Results of this study suggest that at LGA newborns have higher insulin levels and HOMA-IR than AGA and SGA newborns.
It has been suggested that the association between LGA and insulin resistance may reflect the influence of several factors like maternal diabetes and gestational age [21
]; in the second phase of the study, to control these potential confounders, the mother’s age, parity, and gestational age of mothers in the groups of newborns with SGA, AGA, and LGA, were matching criteria.
Our results agree with the report by Wolf et al. [15
] who found in the cord blood of 175 neonates with a birth-weight of 3376±588g, insulin levels of 7.4±4.3 μU/mL as compared with our results in 107 at term newborns with birth-weight and insulin levels of 3313±678g and 7.1±4.3 μU/mL. However, the increase of insulin levels, according birth-weight class, in the study by Wolf et al. (SGA<AGA<LGA) showed a different sequence as compared with our results (AGA<SGA<LGA); finding that could be related with differences in the birth-weight of the LGA newborns, which were heavier (3824±364g versus 4175±197g), and/or the birth-weight of SGA newborns (2516±352g versus 2083±356g) that were lighter in our study as compared with the report by Wolf et al. [15
]. Finally, in addition to leptin, also differences in glucose levels could explain differences in the sequence of insulin concentration; given that we did not measure leptin and Wolf et al. [15
] did not report glucose levels, the results are not comparable in this issue. However, because criteria to define SGA and LGA at birth were not the same, the comparison between our results and those from Wolf et al. should be taken cautiously.
Furthermore, in our study the sequence by weight class in the newborn at birth, for glucose and HOMA-IR was LGA<AGA<SGA and AGA<LGA<SGA, respectively, suggesting that LGA newborns at birth exhibited a decrease of insulin sensitivity, that requires an elevated compensatory insulin secretion from fetus to maintain euglycemia.
Recently, Catalano et al. [25
] conducted a comparative study that included 53 lean and 68 obese pregnant women to verify if fetuses of obese women have increased insulin resistance. Although we did not analyze our data based on the BMI of pregnant women, our results agree with the report by Catalano et al. On this regard, the birth-weight of newborns of lean and obese women (3217±452g and 3320±460g) in the study by Catalano et al. [25
] is similar to the birth-weight of AGA newborns in our study (3270±380g), as was the HOMA-IR (1.51±0.86 vs. 1.06±0.70 in the newborns of obese and lean mothers, in the study by Catalano) and 1.7±1.3 in the AGA newborns of our study. Furthermore, Catalano et. al. [25
] also reported that there was a positive correlation between fetal insulin resistance and fetal adiposity supporting our finding that suggest that elevated body-weight at birth might be the main risk factor involved in the development of hyperinsulinemia and elevated HOMA-IR in the newborns.
Interestingly, although SGA and AGA newborns exhibited similar HOMA-IR, blood cord glucose levels were higher in the SGA newborns, suggesting that could be an impaired insulin secretion and or insulin action related with the low birth-weight at birth. Additional factors related with the insulin production by fetus, such as elevated leptin and decreased adiponectin levels also could explain the findings of our study in regards SGA newborns. Further studies with the appropriate biochemical measurements are needed.
Recently, in the Hyperglycemia and Adverse Pregnancy Outcome Study, it has been reported that cord serum C-peptide levels are strongly related with percent body fat of newborns and mothers glycemia [26
]; although we have not data about mothers glycemia, our results based in measurement of cord insulin levels immediately after delivery, are in agree with these finding and support the statement of a strong relationship between increase of insulin by fetus and the increase of body weight at birth.
In regards the association between birth-weight and insulin resistance later in life, children and young adults with history of very low birth-weight and preterm delivery exhibit higher indices of HOMA-IR as compared with newborns delivered at term [27
]; given that SGA newborns have accelerated weight and fat mass gain during first years of life is expected the development of insulin resistance later in life in these children [30
]. Taking into account that SGA newborns in our study exhibited low HOMA-IR, is probable that the increase of leptin levels related with the increase of fat tissue have an important role in the transition from low insulin secretion and HOMA-IR at birth to high insulin secretion and HOMA-IR at childhood. Further research is needed in the field to clarify underwent mechanisms involved in the fetal insulin secretion.
An additional contribution of our study was the reference values of cord insulin levels and HOMA-IR in AGA newborns. On this regard, because there are not reference values of insulin and HOMA-IR at birth, we evaluate its distribution in an independent group of AGA newborns who had the same inclusion and exclusion criteria that newborns enrolled in the second phase of this study. In this way, we obtained the percentile distribution of insulin levels and HOMA-IR as reference value of both variables.
Finally, it is necessary to keep in mind that mechanisms involved in the relationship between fetal growth and insulin resistance have not been established with certainty; therefore, further studies in the field are necessaries.
Several limitations deserve to be mentioned. First, the small sample size of LGA or SGA newborns in this study could be a source of bias that affects our conclusion. However, given the inclusion and exclusion criteria of this study, is unlikely to find newborns at term with LGA or SGA; further research that includes multiple centers is necessary to increase the sample size in order to confirm our results. Second, the cause-effect relationship cannot be established with certainty in the cross-sectional design studies; however, other methodological designs could involve ethical issues.
The homogeneous target population and the strict matching criteria for controlling the main confounders were the main strengths of this study. In addition, given that pregnant women were in prenatal control, we have reliable data about maternal characteristics.