Description of the condition

Large‐for‐gestational‐age (LGA) is used to describe fetuses or infants with a suspected disturbance of growth, where actual or estimated growth is more than expected for gestational age (Xu 2010).

There are various definitions used to describe growth of the baby during pregnancy. LGA is usually defined as an estimated fetal weight more than the 90th percentile (Pundir 2009). Macrosomia, a frequently‐used term, refers to growth beyond a specific threshold, usually defined as birthweight of 4000 g or more (Delpapa 1991), although some studies have suggested grading macrosomia depending on birthweight, into grade one (4000 to 4499 g), grade two (4500 to 4999 g) and grade three (5000 g or more) (Boulet 2003).

There is a wide variation in the prevalence of macrosomia in different countries. Amongst European countries, prevalence of grade one macrosomia varies between 8% and 21% (Bergmann 2003; Gyselaers 2012; Orland 2012). Macrosomia prevalence in Asian countries tends to be much lower, at 1% to 8% (Li 2014; Morikawa 2013). The worldwide prevalence of grade one macrosomia is approximately 9% and the prevalence of grade three macrosomia is approximately 0.1% (Chauhan 2005).

LGA is associated with a number of maternal or fetal factors. The majority of cases of LGA infants are associated with maternal factors including maternal height, weight, body mass index (BMI), gestational weight gain, ethnicity, parity and maternal age (Bergmann 2003; Stotland 2004), as well as the presence of pre‐gestational or gestational diabetes (Ehrenberg 2004). Fetal factors contributing to LGA infants include fetal sex, post‐term birth, uncertain dates and, less commonly, fetal genetic abnormalities (Vora 2009).

LGA is known to be associated with increased risks to both the mother and baby. Fetal and neonatal implications include increased risk of intra‐uterine death and stillbirth, fetal hypoxia, shoulder dystocia, brachial plexus injuries, bone fracture, hypoglycaemia, and admission to the neonatal intensive care unit (Esakoff 2009; Ju 2009). Maternal complications include prolonged labour, operative births including caesarean section, perineal trauma, postpartum haemorrhage and uterine rupture (Ju 2009; King 2012; Stotland 2004).

It can be difficult to differentiate accurately between a constitutionally LGA infant, due to maternal height, weight and ethnicity, and an infant with abnormally elevated growth for gestation.

Description of the intervention

Management of suspected LGA fetuses requires several steps. Initially, those infants at risk of LGA are identified either by clinical assessment, maternal estimation, or knowledge of risk factors such as diabetes. Once a fetus is suspected of being LGA, further assessment to assist in the diagnosis is performed, such as an ultrasound scan. Following this, antenatal care and surveillance is ongoing and eventually decisions must be made regarding the mode and timing of birth.

Once a baby is suspected to be LGA, antenatal care providers may utilise a variety of methods, or regimens, to monitor the baby’s growth and well‐being. The type and frequency of monitoring techniques vary considerably, and the choice of methods may be affected by other factors such as the presence of pre‐gestational or gestational diabetes mellitus (Campaigne 2007).

Antenatal fetal surveillance methods aim to provide information about the health of the baby (Grivell 2012). The results of surveillance methods then guide decisions about ongoing surveillance and the mode and timing of delivery, which may in turn reduce the risks to both the baby and the mother associated with fetal overgrowth and suspected LGA fetuses.

There are many methods that may be used in clinical practice, alone or in combination, to monitor suspected LGA fetuses and to assess those at risk (Grivell 2012). These include abdominal palpation, measuring symphysial fundal height (SFH), ultrasound assessment of baby’s growth, fetal movement counting, fetal heart rate assessment (cardiotocography), ultrasound assessment of the baby’s well‐being by serial growth measurements or liquor volume, and assessment of maternal glycaemia (Aye 2010; Chauhan 2005; Grivell 2012).

When LGA is suspected in a fetus, either from clinical assessment or maternal estimation, ultrasound is often used to assess growth using various standard measurements. These may include: abdominal circumference, head circumference, bi‐parietal diameter and femur length. These measurements can then be used to estimate fetal weight. Many formulae have been proposed, with the most widely used being Hadlock's (Hadlock 1985). Additional parameters to identify LGA infants and assess fetal well‐being include subcutaneous fat measurements on ultrasound, amniotic fluid volume, three‐dimensional (3D) volumetric estimates, umbilical artery dopplers and fetal magnetic resonance imaging (MRI) (Aye 2010; Chauhan 2005; Orland 2012). Ultrasound measurements can then be plotted against gestation along known percentile lines on a growth curve, to assess severity and timing of increasing growth. Serial ultrasound measurements can then identify those infants who are suspected to be LGA as well as those with acceleration of growth over time.

Ultrasound surveillance has variable accuracy in identifying LGA fetuses. The post‐test probability of detecting a macrosomic fetus on ultrasound has been assessed in a systematic review (Chauhan 2005); in an uncomplicated pregnancy it is variable, ranging from 15% to 79%. In comparison, the post‐test probability of clinical estimate in predicting a macrosomic fetus is 40% to 52% (Chauhan 2005). The use of a single early third‐trimester fetal biometry for the prediction of birth weight deviations showed 90% sensitivity for detection of macrosomia (De Reu 2008).

Macrosomia detection in higher‐risk populations such as women with diabetes has more consistency. The above systematic review found sensitivities ranging from 33% to 69%, specificities from 77% to 98%, and post‐test probabilities from 44% to 81%, in detecting macrosomia in women with diabetes (Chauhan 2005).

3D ultrasound has been used to identify suspected LGA fetuses, in particular by measuring subcutaneous fetal fat. There is some evidence that this improves weight prediction to within 5% of the actual weight (Lee 2001). However, this technique is not widely used.

Fetal MRI may have better predictive value in fetal weight estimation compared with ultrasound (Uotila 2000). However, due to limited availability and high cost, it is not routinely used in the diagnosis of macrosomia.

Although accurate, ultrasound is expensive when used as a screening tool for abnormal growth detection. A low‐cost option for detecting abnormal fetal growth is SFH.

A Cochrane review found insufficient evidence to determine whether SFH measurement is effective in detecting abnormal fetal growth, but only one trial was eligible for inclusion (Robert Peter 2012).

There is concern that identifying suspected LGA fetuses may lead to unnecessary maternal anxiety, investigations and interventions. A prospective observational study found that after an ultrasound, women with a suspected macrosomic fetus were more likely to have induction of labour or caesarean section, compared with actual macrosomia in the comparison group of all other births (Sadeh‐Mestechkin 2008). A previous Cochrane review found that induction of labour is not recommended for suspected macrosomia, as it does not alter the risk of maternal or neonatal morbidity (Irion 1998). Our review will therefore assess the risks as well as benefits of regimens of fetal surveillance for suspected LGA fetuses.

The estimated fetal weight from an ultrasound, or a SFH measurement, can be plotted on a customised growth chart, which adjusts for factors such as the mother's weight and height, ethnicity and number of babies she has had. This may make assessment of fetal growth more precise, differentiating between constitutional LGA and fetal overgrowth, reducing unnecessary consultations for investigations, and parental anxiety (González 2013).

How the intervention might work

We anticipated that by increasing the frequency of monitoring of a suspected LGA fetus, or using different methods of surveillance in combination, it would be possible to identify babies that would either benefit from earlier birth, or benefit from intervention to slow growth acceleration. Interventions to slow growth acceleration include dietary advice, lifestyle modification, and in women with diabetes or gestational diabetes blood glucose monitoring and insulin therapy (Aye 2010; Campaigne 2007; Chauhan 2005). This in turn may be associated with improved health outcomes for the mother and baby.

Why it is important to do this review

There has been a rise in the prevalence of LGA babies over the past few decades in many countries (Bergmann 2003; Gyselaers 2012; Lahmann 2009). In high‐income countries, there has been an overall 15% to 25% increase in the proportion of women giving birth to large infants (Aye 2010). Data are not as easily available for low‐ to middle‐income countries, where the risks are even greater due to the reduced availability of operative delivery options (Koyanagi 2013).

There is evidence that LGA is associated with an increased risk of maternal and neonatal complications. However, policies and protocols for fetal surveillance where accelerated fetal growth is suspected vary widely, which has led to uncertainty among clinicians about appropriate surveillance.

This systematic review aimed to assess the benefits and harms associated with different combinations of surveillance methods for the fetus estimated to be LGA.