The INTERGROWTH-21st Project

The INTERGROWTH-21st Project was established to address this fundamental gap in our understanding of early human growth. The Project aimed to determine how fetal growth, preterm postnatal growth, and neurodevelopment in the first 1000 days of life occur under optimal conditions, and whether or not this growth is similar enough around the world to justify the use of a single set of inter­national growth standards (29).

The INTERGROWTH-21st Project was based on uniquely de­tailed methodology (29) and all study protocols and primary find­ings are available freely online (https://intergrowth21.tghn.org). Briefly, eight diverse urban populations living in demarcated geo­graphical or political areas were selected where environments were free from major known pollutants; altitude was less than 1600 m; most women accessed antenatal and delivery care in institutions; mean birth weight was greater than 3100 g; rates of LBW (labour dystocia and the need for operative birth.

Aetiologies and risk factors for LGA

Diabetes in pregnancy

Rates of diabetes in pregnancy around the world have now reached epidemic levels. The vast majority of cases of diabetes in pregnancy are GDM, with smaller proportions of women entering preg­nancy with known type1 or type 2 diabetes (50). The International Diabetes Federation estimated in 2015 that one in seven pregnant women had GDM (51). Most of these women are unaware of their diagnosis and live in societies with poorly developed health systems to screen or manage diabetes and the associated complications in pregnancy.

During pregnancy, placental hormones create a state of relative insulin resistance in the mother to facilitate the transfer of amino acids and glucose to the fetus. The degree of insulin resistance that develops by the third trimester is similar to that observed in people with type 2 diabetes, requiring the maternal pancreas to increase insulin secretion by almost 250% to maintain, who are either unable to increase their insulin secretion, or who already have borderline levels of peripheral insulin resistance, or both.

The high proportion of women with GDM has been attrib­uted to current unprecedented rates of overweight and obesity in reproductive-aged women (52); women delaying childbearing until the fourth decade; and a relative abundance of calorie-r ich and nutrient-poor foods, particularly in deprived social areas and in countries undergoing economic transition (53).

The other important contributor to the rate of GDM relates to changes in the definition of the condition itself. The landmark Hyperglycemia and Adverse Pregnancy Outcomes (HAPO) study demonstrated a linear and increasing association between maternal glycaemia at 28 weeks and birth weight, need for primary caesarean section, cord blood C peptide (a marker of fetal hyperinsulinaemia), and neonatal hypoglycaemia (54). These findings formed the basis for a revised definition of GDM. This was based on an increased relative risk of 1.75 of one or more adverse perinatal outcomes as­sociated with GDM, including LGA (based on birthweight >90th centile on local charts), primary caesarean section, neonatal hypo­glycaemia, and cord C peptide (55).

In 2015, the National Institute for Health and Care Excellence (NICE) in the United Kingdom updated its recommendations on the screening and treatment of GDM (50). Women who experienced GDM in a prior pregnancy are offered oral glucose tolerance testing as early as possible in pregnancy, while other women with risk fac­tors are screened for GDM around 26-28 weeks' gestation. While there remains controversy around diagnostic thresholds for GDM, values higher than the current diagnostic criteria should trigger a system to provide dietary advice, advice on blood glucose moni­toring, and multidisciplinary care (50). Frequently, the diagnosis of LGA is not made until late in pregnancy in women without GDM. Interpreting results from glucose tolerance tests very late in preg­nancy can be challenging and there is little evidence to guide ap­propriate screening for diabetes in women detected to have a large baby close to term.

The aim of clinical management in women with all forms of dia­betes in pregnancy is to normalize blood glucose to decrease the risk of complications associated with hyperglycaemia. In women with GDM there is level 1 evidence that screening and treatment of GDM can reduce birthweight (56) and severe adverse outcomes in the offspring (death, brachial plexus injury, skull fracture) (57). Normalizing glycaemic control in women with diabetes is one of the few interventions that have been shown to decrease birth size, although tight control of blood glucose in pregnancy can increase the chance of maternal hypoglycaemia, poor fetal growth, and low newborn body fat percentage (58), highlighting how difficult this condition can be to optimally control.

Maternal overweight and obesity

Another risk factor for LGA is maternal obesity, with estimates indicating the risk of macrosomia to be approximately doubled when compared to women with a normal body mass index (BMI) (odds ratio 2.1; 95% CI 1.6-2.6) (59). Several approaches have been taken to reduce fetal overgrowth associated with obesity, including intensive lifestyle and pharmacological interventions; however, re­sults have been disappointing.

The UK Pregnancies Better Eating and Activities Trial (UPBEAT) and the Limiting Weight Gain in Overweight and Obese Women during Pregnancy to Improve Health Outcomes (LIMIT) trial, both assessed the effects of lifestyle interventions to improve pregnancy outcomes in obese pregnant women. The UPBEAT trial was con­ducted in the United Kingdom and involved 1555 obese pregnant women (772 allocated to standard antenatal care and 783 to the intervention). The intervention involved weekly lifestyle and dietary counselling informed by social change theories. There was no dif­ference between the groups in the primary endpoints of a difference in rates of GDM and LGA babies, although some secondary out­comes such as maternal weight gain were improved in the interven­tion group (60).

The effect of metformin to reduce birthweight in obese women without gestational diabetes has also been assessed in two random­ized controlled trials from the United Kingdom. The Efficacy of Metformin in Pregnant Obese Women, a Randomised controlled (EMPOWaR) trial assessed the effect of 2.5 g of metformin daily from 16-18 weeks gestation on birth outcomes in 449 women. A second United Kingdom study assessed the effect of 3 g of metformin in 450 obese women. Neither of these studies demonstrated any evidence of effect on birthweight (62).

The effects of maternal obesity on fetal growth may have their ori­gins very early in pregnancy. In animal models there is evidence that exposure to obesity in either parent changes the molecular structure of the oocyte or sperm, and that this can alter the process of epi­genetic reprogramming that occurs after conception (63). Current interventions targeting gestational weight gain in obese gravidas may therefore be coming too late, with a life course focus on women's health needed if we are to seriously tackle the problem of intergen- erational obesity.

Gestational weight gain

Excessive gestational weight gain is also associated with acceler­ated fetal growth. Despite decades of research, however, there re­mains controversy in the value and recommendations for optimal gestational weight gain in pregnancy with most of the evidence which current recommendations are based on being poor (64). The INTERGROWTH-21st Project has published weight gain trajec­tories in low-risk women with normal BMIs (65). Robust evidence for women in other weight groups is lacking, and it is often women in these weight groups who pose the greatest clinical challenge. The Institute of Medicine 2009 guidelines for gestational weight gain are widely used in clinical practice. These guidelines are an expert synthesis of the best available evidence and try to balance the peri­natal risks of inadequate or excessive weight gain with the risks of postpartum weight retention, a significant contributor to obesity in many women.

Fetal genetic overgrowth syndromes

Detection of very severe LGA, particularly early in pregnancy or in the presence of other abnormalities, should prompt expert re­view in a fetal medicine unit. Somatic overgrowth syndromes, while relatively rare, can begin in the prenatal period and can be associated with genetic syndromes that are associated with neurodevelopmental delay and increase the risks of neoplasia. Generalized overgrowth syndromes that can be detected in utero in­clude Beckwith-Wiedemann syndrome, Simpson-Golabi-Behmel syndrome, Sotos syndrome, Weaver syndrome, Marchall-Smith syndrome, and Perlman syndrome. For an overview of these rare conditions, the reader is referred to the review by Yachelevich, pub­lished in 2015 (66).

Detection of LGA

Traditionally, the detection of LGA babies was made clinically, ei­ther by palpation, symphysiofundal height measurement, or by ma­ternal suspicion; however, all these methods have been reported to have poor detection rates of less than 50% (67). If a large baby was suspected, commonly ultrasound biometry would be performed to measure the size of the baby and approximate the volume of amni­otic fluid. The diagnosis of LGA from ultrasound is usually based on the finding of the estimated fetal weight at greater than the 90th or 97th centile, depending on local practice. As this value is based on regression equations combining two or more measures of fetal biometry, it can be inaccurate at the upper weight limits. The density of babies is not uniform across the size distribution, and typically ac­curacy of estimated weight reduces in larger babies. A review of the published literature found the post-test probability of ultrasound to predict LGA to range from 15% to 79% (68).

Given these limitations, several groups have attempted to improve the diagnosis of LGA by utilizing three-dimensional volumetric measures of fetal structures, such as the thigh volume, or novel two­dimensional measurements of fetal fat deposits, such as cheek-to- cheek diameter, abdominal wall thickness, and upper arm diameter.

While these procedures show promise in improving diagnostic accuracy, due to technical difficulty, cost, and time they are currently limited largely to research settings.

Evidence for clinical management

The clinical management of LGA babies in pregnancy depends on the individual factors likely to be driving the growth and other co- morbid complications. For obstetricians, the most important de­cisions need to be made around the timing and mode of delivery. These decisions will be different in women with LGA and diabetes compared to those women with LGA who are not known to have diabetes.

LGA in pregnancies complicated by diabetes

Decisions for delivery timing and mode for women with all forms of diabetes in pregnancy are based on concerns of excessive fetal growth, an increased risk of shoulder dystocia with disproportionate fat deposition in the fetus, as well as a fivefold increased risk of still­birth in women with pre-existing diabetes. The 2015 NICE guideline recommends delivery by 41 weeks for women with well-controlled GDM with dietary modifications (50). Delivery should be sooner in women with pre-existing diabetes (37-38 weeks), or in women with GDM requiring medication with either poor glycaemic control or signs of macrosomia (50).

Non-diabetic pregnancies

In women without diabetes carrying a LGA baby, there has been little evidence to support induction before term to prevent shoulder dystocia. In 2015, however, a multicentre randomized controlled trial published in The Lancet and examining whether induction of labour by 39⁺⁶ weeks in babies estimated to have a birth weight greater than the 90th centile compared to expectant management could actually reduce this risk. The trial recruited 822 women (409 in the induction group and 413 in the expectant management group) and demonstrated a significant reduction in the primary outcome, shoulder dystocia (RR 0.32; 95% CI 0.15­0.71; P = 0.004) with planned induction of labour. The women in the induction group did not experience a higher rate of caesarean section and were actually more likely to delivery spontaneously (RR 1.14; 95% CI 1.01-1.29) (69). This trial is the largest and most rigorous to address the management of LGA babies. While none of the babies who experienced the primary outcome of shoulder dystocia in labour suffered permanent brachial plexus injury or death, the lack of harm from the intervention and potential for improved spontaneous vaginal birth rates are important. Based upon this evidence, a policy of detection of LGA in the third tri­mester and induction of labour before 39 weeks would seem sens­ible, although international guidelines have not yet endorsed this change in practice. Before clinical practices are changed there must be consideration of local resource constraints for the provision of ultrasound to detect LGA, as well as the capacity of the unit to provide safe induction of labour. As with all evidence from ran­domized trials, there must also be consideration as to whether this evidence is applicable to individual patients, for example, the small risk of shoulder dystocia may be outweighed by the risk of induc­tion of labour in women with a previous caesarean section hoping to have a vaginal birth.

Longer-term outcomes associated with LGA

Around the world, children are getting heavier and childhood overweight and obesity is recognized to be a significant global problem (70). Being born large is associated with an increased risk of childhood obesity, although establishing the direct causal link is challenging due to multiple potential confounding dietary and so­cioeconomic factors (71). The observational epidemiological data consistently support a positive association between size at birth and childhood weight. Part of the effort to reduce childhood obesity must include strategies to limit excessive fetal growth in pregnancy where possible (72). As previously discussed, this is not an easy out­come to change, and any advice to expectant mothers must be given in a way that balances the real risks of fetal overgrowth with risks of mothers drastically altering their food intake to prevent these out­comes and causing problems of inadequate fetal growth.