who develop gestational diabetes have larger fetuses already by 20 weeks of gestation. 14,15 Second, there is evidence that neonatal girls might be more insulin resistant than boys, 16,17 and mechanisms underlying the relationship between maternal metabolism and neonatal adiposity might differ depending on the neonatal sex. 18 The aim of this investigation was to evaluate the direct and indirect associations between insulin resistance during pregnancy and neonatal adiposity, as measured by sum of skinfolds. In addition, the direct and indirect associations between fasting glucose, triglycerides, non-esterified fatty acids, and neonatal adiposity were also assessed. This is a novel approach because it assesses the complexity of the relationships among different maternal health Abstract Background: Although previous studies evaluated the association of maternal health parameters with neonatal adiposity, little is known regarding the complexity of the relationships among different maternal health parameters throughout pregnancy and its impact on neonatal adiposity.
Objectives: To evaluate the direct and indirect associations between maternal insulin resistance during pregnancy, in women with obesity, and neonatal adiposity. In addition, associations between maternal fasting glucose, triglycerides (TG), non-esterified fatty acids (NEFA), and neonatal adiposity were also assessed.
Methods: This is a longitudinal, secondary analysis of the DALI study, an international project conducted in nine European countries with pregnant women with obesity. Maternal insulin resistance (HOMA-IR), fasting glucose, TG, and NEFA were measured three times during pregnancy (<20, 24-28, and 35-37 weeks of gestation).
Offspring neonatal adiposity was estimated by the sum of four skinfolds. Structural equation modelling was conducted to evaluate the direct and indirect relationships among the variables of interest. parameters throughout pregnancy and its impact on neonatal adiposity. Although the observational design does not allow to assert on causality, it is a valuable approach for generating hypothesis to be tested in clinical trials.

| Design and participants
This is a longitudinal, secondary analysis of the DALI study which was a multicentre parallel randomised trial conducted in nine European countries (Austria, Belgium, Denmark, Ireland, Italy, Netherlands, Poland, Spain, and United Kingdom) during 2012-2015. Pregnant women with a pre-pregnancy body mass index (BMI) of ≥29 kg/m 2 , before 20 weeks of gestation, with a singleton pregnancy, and aged ≥18 years were invited to participate. 19 Exclusions included diagnosis with early gestational diabetes mellitus (GDM), pre-existing diabetes, and chronic medical conditions. The study was prospectively registered as a randomised controlled trial (RCT) on 21 November 2011 (ISRCTN70595832).

| Randomisation, masking, and interventions
Using a computerised random number generator, women were randomised to one of the following groups, pre-stratified for site: (a) healthy eating; (b) physical activity; (c) healthy eating + physical activity; (d) healthy eating + physical activity + vitamin D; (e) healthy eating + physical activity + placebo; (f) vitamin D; (g) placebo; and (h) control. Staff involved with measurements, but not participants, were blinded to the intervention. Since methodology has been extensively described elsewhere, 19 only variables of interest will be detailed in this manuscript.

| Neonatal adiposity
Triceps, subscapular, supra-iliac, and quadriceps skinfolds were measured with a Harpenden skinfold calliper, and values summed. Each skinfold measurement was repeated once, and if a difference of more than 0.2 mm was registered, a third measurement was performed and the average of the three was taken. Time between birth and measurements was registered in hours-neonatal age at the measurement. After fasting for 10 hours, blood was collected in three different periods during pregnancy (<20, 24-28, and 35-37 weeks of gestation).

| Maternal measurements
The procedures were identical in all pregnancy periods. All the samples were centrifuged and separated aliquots (1000 or 250 μL) placed in microrack tubes and stored at −20 or −80°C in the central trial laboratory, prior to analysis, in Graz, Austria. The maternal plasma concentrations of fasting glucose, triglycerides, and non-esterified fatty acids were quantified. For insulin and leptin, commercially available enzyme-linked immunosorbent assays (ELISAs) were used. Insulin resistance was derived from homeostasis model assessment (HOMA-IR). 20

Study question
How is the synergic relationship of maternal insulin resistance, fasting glucose, triglycerides, and non-esterified fatty acids with neonatal adiposity in pregnant women with obesity?

What's already known
Maternal insulin resistance and fasting glucose have a positive association with neonatal adiposity, and the role of maternal triglycerides and non-esterified fatty acids on neonatal adiposity remains controversial.

What this study adds
The association between maternal insulin resistance and neonatal adiposity differed depending on the infant sex.
While adiposity in boys was associated directly with maternal insulin resistance in the beginning of pregnancy, adiposity in girls was associated with maternal insulin resistance at 24-28 weeks of gestation via a chain of relationships through fasting glucose, triglycerides, and free fatty acids.

| Statistical analyses
Prevalence, means, standard deviations (SD), medians, and interquartile ranges (IQR) were used to describe the sample. Descriptive mean differences are presented with respective 95% confidence intervals.
Structural equation modelling was performed to evaluate the direct and indirect associations among maternal HOMA index, glucose, triglycerides, non-esterified fatty acids, and neonatal sum of skinfolds. Direct association refers to the association between two variables, whereas indirect association refers to the association between two variables that is mediated via a third variable. All the standard errors in the analyses were adjusted for the cluster structure of the data (individuals nested within site of recruitment). Table 1 describes the role of each variable in the diagram in detail, including the variables of interest and the adjustments. We have more than

| Missing data
We used the maximum likelihood for missing values method, which does not exclude a participant in the analysis because of a missing value in one of the variables. Thus, we avoided selection bias in our analysis. 21 Therefore, all the 657 neonates and their mothers in the study were included in the main analysis.

| Ethics approval
We obtained the local ethics committee approvals in all the nine countries, in which the study was conducted. Moreover, written informed consent of all women was acquired.

| RE SULTS
In total, 657 neonates (50.7% boys) and their mothers with obesity were included in the study. On average, the mothers were 32

| Maternal insulin resistance in relation to neonatal adiposity
In boys, maternal HOMA-IR at <20 weeks of gestation was directly associated with neonatal sum of skinfolds (β = 0.35 mm, 95% CI 0.01, 0.70). No other associations of maternal HOMA-IR with neonatal sum of skinfolds in boys were found (see Table 3). In girls, maternal HOMA-IR was not directly associated with neonatal sum of skinfolds at any time during pregnancy. However, maternal HOMA-IR at 24-28 weeks of gestation was indirectly associated with neonatal sum of skinfolds (β = 0.26 mm, 95% CI 0.08, 0.44).  (Table 4).
In girls, maternal fasting glucose at <20 weeks was inversely and

| Principal findings
The primary aim of this investigation was to evaluate the potential pathophysiological pathways underlying the association of maternal insulin resistance with neonatal adiposity in pregnant women

| Strengths of the study
Strengths of our study are that we have measures of maternal metabolic parameters at three time points in pregnancy, enabling us to look at longitudinal associations. Data were collected in nine European countries, increasing the external validity of our findings. Importantly, the mechanistic path analysis adds substantial insights on how some of the metabolic parameters during pregnancy in women with obesity could affect neonatal adiposity, although we cannot assert on causality and our model should be tested in clinical trials.

| Limitations of the data
Our findings should be interpreted with caution, since women were exposed to different interventions that could influence our results. However, the analyses were adjusted for the intervention component. Another limitation is the lack of normal or overweight women in the study; thus, we can only generalise our findings to Caucasian obese women. However, the prevalence and the increasing rate of obesity in women in reproductive age are alarming and so the complication related to obesity during pregnancy. 22

| Interpretation
Regarding the pathophysiological pathways on the relationship between maternal metabolism and neonatal adiposity, Shapiro et al 7 observed that 21.0% of the association between maternal pre-pregnancy BMI and neonatal adiposity was mediated by HOMA-IR and fasting glucose levels, but not by triglycerides or non-esterified fatty acids. 7 This is partly in line with our results, although Shapiro et al 7 included pregnant women with and without obesity whereas we only included pregnant women with obesity.
We observed that fasting glucose mediated 81.0% of the association between maternal HOMA-IR and neonatal adiposity, but only in girls.  Note: Indirect association refers to the association between the exposure (maternal HOMA-IR at different periods during pregnancy) and the outcome (neonatal sum of skinfolds) which is mediated by all the variables in the pathway (see Table 1 for detailed description). Note: Indirect association refers to the association between the exposure and the outcome (neonatal sum of skinfolds) which is mediated by all the variables in the pathway (see Table 1 for detailed description). The empty cells on the indirect relationships between fasting glucose, triglycerides, non-esterified fatty acids at 35-37 wk of gestation and neonatal sum of skinfolds are due to the lack of possible indirect relationships to be estimated (there are no intermediate variable between these exposures and the outcome).

F I G U R E
The direct association between maternal glucose and neonatal adiposity is well established in the literature. 6,[8][9][10]  The importance of triglycerides and non-esterified fatty acids levels during pregnancy on neonatal adiposity is not well established. 6,8,11 When associations are found they are usually positive. 27 In our sample with pregnant women with obesity, maternal triglycerides were negatively associated with neonatal adiposity in boys at <20 weeks (indirectly) and at 24-28 weeks (directly). In girls, maternal non-esterified fatty acids at 35-37 weeks showed a negative association with neonatal adiposity. Pregnant women with obesity from DALI exposed to motivational sessions on healthy eating presented higher levels of NEFA, and higher levels of NEFA were also found in the cord blood of the offspring when compared to controls. 28 In addition, maternal diet has been shown to influence triglyceride and free fatty acid levels, even after a period of fasting. 29 Whether this influenced the association between lipids and neonatal adiposity is unknown. However, there seems to be a complex interaction between maternal glycaemia and lipids in their influence on fetal growth, since maternal lipids were only associated with fetal growth in women with hyperglycaemia but not in those with normal glycaemia. 30 Whether this is due to differences in insulin resistance and therefore maternal lipolysis needs further study.
Sex differences in the associations between maternal metabolism and neonatal adiposity have been reported previously, but are inconsistent. 16,17,31 Inconsistencies might be due to study population (ie women with obesity versus with normal weight, with normal glycaemia versus with hyperglycaemia) or due to differences in the timing in pregnancy. Theoretically, male and female fetuses seem to experience different growth strategies in utero. 32,33 It is not well established yet, but male embryos tend to have more rapid cell divisions compared to females, 34 generating a more rapid growth. 32,33 This higher growth rate might explain the higher response to changes in nutrient supply reported in male fetuses. 32,33 These different growth strategies might result in differences in nutritional needs, at different times in pregnancy and our findings support this concept. However, more systematic assessment of sex differences in the relationship of maternal metabolic parameters with neonatal adiposity is needed.
We evaluated the pathophysiological pathway from maternal insulin resistance, through maternal glucose, triglycerides, and non-esterified fatty acids to neonatal adiposity in pregnant women with obesity. It indicates again the relevance of the early pregnancy period for neonatal outcomes in boys and in the middle of pregnancy for girls. Although our model explained most of the pathophysiological pathways from insulin resistance to neonatal adiposity in girls, the pathways in boys were still uncertain. A next step would be to investigate the associations of maternal metabolism with cord blood parameters, to confirm the Pedersen-Freinkel hypothesis that maternal glucose and lipids increase fetal insulin levels and thereby increase fetal fat accretion. 35

| CON CLUS IONS
Maternal insulin resistance in women with obesity was associated with adiposity in neonates. More specifically, higher insulin resistance in the beginning of pregnancy was related to higher adiposity in boys, whereas this association in girls was observed with insulin resistance in the middle of the pregnancy period.
Obstetricians, midwives, and health professionals should be aware of the deleterious impact of maternal insulin resistance on the offspring health and support pregnant women with life style changes (nutrition and physical activity) which could lower insulin resistance during pregnancy and consecutively benefit neonatal body composition. 36

ACK N OWLED G EM ENTS
We would like to thank all participants for their time on participating in this study. In the United Kingdom, the DALI team acknowledge the support received from the NIHR Clinical Research Network: Eastern, especially the local diabetes clinical and research teams based in Cambridge.