The main maternal factor associated with increased birth weight was greater gestational weight gain while the main maternal factor associated with greater birth length was non-smoking status. Neonatal central adiposity, determined using waist:length ratio, was negatively associated with maternal age, and positively associated with maternal smoking status, pre-pregnancy mid-upper arm circumference, trimester 3 saturated fat intake, postprandial glucose at 28 weeks gestation and membership of the control group and showed a trend towards a positive association with trimester 2 glycaemic load.
Similar to other studies [45–47], gestational weight gain was found to be positively associated with birthweight in this cohort. Guidelines from the Institute of Medicine (IOM) rely on pre-pregnancy BMI to determine an appropriate range of gestational weight gain . It is well established that those who exceed the IOM guidelines are at risk of delivering a macrosomic infant . Excess gestational weight gain increases the normal insulin resistance that occurs in pregnancy and may also affect other hormones that regulate nutrient transport across the placenta resulting in increased fetal insulin secretion, growth and adiposity . Research by Ludwig et al. into multiparous women throughout successive pregnancies has found that gestational weight gain is responsible for increased birthweight despite controlling for genetic and sociodemographic factors .
Our finding that maternal smoking during pregnancy was associated with decreased birth length is well established in the literature [50, 51]. Maternal smoking was also found to be positively associated with waist:length ratio, the equivalent of which (waist:height ratio) has been found to be a good measure of central adiposity in adults and children with a ratio of ≥0.5 indicating excess central adiposity . A recent study by Brambilla et al. found it to be a better measure of adiposity than waist circumference or BMI in children and adolescents . A meta-analysis by Ino  found that maternal smoking is associated with childhood overweight and obesity, possibly through a combination of the thrifty phenotype and catch-up growth during early infancy. Although maternal smoking during pregnancy is associated with reduced birth weight, length and relatively unchanged ponderal index in the literature [55, 56], we were unable to identify any studies that had measured waist:length ratio at birth in relation to maternal smoking status. The only studies identified that had measured waist:length ratio at birth were from the same group and involved creation of normative waist:length centile charts at birth . The increased waist:length ratio observed in the off-spring of smokers in this study appears to reflect the reduced height also observed and likely indicates that, while birthweight is not increased, central and visceral adiposity may be, increasing the risk of metabolic syndrome in later life.
Maternal early pregnancy MUAC was positively associated with neonatal waist:length ratio. MUAC has been found to be well correlated with maternal weight and BMI and remains stable in pregnancy i.e. unaffected by length of gestation . As a measure of maternal overweight and obesity, MUAC has been found to be positively associated with birthweight [59–61]. Again, there is no information on its association with neonatal waist:length ratio to date to the best of our knowledge.
Trimester 3 SFA intake was positively associated with abdominal circumference and waist:length ratio. Trimester 3 PUFA intake showed a trend towards a negative association with abdominal circumference and subscapular skinfold thickness. There is a paucity of similar data into the effect of maternal fat intake in pregnancy on neonatal adiposity. High fat isocaloric diet in rats has been found to result in no difference in birthweight of pups . Similarly, Brion et al.  found no association between maternal diet and child adiposity at 9 or 11 years of age in humans. However, the quality of dietary macronutrient intakes may be more important than absolute intakes and high SFA diet in pregnancy has been found by Murrin et al.  to be positively associated with child weight at age 5. Maternal intake of trans fatty acids in trimester 2 of pregnancy was also found by one group to be positively associated with birthweight  while other studies have reported conflicting results possibly due to examination of different trans fatty acids [17, 66, 67]. Fetal fat deposition increases with gestational age, therefore our observation that maternal fatty acid intake in trimester 3 is associated with neonatal adiposity is reasonable . Although there is little research into the effect of SFA in pregnancy there is more interest into the effect of maternal dietary PUFA intake, in particular the possible anti-obesogenic effect of a greater omega 3:omega 6 ratio. However, a recent review by Hauner et al.  found that there is little evidence of this to date due to conflicting study results. Results of a prospective intervention study involving omega 3 supplementation and dietary reduction of omega 6 showed no effect on fat mass at age 1 . This study, the INFAT study, similar to others involving fish oil supplementation found that birthweight and length of gestation were increased, however adiposity was not affected . We believe ours is the first study to show an association between neonatal central adiposity and dietary fat quality.
Trimester 2 Glycaemic load (GL) was included in the multiple linear regression model and showed a trend towards a positive association with waist:length ratio. However, while the overall model, which included trimester 2 GL, was statistically significantly associated with waist:length ratio, T2 GL was not independently significantly associated and lost significance when underreporting was controlled for. Maternal postprandial blood glucose and membership of the control group were significantly positively associated with waist:length ratio although the relationship between postprandial glucose and waist:length circumference was reduced to a trend (p= 0.050) when underreporting was controlled for. There has been much research into glycaemic control and into the effect of dietary glycaemic index and load during pregnancy on birthweight due to the well-established risk of macrosomia in gestational diabetes . The risk of macrosomia and increased neonatal adiposity has also been found to be increased towards the upper limits of normal blood glucose control in pregnancy [9, 71, 72] which is in line with our results. Low glycaemic index diet has been found to ameliorate the normal pregnancy-related increase in glucose intolerance associated with pregnancy resulting in fewer peaks in maternal postprandial glucose concentration . The positive trend towards association between postprandial glucose and waist:length ratio in this study may indicate that the reduction in postprandial blood glucose levels observed with low glycaemic index diet was associated with reduced glucose transfer to the fetus and therefore less deposition of fetal adiposity. While the ROLO study found that the intervention group reduced the GI and GL of their diet, even when underreporters were excluded from the analysis , and had reduced glucose intolerance, no difference in birthweight was observed. In contrast, a recent retrospective analysis of members of the Danish National Birth Cohort has found that those in the highest GL quintile had significantly higher offspring birthweight than those in the lowest quintile  indicating that effects may be observed at extremes. The finding by the current study, that neonatal waist:length ratio was lower in the intervention group indicates that improved dietary carbohydrate quality may be associated with reduced central adiposity rather than birthweight at less extreme levels as the reduction in GI and GL observed in this study was quite modest.
This was a large, well-powered clinical trial designed to examine the effect of diet and lifestyle on neonatal size and adiposity. Rich dietary data was available at each trimester in combination with biochemical measures of glycaemic control. One limitation of this study was that detailed neonatal anthropometric data was not available for the full cohort, however weight and length measurements were taken for all infants. It should also be noted that this was a cohort at risk of macrosomia, therefore care should be taken regarding generalisation of results to other populations. A further limitation was that fatty acid composition was not broken down into omega 3, omega 6 and trans fats. This clinical trial was originally focused on the effect of GI, GL and glycaemic control on birthweight but nevertheless, data on SFA, MUFA, PUFA and total fat was available and this is the first human study to report an association between dietary fat quality and neonatal central adiposity.