One Fetal Movement by benbenzhou


More Info
									Fetal Growth Assessment                                                          Background

                                    Chapter One


1.1     Introduction

Fetal size is associated with pregnancy outcome. The growth-restricted fetus is at

increased risk of pre term delivery, perinatal death and infant morbidity and mortality

due to hypoxia, acidosis, hypoglycemia and hypocalcemia. The fetus with macrosomia

(birth weight > 4000g) is at risk of stillbirth, perinatal asphyxia, meconium aspiration

and birth injuries such as shoulder dystocia and fractures. Neonatal complications for

these large babies include hypoglycemia, hypocalcemia and jaundice whilst the mother

may suffer pre-eclampsia, postpartum haemorrhage and perineal trauma. Features that

predispose to a large baby include gestational diabetes mellitus (GDM), a large

maternal stature, increased maternal age, sedentary lifestyle or maternal obesity.

Estimating fetal size/growth should be an important component of obstetric care,

particularly in pregnancies complicated with gestational diabetes and fetal overgrowth

that can lead to a macrosomic baby with its associated birth complications. The

following literature review confirms that fetal size estimation can be difficult due to

many contributing factors such as maternal body habitus, uterine distortions due to

pathology, maternal ethnic variations as well as different techniques used to calculate

the size of the fetus.

        This thesis reviews the variability in assessing gestational age, fetal size, growth

and weight and the variations seen in different racial groups. It then considers the

diverse approach to obtaining accurate ultrasonic measurements for many parts of the

Fetal Growth Assessment                                                        Background

fetal body and the mathematical modelling for fetal size/growth curve formatting. The

thesis also examines fetal growth and ultrasonic estimation of fetal size, in the third

trimester of pregnancy in both Caucasian and Chinese women living in Australia. The

Chinese population was chosen due to the increasing number of immigrants settling

within the Northern Sydney Health Area in New South Wales, Australia. The statistical

variations seen in birth weight, interventions and complications between the Caucasian

and Chinese at a local hospital during a routine audit prompted the question of ethnic

differences influencing birth outcomes. This instigated the research that led to this

thesis, which investigates pregnancies at risk of fetal macrosomia, both with and

without gestational diabetes mellitus and the prevalence of birth intervention and

complications, such as post partum haemorrhage in pregnancies thus affected. Finally

the chapter summarises the works of the authors and lists the objectives of this study.

1.2    Fetal Growth

Fetal growth is determined by a complex interaction of genetic, environmental and

socio-economic factors (Vorherr 1982) with normal birth standards based on a

combination of gestational age at birth, head size, length and birth weight (Varner

1987). Maternal risk factors strongly influence fetal size (Goldenberg 1993) and when

these are combined with individual maternal profiles and population variances it could

help explain the difference seen in fetal growth. The recommended maternal weight

gain in Caucasian pregnancies, according to Taylor and Pernoll (1976) is between 10-

12 kg with the average fetus accounting for approximately 3500 g of this gain. The

remainder of the weight gain is from amniotic fluid (750g), placenta (1000g), blood

volume (1500g), interstitial fluid (1500g), breasts (500g) and maternal fat upwards of

Fetal Growth Assessment                                                        Background

1500g. Taylor and Pernoll (1976) further described the various birth weight outcomes

as follows. An immature infant is between 20 and 28 weeks gestation and weighs less

than <1000grams whilst a premature infant is between 28 and 38 weeks gestation and

weighs between 1000-2500 grams. A low birth weight infant is less than 2500g, a

small for gestational age infant is less than two standard deviations below the mean

weight for gestation, a mature infant weighs more than 2500g and is between 38-40

weeks gestation, a post-mature infant is more than 42 weeks gestation and finally the

excessive size infant which weighs more than 4500 grams and termed macrosomic.

       Normal fetal growth is not only reflected in birth weight but is defined

according to population standards and percentiles (Brooke et al 1981, Dunn 1985,

Gardosi et al 1992a, 1995b). Fetal measurement graphs and birth weight graphs are

divided into standard deviations or percentiles where 50% of the measured population

will be above the mean and 50% below the mean of the relevant graphs. The smallest

10% will be classed as less than the 10th percentile and therefore small for gestational

age and the largest 10% greater than the 90th percentile will be large for gestational age

(Deter 1981, Gallivan & Robson 1993, Jeanty 2001). Eight to ten percent of all

Caucasian births are macrosomic, or have a birth weight above 4000 grams, which

places them above the 90th percentile (Benson and Doubilet 1998) and 2% of these

macrosomia births are greater than 4500 grams. Individual fetal growth curves are

created by plotting gestational age and estimated fetal weight onto a weight graph

displaying at least the 10th, 50th and 90th percentile limits (McCalum & Brinkley 1979,

Hadlock 1990, Jeanty 2001).     Brandt (1963) and Naey & Tafaril (1985) supported the

theory that nutrition and lifestyle during pregnancy influenced birth weight. Taylor et

al (1976) was of the opinion that any increase in mean birth weight over time may be

Fetal Growth Assessment                                                       Background

due to better nutrition but there has been no corresponding change in female pelvic size.

This fact alone can account for any increase in birth intervention and complications.

1.2.1 The Small for Gestational Age Fetus

The small for gestational age (SGA) fetus as mentioned previously, is less than the 10th

percentile below the population mean for the stage of gestation (Hughey 1984). SGA

may be as a result of a combination of factors including obstetric background (Visser

1986), maternal stature, genetic disorders and maternal diseases such as hypertension,

anemia, heart disease, renal disease, malabsorption and autoimmune disease (Varner

1987).   Chesley (1978) termed the phrases “disease of theories” to describe pre-

eclampsia with one of the theories being compromised placental perfusion. Unstable

maternal pre-eclampsia can cause utero-placental insufficiency and placental abruption,

which may result in intrapartum fetal distress or stillbirth (Fox 1987). Spirt and Gordon

(2001) described placental insufficiency as being a “nonspecific clinical term with no

known morphologic basis” and said that decreased uteroplacental blood flow may be

due to anoxia secondary to maternal vascular problems. Sebire and Talbert (2001)

analysed the placenta with a closer look at the pathophysiology of placental

hemodynamics in uteroplacental compromise and concluded that the dynamic placenta

allows local control of vascular smooth muscle tone, which in turn regulates the

ventilation and perfusion differences. Up to 30% of placental attachment can be lost

with no corresponding loss of function (Crawford 1962). Reduced placental function is

the major cause of small for gestational age (Spirt & Gordon 1984a, 2001b) as a result

of placenta previa, infarction and abruption (Varner 1987). Placental abruption occurs

in up to 1 in 89 deliveries with fetal death occurring in 1 in 500 to 750 deliveries

Fetal Growth Assessment                                                      Background

(Pernoll 1987) and can be diagnosed with ultrasound (Nyberg et al (1987). Mabie and

Sibai (1987) cited the association of maternal pre-eclampsia with pre-term delivery and

the complications arising from pre-maturity including hypothermia, meconium

aspiration syndrome, hypoxia and acidosis, polycythemia and long term complications

such as a lower IQ and learning disorders (Varner 1987). Nicolaides et al (1988)

showed that an absence of end diastolic frequencies in the umbilical artery was a sign

of fetal hypoxia and acidosis. Detection of the small fetus is critical for an improved

obstetric outcome and a number of authors including Hadlock (1982), Seeds (1984),

Hughey (1984) Deter (1992), Benson & Doubilet (1998) and Jeanty (2001) used

ultrasonic fetal measurements to assess fetal size so that early clinical management

could be implemented.       Chang et al (1993) went further by comparing doppler

waveform indices and serial ultrasound measurements of abdominal circumference and

fetal weight to identify fetal growth retardation.

1.2.2 The Large for Gestational Age Fetus

The large for gestational age (LGA) fetus is defined as being above the 90th percentile

at any stage during pregnancy (Varner 1987). This is the same definition given for fetal

macrosomia, or hypersomatism, by Golditch and Kirkman (1978), Hadlock (1989),

Shah (1993), Benson (1998) and Jeanty (2001). Macrosomia is described in greater

detail in section 1.4. Normal fetal growth in the third trimester may be assessed not

only clinically but by ultrasonically measuring fetal parameters, such as head, abdomen

and femur, then applying these measurements to either a graph or a fetal weight

formula. Hadlock (1982) thought that a fetus can be defined as having normal growth

if the head circumference (HC) / abdominal circumference (AC) ratio is 10 - 90%.

Fetal Growth Assessment                                                        Background

Hadlock‟s definition was confirmed by Deter (1982), Doubilet and Benson (1990) and

Jeanty (2001) amongst others. A large fetus will have a head circumference (HC) to

abdominal circumference (AC) ratio of greater than 90% whilst a fetus with

macrosomia will show the AC > HC > 90% (Hadlock 1985). Schlater (1987) said that

a difference between the head circumference and chest circumference of 1.6cms was an

indicator of macrosomia but the clinical usefulness of this theory is questionable as the

measurement has not been implemented into any routine ultrasound scanning protocols

currently in use by Australian ultrasound centres. Fetal weight gain can be estimated

by the difference in weight between two ultrasound examinations (Garrett & Robinson

1971, Ogata 1980, Gallivan 1993).     Benson and Doubilet (1998) created fetal weight

percentiles in the third trimester, with the resultant data showing mean fetal weight gain

increases by approximately 240g per week until 37 weeks gestation, after which the

rate of gain in a normally growing fetus gradually declines.

1.3    Clinical Evaluation of Fetal Size and Growth

Physicians and midwives rely on physical examination of the pregnant woman to

estimate fetal size by uterine palpation and measurement of the fundal height. This

technique is made more difficult by large maternal body habitus and uterine fibroids or

other pelvic masses, which can give a false indication of size, as well as amniotic fluid

and the bulk of the placenta (Benson 1998). Multiple pregnancies are particularly

difficult to evaluate (Benson 1987, MacGillivray 1978). Taylor and Pernoll (1976)

described the fundal height, measured from the symphysis pubis, as one of the more

reliable clinical examinations to assess fetal size with the height of the fundus in

centimetres, being equivalent to the week of gestation from about 22 to 36 weeks.

Fetal Growth Assessment                                                          Background

When estimating fetal size for mode of delivery one of the methods historically used

for calculating fetal weight was Johnson‟s formula, which could only be used if the

fetus is in the vertex position:

                Fetal weight (grams) = ( fundal height (cms) – n) x 155

If the head is above the ischial spines then n = 12 or, if head is below the ischial spines

then n = 11. For patients over 90kg 1cm is subtracted from the fundal height prior to

calculation. Although clinically irrelevant nowadays Pernoll (1987) claimed that this

formula was accurate to within 375 grams in 75% of cases even though it did not make

allowances for the fetus in breech or transverse lie.

        Infant size can be related to maternal stature (Varner 1987, Hardy 1999) and

Roher‟s ponderal index (PI) is used to assess infant symmetry with a weight for length

relationship at birth. The normal ponderal index is 1.8 at 28 weeks gestation, which

increases by 0.05 per week to be 2.8 at 40 weeks gestation. An asymmetric infant will

have a low PI if it is small for gestational age or a high PI if it is large for gestational


                        PI = birth weight x 100 / (crown-heel length)3

1.3.1 Predictors and Consequences of Excessive Growth

Although the mean weight of Caucasian infants has remained relatively constant over

the past ten years the rate of macrosomia births (greater than 4000g), according to

Northern Sydney Health statistics (2003), is increasing with 10% of births with

Fetal Growth Assessment                                                    Background

macrosomia in 1992, 14% in 2000 and over 15% in 2003. The Australian Bureau of

Statistics Report (2003) shows this trend is reflected in other centers throughout


       Catanzarite et al (1976) stressed the importance of antenatal fetal assessment,

suggesting, like Taylor and Pernoll (1976) that fundal height measurements could give

an indication of excessive growth.    Catanzarite was in favour of fetal monitoring,

where appropriate, with the aim of detecting those fetuses being compromised by the

pregnant environment. Excessive fetal growth can often be associated with maternal

profile with a high body mass index (>30), large stature, obesity, increased age,

previous infant with macrosomia, postdatism and diagnosed diabetes or gestational

diabetes being indicative of a pregnancy at risk of fetal overgrowth. Varner (1987)

described some of these risk factors, giving rates of increased risk with maternal

obesity for example having a 4 to 12 fold risk of an infant with macrosomia. Hardy

(1999) assessed macrosomia predictors amongst a multiethnic group, concluding that

the women‟s non-pregnant body mass index followed by blood glucose levels could be

used to identify the pregnancies at risk of macrosomia. Varner (1987) maintained that

between 28% and 53% of fetal macrosoma can be identified by a thorough clinical

history and physical examination of the mother and also suggested that birth weight

correlated more closely with maternal stature rather than maternal weight. Schlater

(1987) disagreed with this, believing that diagnosing fetal macrosomia by abdominal

palpation to be “notoriously inaccurate”. If excessive fetal growth is identified then

appropriate action can be taken to lessen the risks to both the mother and baby during

birth. The consequences of excessive growth can be divided into maternal, fetal and

neonatal risks.

Fetal Growth Assessment                                                       Background

1.3.2 Maternal Risks

The risk to the mother with a large for gestation (LFG) infant include postpartum

haemorrhage, reproductive tract injury, perineal injury with vaginal delivery and the

risks of an emergency caesarean section and postpartum hysterectomy (Bhatia et al

1987). Failure to progress in labour is a precursor for fetal distress and emergency

caesarean section (Varner 1987). Pre-eclampsia tends to occur in the second half of

pregnancy and is a condition combining hypertension, edema and proteinuria which,

according to Mabie and Sibai (1987), affects up to eight percent of all pregnancies and

accounts for nearly 20% of maternal deaths (McDonald 1978).              The risk of a

subsequent large baby is up to four times greater (Varner 1987) and so it is important to

check for latent gestational diabetes mellitus in the early second trimester (Crowther et

al 2005). Post partum haemorrhage (PPH), defined by Kapernick (1987) and Novy

(1982) as blood loss greater than 500ml following vaginal delivery, occurs in up to 8%

of vaginal deliveries and is directly responsible for approximately 16% of maternal

deaths. PPH may be caused by an atonic uterus due amongst other things, to over

distension, lacerations or episiotomy from delivery of a large baby, coagulation

problems or retained placental tissue. Episiotomies and lacerations account for 20% of

PPH cases, the atonic uterus is responsible for 50% of PPH cases and retained placenta

5-10% (Kapernick 1987). PPH may occur up to six weeks post delivery and this

influenced Lee et al (1981) to assess the postpartum uterus with ultrasound to detect

any retained products of conception such as placental remnants.                Placental

abnormalities can increase the likelihood of PPH. One such abnormality is placenta

accreta, which can be assessed with ultrasound, which is where the placenta adheres

Fetal Growth Assessment                                                         Background

directly to the myometrium instead of being separated by a layer of deciduas (Tabsh

1982, Pasto 1983). Placenta accreta may be present in as many as 1 in 2000 births. In

Australia PPH is responsible for a maternal mortality of 1:100,000 births (National

Health Statistics 1998) and for significant morbidity, including the need for transfusion

and/or postpartum hysterectomy with further associated risks.

1.3.3 Fetal Complications

The higher the fetal weight the higher the incidence of birth complications (Jung 1997

& Hod 1998). Excessive fetal growth results in increased perinatal morbidity and

mortality due to delayed progress in labour, traumatic vaginal deliveries, which are up

to 23% of all macrosomia births, and emergency caesarean section (Dunsted et al

1985).    The problems of difficult labour due to the large fetus may cause fetal

complications including the risk of still-birth, increased incidence of birth injuries such

as shoulder dystocia, fractures, facial palsy and birth asphyxia leading to cerebral palsy

and mental retardation (Schlater 1987). Excessive uterine distension due to a large

fetus can be a predisposing factor to placental abruption (Schlater 1987), which, as

mentioned earlier, occurs in up to 1 in 89 of deliveries with fetal death occurring in 1 in

500 to 750 of deliveries (Pernoll 1987). Pulsed doppler assessment of the umbilical

cord arteries can assist with identifying the macrosomic fetus that is continuing to grow

(Trudinger (1991). The flow velocity waveforms from the umbilical cord arteries will

display a high resistant pattern when the fetus ceases to grow. Golditch and Kirkman

(1975) proposed that the large fetus could be managed by offering elective caesarean

section. Gonik et al (1983) and Harris (1984) suggested maneuver techniques to assist

with avoiding shoulder dystocia but even with the implementation of these ideas

Fetal Growth Assessment                                                      Background

shoulder dystocia still occurs in up to 10% in cases of macrosomia. Varner (1987)

surmises that males account for sixty percent of macrosomic births and that there is

about 150 gram difference between the sexes for each week of gestation in late


1.3.4 Neonatal Risks

Neonatal complications of excessive growth include low apgar score, birth injury,

hypoglycemia, which increases the risk of neuro-behavioural problems (Bhatia and

Sokol 1987), hypocalcaemia, polycythemia, jaundice, feeding difficulties and

admission to neonatal intensive care unit.

1.4    Macrosomia

The Oxford Medical Dictionary defines macrosomia as:

“ Macrosomia. n. abnormally large size. In fetal macrosomia a large baby is

associated with poorly controlled maternal diabetes. The increased size is due to

excessive production of fetal insulin and thence to increased deposition of glycogen in

the fetus.”

Fetal macrosomia is a term used to describe fetuses that are greater than the 90th

percentile on fetal growth charts, with babies that are born greater than 4000 grams in

weight described as being macrosomic (Benson 1988, Hadlock 1989, Shah 1993).

Macrosomia is a clinical problem that can lead to birth injury, intervention, postpartum

Fetal Growth Assessment                                                         Background

haemorrhage and other adverse outcomes. Northern Sydney Area Health Statistics

show that macrosomia births are increasing not only for Caucasians but for other ethnic

groups. Although Rodrigues et al (2000) and Hadlock (1989) defined macrosomia as

birth weight greater than the 90th percentile for gestational age, the actual weight at

which macrosomia is called is debatable. Whilst some authors such as Modanlou,

Dorchester et al (1980), Boyd and Usher (1983) and Varner (1987) defined macrosomia

as birth weight greater than or equal to 4500 grams others such as Schlater (1987),

Bhatia, Sokol, Pernoll (1987), Wollschlaeger et al (1999), Parry et al (2000), Chauhan

et al (2000), Hadlock (1992) and Benson and Doubilet (1998) called it a weight of 4000

grams and above. Rouse and Owen (1999) used both a 4000 gram and 4500 gram cut

off to study prophylactic caesarean delivery for fetal macrosomia determined by

ultrasonic weight estimation. Neither weight cut off made a difference to the results

due to the inaccuracy of the weight estimations. Hadlock (1988) and Jeanty (2001)

claimed that the limitations of estimating fetal weight include the difficulty in obtaining

ultrasonic images at the correct imaging plane for measuring the abdominal

circumference. Benson and Doubilet (1998) agreed with Hadlock and went further by

saying that in the diabetic pregnancy, weight formulas using a combination of head,

abdomen and femur have a 95% confidence range of plus/minus 24%.                     More

importantly, according to Hadlock (1984), fetal weight charts are usually based on

fetuses with normal body proportions whilst fetuses with macrosomia have an increase

in body fat and thus there is a tendency to over estimate weight by up to 4%. This was

in agreement with Bernstein et al (1992), who concluded that fetal fat influenced the

ultrasound estimation of fetal weight in diabetic mothers. What was not taken into

account in all of these studies was ethnicity, maternal stature and genetics which can

strongly influence fetal size and birth weight (Brooke 1981, Wan & Woo 1984,Wilcox

Fetal Growth Assessment                                                       Background

et al 1993, Lai & Yeo 1995, de Jong 1998). Whilst a birth weight of 4000 grams may

be called macrosomic in one population, the potential birth complications due to

macrosomia may be present at a lower birth weight in a population with smaller

maternal stature.

       As mentioned earlier fetal macrosomia is associated with increased risks,

obstetric intervention, shoulder dystocia and post partum haemorrhage. Risk factors

include maternal obesity, increased maternal age, previous large baby or diabetes

(Spellacy 1985). Attempts have been made to detect/predict antepartum macrosomia in

order to reduce birth intervention and complications. Benson and Doubilet (1998)

assessed normal and abnormal fetal growth and determined that ultrasonic weight

estimation was less accurate in the larger fetus with a 51% positive predictive value for

LGA and 67% for macrosomia. Varner (1987) claims that the rate of macrosomia is

increased 4 fold by maternal obesity, 2.5 times by post term and gestational diabetes

mellitus and that up to 53% of babies with macrosomia could be identified by a

thorough obstetric history and physical examination, with a 10% false positive rate.

This relied on the ability to assess fundal height and was inaccurate if the fetus was in

any position except vertex. The method also did not allow for ethnic influence.

1.5    Fetal Growth and Gestational Diabetes Mellitus

Gestational diabetes mellitus can be associated with increased fetal growth and is

defined in the Report of the Expert Committee on Diagnosis and Classification of

Diabetes Mellitus. (1997) as: “any level of carbohydrate (glucose) intolerance with

onset or first recognition during pregnancy.”

Fetal Growth Assessment                                                     Background

       There is no differentiation between those women with pre-existing unrecognised

non-insulin dependent diabetes mellitus (NIDDM) and those who develop glucose

metabolism impairment during pregnancy. Fifty percent of women with gestational

diabetes mellitus (GDM) are at a high risk to develop type-two diabetes, or NIDDM,

within ten to twenty years of pregnancy. NIDDM accounts for up to 90% of all


       In Australia, according to Moses et al (1994), between 4 – 8% of all pregnancies

are affected by GDM with a 30 to 37% chance of recurrence in subsequent pregnancies.

Another Australian study by Foster-Powell and Cheung (1998) gave a 70% incidence

of GDM recurrence with increased maternal age and/or insulin therapy in both the

index and subsequent pregnancy being the strongest predictors for recurrence. Moses

(1994) listed the risk factors for GDM as including a body mass index greater than 30,

maternal age greater than 35 years, a previous baby with macrosomia, previous GDM,

glycosuria, hypertension, polyhydramnios and racial susceptibility. Fulcher, Gunton

and Chippendall (1998) described the implications, diagnosis and management of GDM

with the maternal risks of GDM being an incidence of caesarean delivery 1.5 times

greater than the normal population due to fetal size and risk of future development of

diabetes mellitus. Fetal risks include perinatal mortality (2.5x greater), prematurity,

large for dates (2x) and macrosomia (1.5x) and an increased rate of polycythaemia,

hypoglycaemia, neonatal jaundice, risk of future diabetes and obesity. Fulcher also

cites the controversy and variances between countries in diagnosing GDM. In Australia

most pregnant women in antenatal care are offered a 50 gram glucose challenge test

(GCT) around 28 weeks of gestation, which has reduced the risk of the undiagnosed

gestational diabetic proceeding through pregnancy without the opportunity of

Fetal Growth Assessment                                                      Background

monitoring and treating abnormal glucose levels. Those women with a level > 7.8mm/l

proceed to a 75gram glucose tolerance test, where a positive diagnosis of GDM is made

if 2 of the 3 levels are increased (Fasting >5.2mm/l, 1hr >9.9mm/l, 2hr >8.5mm/l).

This compares with North America, for instance, where a 100 gram, 3-hour GGT with

values measured at 0, 1, 2 and 3 hours is recommended with GDM diagnosed if any

two of these levels are raised. Langer (1987) has shown that women with only one

abnormal level have higher rates of pre-eclampsia and macrosomia than women with a

normal GTT and that the higher the two hour level, the higher the risk of macrosomia.

Women with GDM usually have no symptoms and most can self monitor their glucose

level, controlling it with moderate exercise, diet and/or insulin therapy. Bartha and

Martinez-Del-Fresno (2000) concluded that the earlier women are diagnosed with

GDM the higher the risk they are for problems such as hypertension and the need for

insulin therapy. A study by Bevier, Fischer and Jovanovic (1999) looked at how

dietary treatment of women with an abnormal glucose challenge test, but a normal

glucose tolerance test (GTT) reduced the incidence of macrosomia and the number of

caesarean deliveries.

       O‟Sullivan et al (1973) and Oats and Beischer (1986) suggested that undetected

GDM is responsible for an increase in perinatal mortality. Hoffman et al (1998) gave

the problems associated with increased morbidity with GDM as macrosomia,

respiratory distress, hypoglycaemia and hyperbilirubinaemia. The GDM management

guidelines of 1998 (Hoffman) stressed that a team approach to patient care with

antepartum fetal monitoring and dietary care was paramount to a successful outcome.

Individualising the diet of women with GDM depending on maternal height and body

mass index helps reduce the need for insulin therapy. Insulin therapy is not required if

Fetal Growth Assessment                                                      Background

fasting blood glucose can be kept < 5.5mmol/L, one hour post prandial < 8mmol/L and

<7mmol/L two hours post prandial but is recommended if blood glucose levels are

raised on two occasions during a one to two week period (Hoffman 1998). Langer et al

(1994) believed that self monitoring and insulin therapy, where required, reduced the

incidence of macrosomia and subsequent perinatal complications.

       Chung and Myrianthopoulos (1975) questioned the possibility of a genetic

factor in diabetes but found that women with pregnancy induced glucose intolerance

were at no higher risk of giving birth to infants with congenital abnormalities than

women with normal glucose tolerance. Dooley, Metzer and Cho (1991) found that

ethnicity had a significant independent effect on birth-weight of babies born to mothers

with diet controlled GDM with mean birth-weight being lowest in Asians and highest in

Hispanics. They also concluded that undiagnosed or untreated gestational diabetes can

be a major cause of fetal macrosomia. Cheung et al (1998) also assessed the high risk

of developing abnormal glucose tolerance in non-English speaking women with Asian

women having up to a 15% incidence of GDM. This agreed with the results of Yue et

al (1996) who showed Caucasians to have a 5% incidence of GDM compared with up

to 15% of Chinese pregnancies.

       Fetal macrosomia is 1.5 times greater in GDM affected pregnancies (Fulcher et

al, 1998) and can be secondary to raised insulin levels in the fetus as a result of

maternal hyperglycemia. Glucose is readily transferred across the placenta and so an

increase in maternal glucose can result in fetal hyperglycemia. This increase in glucose

level stimulates the fetal pancreas to raise its production and release of insulin and

according to Pedersen (1964) increases lipogenesis and the synthesis of glycogen and

Fetal Growth Assessment                                                        Background

protein. Shah (1990) claims that with fetal macrosomia the various organs grow at

different rates and as such the brain, being insulin-insensitive, doesn‟t display the

accelerated growth of the abdomen, which is increased in size due to fat deposition. As

a result there is a disproportional growth of the fetal abdominal organs compared with

the head and limbs. Shah also found the long bones to have an increased fat layer and

that a combination of these features assists in the ultrasonic determination of fetal

macrosomia. Mintz (1988) and Landon (1989) demonstrated the increase in growth of

the fetal abdomen and chest between 28 and 32 weeks gestation in GDM pregnancies.

Measuring of the fetal chest is not a routine ultrasound procedure and so the clinical

implications of this finding have not been tested in mainstream management. Doubilet

(1990) said that although between 8% and 10% of all babies in the general population

are born with macrosomia, the rate is between 10 and 15% in babies born to diabetic

mothers. The authors did not state whether these figures included both stable and

unstable diabetics and seem over estimated when compared with the figures stated by

Hadlock (1991) that only 2% of babies with macrosomia are born to diabetic mothers.

A study of fetal macrosomia by Wollschlaeger, Nieder, Koppe and Hartlein in 1999

compared both maternal and fetal complications directly related to fetal overgrowth in

non-diabetic, macrosomic deliveries with a control of normal deliveries. The group

showed the relationship between increased maternal age, parity and body mass index to

increased risk of macrosomia. Of the infants with macrosomia, 34% recorded high

insulin levels in umbilical blood indicating a possible failure to correctly identify some

pregnancies with glucose intolerance. The Wollschlaeger findings were similar to the

work of Dooley et al (1991), which supported the theory that undiagnosed or untreated

gestational diabetes can be a major cause of fetal macrosomia. Langer et al (1987)

concluded that women with only one raised level at the oral glucose tolerance test

Fetal Growth Assessment                                                           Background

(OGTT) have an increased risk of macrosomia compared with women with a normal

OGTT and Tallarigo (1986) showed that the higher the level for the abnormal OGGT

the higher the likelihood of an infant with macrosomia. Hardy (1999) concluded that

although non-obese women with unstable GDM may be at risk of having a macrosomic

infant, the greater risk group for macrosomia were the GDM women with high blood

glucose levels who were obese both before and during pregnancy. Combs, Gunderson

et al (1992) showed a strong relationship between fetal macrosomia and unstable

maternal postprandial glucose control whilst Naylor et al (1996) claimed that the risk of

macrosomia could be reduced by up to 65% with tight glycaemic control. Murata and

Martin (1973) were amongst the first groups to study the effect of diabetes on fetal size

and, by measuring the BPD of fetuses from 13 weeks to term, they determined that

there was no significant difference in head size between normal pregnancies and those

affected by diabetes. Hadlock (1991) observed that babies with macrosomia born to

diabetic mothers had asymmetric growth, with the abdomen being proportionately

larger due to increased fat deposition, than the rest of the body, compared with a

symmetric macrosomic baby of a non-diabetic mother. This highlights the benefit of

the head and abdominal circumference ratio when ultrasonically assessing fetal size in

late pregnancy. Benson and Doubilet (1998) found weight prediction using ultrasound

was less accurate with GDM, with the 95% confidence limit being +/-24% compared

with +/-15% in normal pregnancies.

       The 1998 consensus guidelines on management of GDM from the Australian

Diabetes in Pregnancy Society were implemented due to a „lack of quality randomised

controlled clinical trials in the area of gestational diabetes mellitus‟. This led to a group

of researchers establishing an Australian carbohydrate intolerance study in pregnant

Fetal Growth Assessment                                                      Background

women trial group (ACHOIS). The main aim of ACHOIS was to establish the effect of

treatment of GDM on pregnancy outcomes. One thousand pregnant women with a

gestation between 24 and 34 weeks and confirmed GDM were randomly assigned to

either routine antenatal care or to an intervention group. The latter group received

dietary advise, blood glucose monitoring and insulin therapy if required. The main

outcomes of the study (Crowther et al 2005) included a reduction in serious perinatal

complications such as dystocia and nerve palsy, from 4% in the control group to 1% in

the intervention group and a drop in LGA births from 22% to 13%.            The study

concluded that treatment of GDM led to a decrease in perinatal morbidity. The study,

although only published in June 2005, has already received praise from Greene and

Solomon (2005) who believed it was now time for universal treatment of GDM. An

editorial in the Australian Medical Journal (2005) cites that although universal

screening for GDM has merits, the costs involved in not only the initial screening but

also in providing the extra health care support required, such as diabetes education

staff, dieticians and clinicians, may make large scale implementation throughout

Australia difficult. HAPO study cooperative research group, the hyperglycemia and

adverse pregnancy (HAPO) study (2002) is another group with aims of establishing

diagnostic   thresholds   for   GDM   with   an   international   blinded   prospective

epidemiological study due to be reported on in mid 2007.

1.6    Ultrasonic Evaluation of Pregnancy

Diagnostic ultrasound refers to high frequency sound waves up to fifty megahertz. In

medical ultrasound imaging, the sound waves are mechanical disturbances generated by

a crystal in a hand held transducer. The crystal converts electrical energy into sound

Fetal Growth Assessment                                                      Background

energy using a pulse-echo technique and this sound wave travels through the body,

bouncing back from the different tissue interfaces to be converted back to electrical

energy.     The returned echo is mapped for position and intensity to build up an

anatomical image.

          Obstetric ultrasound has been in use since the late nineteen fifties when Ian

Donald and his team at the University of Glasgow‟s Department of Midwifery began

work on measuring the fetal head. In 1957 Donald and Brown, an engineer, developed

the worlds first compound contact scanner, which was a hand held system, followed in

1964 by the first commercial mechanical system.          In Australia in 1962 Kossoff,

Robinson, Jellins and Dadd developed a water bath compound scanner for use at the

Royal Hospital for Women in Sydney by Dr William Garrett (1966, 1970, 1971) for

examining pregnant women. This group of researchers were recognised internationally

as being at the leading edge of obstetric ultrasound.

          Donald and MacVicar (1962) first measured the growth of the fetal head with

ultrasound.     This milestone in obstetrics was followed by a steady stream of

investigators including Sunden (1964) and Thompson et al (1965), who investigated

abnormal fetal growth by measuring the biparietal diameter (BPD) with the use of

amplitude and brightness (A and B mode) techniques. Campbell (1969) also developed

a reproducible method for measurement of the BPD, which was also used in the works

of the Australian team of Garrett and Robinson (1971).

Insler (1967) used the BPD and trunk measurements to estimate fetal weight but this

was found by the Sydney group of Garrett, Kossoff and Robinson to be of little use.

Fetal Growth Assessment                                                        Background

Taking the fetal cadaver measurements of Scammon and Calkins (1929), Garret and his

team produced graphs for cross sectional areas of the head and abdomen. Garrett et al

(1971) proposed relating the:

        “head and chest sizes to a period of gestation and express the result in weeks

rather than attempt an estimation of weight.”

       From these early results, produced from ultrasounds performed on A mode and

bistable static B mode, technology advanced to more sophisticated grey scale B mode

scanners, real time and three dimensional ultrasound.

       Dr William Garrett, an obstetrician and Dr David Robinson, a scientist, were

two pioneers of Australian Ultrasound who stressed the importance of fetal


“ Knowledge of fetal size has two main applications in obstetric practice. The first is

to compare the size of a fetus of unknown gestational age with normal figures and so

obtain an estimate of the maturity of the fetus. The second application is to compare

the size of a fetus of known gestational age with known normals either as a single

reading to tell whether the fetus in question is larger or smaller than normal or, better,

as a series of readings. A series of readings is to be preferred since it not only checks

the accuracy of a single reading but also gives most useful information as to the rate of

growth of the fetus.” (1971)

Fetal Growth Assessment                                                         Background

       Compared with physical examination of the pregnant uterus ultrasound imaging

is the most reliable method for assessing and tracking fetal growth. The prenatal

questions answered by ultrasound include fetal number (singleton, twins, triplets etc),

correct dating which enables decisions for time of elective caesarian section, and the

detection of intra uterine growth restriction or fetal macrosomia in late pregnancy. Due

to the wider availability of obstetrical ultrasound it is now a routine component of

antenatal care. Twenty-five years ago in Australia ultrasound examinations were only

performed if a clinical problem existed and it was only twenty years ago that pregnancy

ultrasounds became the routine and offered to most patients in private practice and

women attending antenatal clinics run by district hospitals. Good quality portable real

time scanning systems are used routinely by most obstetricians, who scan their patients

at each antenatal visit if deemed necessary.       In 2002 the number of self-referred

obstetric ultrasound examinations exceeded 132,000 (HIC Statistics). With the success

of the Fetal Medicine Foundation‟s first trimester screening of the nuchal translucency,

there is now strong justification for routine scanning of obstetric patients at both twelve

and eighteen weeks of gestation (BMUS 1990). In 2002 according to HIC statistics in

Australia over 500,000 obstetric ultrasounds were performed. These figures include

only those services, which qualify for a Medicare benefit and do not include services

provided by hospital doctors to public patients in public hospitals, nor the self-referred

scans. The Australian Bureau of Statistics (ABS) showed the number of births in 2002

to be 246,300.

Fetal Growth Assessment                                                       Background

1.7    Safety of Ultrasound

The safety of ultrasound has often been questioned but after four decades of use on over

three generations of women, obstetrical ultrasound has been declared to be a low risk

examination. The numerous studies of both functional and morphologic ultrasonic

biological effects, seeking adverse effects to the mother or fetus, have failed to define

any significant problems. Johnson (1998) went further by saying that ultrasound is safe

when used in the correct medical situation and performed by trained and accredited


       Most ultrasound systems now incorporate pulsed-wave and colour Doppler,

which operate at higher intensities. Allan et al (2000) described the current output

displays as based on two indices, the mechanical index (MI) and thermal index (TI).

MI is an estimate of the potential for producing both mechanical and non-thermal

biological effects on tissue whilst the TI is an estimate of the potential for producing

thermal biological effects in bone and soft tissue.

The American Institute of Ultrasound in Medicine (AIUM) released this official

statement on clinical safety in March 1993:

“Although the possibility exists that such biological effects may be identified in the

future, current data indicate that the benefit to the patient of the prudent use of

diagnostic ultrasound outweigh the risks, if any, that may be present.”

Fetal Growth Assessment                                                       Background

The AIUM Consensus Statements on Diagnostic Ultrasound: Mechanical Bioeffects

February 2000 found that after more extensive research no adverse effects can yet be

found on the use of diagnostic ultrasound used in a safe environment for clinical use.

“Diagnostic ultrasound has proven to be a valuable tool in medical practice. An

excellent safety record exists in that, after decades of clinical use, there is no known

instance of human injury as a result of exposure to diagnostic ultrasound. Evidence

exists however, to indicate that at least a hypothetical risk for clinical diagnostic

ultrasound must be presumed.”

1.8    Variability in Gestational Age and Fetal Growth

“Regardless of the number of sonographic measurements that we employ in predicting

menstrual age, it is very important to remember that we are simply inferring age from

size and to understand the variability that can be associated with that estimate.”

Hadlock 1991

Hadlock‟s extensive works on fetal age and weight assessment began with publication

of fetal measurement charts for the BPD and femur length followed by the abdominal

and head circumference. He sites the importance of why, once menstrual age has been

established either by dates or early ultrasonic measurements, it should never be


“subsequent fetal measurements become an index of fetal growth rather than

menstrual age.”

Fetal Growth Assessment                                                       Background

Menstrual age, taken as the first day of the last menstrual period (LMP), is used

interchangeably with the term gestational age. Ultrasonic fetal measurement charts for

the various parameters are based on menstrual age / gestational age. Wenner and

Young (1974) reported 66% of women in their study had a reliable LMP compared

with 18% in the article by Herz and Sokol (1978). Geirson and Busby-Early (1991)

used a questionnaire to find that although 76% of respondents claimed a certain LMP,

only 48% had recorded the date. Graham and Sanders (1985) put the 95% confidence

range of an accurate LMP at plus or minus three to four weeks. Assignment of

gestational age based on LMP may cause problems with estimating the date of delivery.

Kramer (1988) investigated the validity of GA estimation by LMP in term, pre- term

and post- term pregnancies and found that one in four pre- term and seven in eight post-

term pregnancies were classified incorrectly. Mean fetal size may vary as a result of

early or late ovulation, which Matsumoto et al (1962) reported as occurring in

approximately 20% of the female population. Hadlock (1991) also found that women

with the same LMP can ovulate at different times, thus giving a variation of fetal age of

plus or minus 1 week. Rossavik and Fishburne (1989) compared pregnancies of known

conception date with pregnancies dated from the LMP and reported that patients with a

certain LMP were usually certain of conception date.         Vaginal bleeding in early

pregnancy can also make dates inaccurate (Campbell et al 1985).

       The accuracy of dating pregnancies by a certain LMP compared with ultrasound

dating was tested by Campbell, Warsof and Little (1985), Waldenstrom and Axelessonl

(1990) and Tunon and Eik-Nes et al (1996) who all found that a BPD between fourteen

and twenty weeks, was more reliable than a certain LMP with Waldenstrom et al (1990)

finding that the dates given by the early ultrasound measurements disagreed with the

Fetal Growth Assessment                                                        Background

LMP in 20% of patients. This was also found to be the case with Bowie and Andreotti

(1983) who reviewed the different clinical estimators of gestational age (Table 1).

        Pedersen (1983) showed that the crown-rump length measured between eight

and twelve weeks gestation was a more reliable indicator of fetal age than the head

measurements but that from twelve weeks onwards, fetal flexion increased the chances

of error. This result was in concurrence with Benson and Doubilet (1998) who also

claimed that the accuracy of ultrasonic gestational age determination was +/- 0.5 weeks

throughout the first trimester. Bovicelli et al (1981) and Selbing (1986) thought that the

head measurements were more accurate than either the abdomen or femur

measurements from 11 weeks gestation.

        Persson (1986), Rossavik et al (1989) and Hadlock and Martinez-Poyer (1992)

all demonstrated how the uncertainty in predicting gestational age increases with

advancing weeks of pregnancy, particularly from the late second trimester. Hadlock

(1991) found that for up to twenty weeks gestation, ultrasound reflects gestational age

with an accuracy where two standard deviations (SD) are equal to plus or minus one

week, however in the late third trimester this variability is reported to be up to plus or

minus three and a half weeks. This was in agreement with the Australian studies by de

Crespigny and Speirs (1989) and Westerway (2000). Hadlock (1991) questioned the

validity of Kurtz and Wapner (1980) who reported a plus or minus two-week variability

in the third trimester but suggests that this difference resulted from working with mean

figures from a number of studies instead of raw data, which would have increased the


Fetal Growth Assessment                                                 Background

Table 1/1

Clinical Parameters in Estimation of Gestational Age

Priority for Estimating Gestational Age 95% Confidence

1    Invitro fertilization                                         Less than 1 day
2    Ovulation induction                                           3 – 4 days
3    Recorded basal body temperature                               4 – 5 days
4    Ultrasound crown–rump length                                  +/- 7 days
5    First-trimester physical examination                          +/- 1 week
6    Ultrasound BPD prior to 20 weeks                              +/- 1 week
7    Ultrasound gestational sac volume                             +/- 1.5 weeks
8    Ultrasound BPD from 20 – 26 weeks                             +/- 1.6 weeks
9    LMP from recorded dates (good history)                        +/- 2 – 3 weeks
10   Ultrasound BPD 26 – 30 weeks                                  +/- 2 – 3 weeks
11   LMP from memory (good history)                                3 – 4 weeks
12   Ultrasound BPD after 30 weeks                                 3 – 4 weeks

Adapted from James Bowie (1983) by Hadlock F (1991)

“It is recommended to always use a more reliable indicator in preference to a less

reliable one.” Hadlock 1991

Fetal Growth Assessment                                                      Background

          One of the major causes of incorrect assignment of gestational age is the

ultrasonic imaging plane used to measure the fetal parameter and the placement of the

calipers. This incorrect measurement of fetal parameters can lead to errors as stated by

Parvey (1990):

“Fetal biometry must be accomplished in a consistent and reproducible manner to

ensure that any differences observed are real and not just procedural artefacts.”

Jeanty (2001), Dudley and Chapman (2002) and Hadlock (1991) all stressed the

importance of good technique with Hadlock saying:

“one must be certain to duplicate the technique of the investigator whose data one is


Correct cursor placement is critical for accurate measurements. Some of the early

ultrasound systems used a very crude form of callipers, which could only measure in

increments of five millimeters. Modern systems measure to an accuracy of half a

millimetre and combined with the high resolution of the scanning equipment, the ease

of measuring fetal parameters has been greatly improved.           The fetal head and

abdominal circumference charts formulated by both Hadlock (1982) and Deter (1982)

used a digitised map measurer on a polaroid image for measuring circumference. The

possibilities for error were enormous and yet their charts were in common use in

Australia until the Australasian Society for Ultrasound in Medicine (ASUM)

recommended the implementation of new charts in 2001 based on an Australian


Fetal Growth Assessment                                                     Background

        Merz (1987) questioned whether the differences in the mean femur length seen

in the second trimester were due to inter-observer error. This was highly probable as

Lawson and Albarcelli (1976) showed the inter-observer error in 14% of patients to be

statistically significant, whilst Docker and Settetree (1977), and Davidson and Lind

(1973) yielded inter-observer errors up to 30%. Deter, Harrist and Hadlock (1982)

showed in a multi-factorial analysis of variance that inter-observer errors of up to 8%

are possible. Gull et al (2002) calculated a three way inter-observer error for fetal

weight estimation at 6.1% and an intra-observer error of 1.5%. Hadlock (1982) had an

intra-observer error of 2% when he measured the same images a second time. Graham

and Sanders (1991) attributed the differences in early charts of European researchers

when compared with British and American figures, to ultrasound systems being

calibrated for a characteristic velocity of 1600m/s rather than the now accepted

1540m/s. This yielded a 4% difference or error in measurements. This error factor was

highlighted by Pedersen (1983) who studied the crown-rump length / BPD relationship

on a B-mode system set at 1556m/sec and adjusted all of his measurements by 0.5 to

1% to compensate for the difference.

1.9     Ultrasonic Determination of Macrosomia

Wladimiroff, Bloesma and Wallenburg (1978) recommended using ultrasound to

measure the fetal chest as it is:

“significantly influenced by accelerated growth of the fetal liver and therefore

reflective of macrosomia.”

Fetal Growth Assessment                                                      Background

They claimed a 47% detection rate for macrosomia using the 95th percentile.

Houchang, Modanlon, Komatsu et al (1982) also used a chest circumference and head

ratio to predict macrosomia and subsequent birth trauma. Schlater (1987) suggests that

a difference of 16mms in the head and chest circumference identifies macrosomia but

for this technique to be successful correct imaging planes for measurement are essential

as angulation can increase the circumference and thus the ratio. Chauhan et al (2000)

assessed the usefulness of weight estimations for macrosomia based on traditional

sonographic measurements of head and abdominal circumference and femur length.

The group also incorporated new ultrasonic measurements of fetal soft tissue such as

upper arm and subcutaneous tissue of thigh and upper arm, cheek-to-cheek diameter

and a ratio of thigh subcutaneous tissue and femur length in the hope of identifying the

fetus with macrosomia. The subcutaneous tissue measurements of the upper arm and

thigh and the subsequent ratio with the femur length proved to be poor predictors of

fetal overgrowth, however the cheek-to-cheek diameter and upper arm soft tissue were

equivalent to clinical evaluations.    In clinical practice inter and intra observer

reproducibility could be questioned as it may be difficult to measure fetal variables

such as the subcutaneous tissue due to interlayer edge perception as opposed to well-

defined end points as seen with long bones and fetal skull.   Parry, Severs and Morgan

et al (2000) retrospectively assessed the rate of caesarean deliveries as a result of

incorrect ultrasonic prediction of macrosomia. There was a 42.3% false positive rate

compared with 24.3% correctly predicted as not macrosomic. Data was reviewed from

two centres but there was no mention by the authors as to what fetal weight formulas

were used, or whether both centres, and the varying ultrasound machines, used the

same formulas. This oversight may have biased the results. A similar study by Rouse

and Owen (1999) for prophylactic cesarean delivery for fetal macrosomia diagnosed by

Fetal Growth Assessment                                                        Background

means of ultrasonography showed that from 100 pregnancies scanned, ultrasound

predicted 16 to be macrosomic, of which only 7 were actually over 4000 grams. Again

fetal weight formulae were not assessed. Conway and Langer (1998), like Rouse and

Owen, assessed the increased rate of elective caesarean delivery in predicted

macrosomics to reduce the rate of possible shoulder dystocia, concluding that the

increased cost associated with the procedure was not warranted.            These studies

highlight the difference in accurate fetal weight prediction between the claims of the

authors of the fetal weight formulas such as Hadlock (1984), Deter (1982) and

Campbell and Wilkin (1975) and the actual results in clinical use. This is where quality

control may be of importance as Dudley and Chapman (2002) found the variations in

imaging planes for the abdominal circumference led to significant differences in weight


       Baker et al (1994) was one of the first groups to try to estimate fetal weight with

magnetic resonance imaging (MRI) but the accuracy was questionable. Uotilia et al

(2000) however, using more modern equipment, utilised (MRI) to more accurately

predict fetal macrosomia with a correlation to actual birth-weight of 95% with MRI

compared with 77% of predicted weight by ultrasound. The ultrasonic result of the

Uotilia study compares with Benson and Doubilet (2000) who claimed that ultrasonic

assessment of fetal weight had a positive predictive rate of up to 67% for macrosomia.

Sabbagha, Ogata and Metzger (1980) used serial ultrasounds in the assessment of

evolving macrosomia and suggested that the abdominal circumference was the fetal

measuring parameter of choice for predicting macrosomia, backed up by Hadlock

(1985) and Tamura (1986). Buchanan et al (1994) also suggested the use of ultrasound

measurement of the abdominal circumference for the detection of the large for

Fetal Growth Assessment                                                        Background

gestational age and macrosomic fetus and determined that at between 29 and 33 weeks

gestation an abdominal circumference greater than the 70th percentile was associated

with a large for gestational age rate of 37% compared with 11% with an abdominal

circumference below the 70th percentile.     Owen et al (2002) aimed to establish a

relationship between amniotic fluid index, measured by ultrasound, and fetal weight but

concluded that there was no clinically relevant correlation between the two.

1.10   Fetal Size/Growth and Ethnicity

Garrett and Robinson added this comment on the 1989 growth charts produced for the

Royal Hospital for Women in Sydney, Australia:

“From twenty five weeks to term, measurements do not necessarily relate to the period

of gestation but simply give a comparison to the mean values for the population at

large. The variation among normal fetuses is considerable, and in the third trimester a

variation of three weeks size on either side of the mean value may be seen.”

Homko, Sivan, Nyirjesy and Reece (1995) looked at the interrelationship between

ethnicity and gestational diabetes in fetal macrosomia concluding that:

“ethnic variation in fetal growth may be due to varying influences of in-utero growth

promoters among these populations as well as underlying genetic factors.”

Fetal Growth Assessment                                                      Background

Rodrigues, Robinson and Cramer (2000) found the rate of macrosomia in Cree Indians

to be over 34%, compared with 11% among non-Natives, with this difference in fetal

growth possibly due to genetics. Deter (1981) believed that it was necessary to identify

these small subgroups with different growth characteristics, within a population where

growth patterns are relatively consistent. Although studies in Canada, Britain and the

USA have shown no significant difference in fetal growth between different ethnic

groups who reside in the same geographical regions (Parker 1982, Sabbagha 1976), a

study by Westerway in Australia in 1997 showed that fetuses of Vietnamese immigrant

women living in Western Sydney were smaller than Caucasian women in the last five

weeks of gestation.     This difference in ultrasonic growth was reflected in the

differences in birth weight between the Caucasian and Vietnamese babies seen in the

Western Sydney Health birth statistics in 1996.

       The majority of early studies in fetal measurements such as Robinson (1975),

Hadlock (1982), Deter (1982), Jeanty (1984) and Merz et al (1987) used middle class

white populations. This led some authors to question whether standard ultrasonic fetal

measurement charts can accurately predict gestational age of fetuses from different

ethnic populations.    As far back as 1976 the feasibility of using general fetal

measurement charts was being investigated, with Sabbagha and Hughey (1976)

comparing the BPD of fetuses of both white and black Americans and showing no

significant difference between the two groups. Hadlock, Harrist and Shah (1987) also

compared the various ultrasonic fetal parameters between white and black American

women, finding that there were no statistically significant differences in measurements

between the groups. In 1990 Hadlock et al asked if the current data on fetal growth

standards were applicable to a racially mixed, indigent population.          After first

Fetal Growth Assessment                                                        Background

determining that there was no clinically significant difference in predicted values of the

biparietal diameter (BPD), head and abdominal circumferences (HC and AC) and

femur length between Negros and Hispanics living in the USA, they then compared the

racially mixed group with a white middle class population with the average differences

across the whole gestational age range being BPD 0.7mm, FL 0.6mm, HC and AC

3mm which were statistically of no significance. It is interesting to note the much

higher AC difference when compared with the head and long bones.

       It was suggested by Spencer et al (1995) that it was inappropriate to use

standard charts for abdominal circumference and estimated fetal weight for the Asian

population due to their lower body mass index. Thompson (1968) showed that that the

stature of the mother affects birth weight whilst Gardosi et al (1992) contradicted this,

finding that ethnicity affected birth weight but that stature did not.       Findings by

Spencer, Chang, Robson and Gallivan (1995) based on their longitudinal investigation

of 20 Bangladeshi pregnancies, concluded that the lower birth weight seen in this

population was a result of the lighter and shorter stature of the Bangladeshi women as

opposed to the Anglo-Saxons. Even though there was a difference in birth weights the

fetuses from both groups grew at a similar rate during the last trimester and so Spencer

concluded that the use of ethnic specific growth charts would not be warranted for the

recognition of IUGR. Naey and Tafari (1985) documented the effect of maternal

nutrition on the fetus, finding that Ethiopian women of the same height, but with higher

skinfold thicknesses, had heavier babies than women with a thinner skinfold. This

supports the claim that maternal diet, lifestyle and physique are all contributing factors

to the well being of the fetus (Naey and Tafari1985). Arenson and Chan (1995)

investigated the value of standardized gestational age charts for fetuses of first

Fetal Growth Assessment                                                      Background

generation oriental immigrants to Canada, determining that it was acceptable to use the

existing growth charts as there were no significant differences in the BPD, HC, AC and

FL measurements between the oriental and non-oriental groups. This was in agreement

with the earlier work of Ruvolo, Filly and Callen (1987) who compared femur lengths

in a white and oriental population, finding no significant difference.

          Meire and Farrant (1981) deduced that fetuses of East Indian women residing in

Britain had a significantly lower mean abdominal circumference after twenty-four

weeks gestation, and a lower birth weight than white European fetuses of the same

gestation, but that the BPD measurements in both groups was the same. This was

contrary to the study by Okupe and Coker (1984) who found that the BPD of term

fetuses of Nigerian women tended to be in the upper range of the standard curves.

Jacquemyn, Sys and Verdonk (2000) showed a statistically significant difference in the

head circumference (p=0.017), abdominal circumference (p=0.0015), femur length

(p=0.0014) and estimated fetal weight between Belgian, Turkish and Moroccan women

living in Belgium, however there was no difference seen in the BPD. The important

difference in this study was that the partners of the women recruited had to be of the

same ethnic origin. In Australia, Westerway (1997) also reported a lower birth weight

in an Asian community, which was thought to be due to the smaller stature of the

women where statistically significant differences were seen in BPD, OFD and FL in the

last five weeks of gestation when compared with both Royal Hospital for Women

charts by Garrett and Robinson (1987), and charts by de Crespigny and Robinson

(1990).     The latter two charts were based on a mainly middle class Caucasian

population.     Lai and Yeo (1995) produced reference charts for BPD, head and

abdominal circumference and femur length for Asians residing in Singapore, which

Fetal Growth Assessment                                                       Background

showed minor differences for all but the abdominal circumference, which had a larger

discrepancy, indicating that if available, local charts should be utilised for fetal

measuring. This was also the findings of Wan and Woo (1984), who tried to estimate

fetal weight in a Chinese population using the abdominal circumference and head

measurements and then correlate them with Caucasian weight charts.

       As has been shown, the controversy regarding population specific fetal

measurement graphs continues.      Dunn (1985) suggested an international reference

range for fetal size, regardless of ethnicity, but with an awareness of differing growth

patterns in the last five weeks of gestation. This appears to be the general consensus of

authors stretching through two decades.

1.11    Measurements of Fetal Parameters for Growth/Weight Assessment.

1.11.1 The BPD, OFD and Head Circumference

Historically the BPD was the first parameter used to assess gestational age by Donald

(1962) and later Campbell (1969) who produced charts using A mode systems. BPD

charts, using B-mode, were produced by Levi et al (1973), Weiner et al (1977),

Sabbagha and Hughey (1978), Kurtz and Kapner (1980), Hadlock et al (1982), Shepard

and Filly (1982) Campbell, Warsof and Little (1985) and O‟Brien and Queenan (1985).

O‟Brien and Queenan (1985) counted that by the early 1980‟s there were over twenty-

five BPD charts in use internationally.

       The plane of section of the fetal head at which the BPD, OFD and head

circumference is measured is the level of the cavum septi pellucidi and the thalamic

Fetal Growth Assessment                                                     Background

nuclei (Benson and Doubilet - 1991). This was the image plane put forward by the

American College of Obstetricians and Gynaecologists (ACOG) and accepted as an

international standard in 1991.     Until this time there had not been international

uniformity in measuring planes. Kurtz et al (1980) found that, due to the wide normal

range at any stage of gestation, it made no significant difference whether the BPD was

measured from the outer edges of the parietal bones, middle to middle, or outer edge of

proximal to inner edge of distal parietal bone. de Crespigny, Garret and Robinson

(1989) prepared the first ASUM Standard BPD chart based on a middle class Australian

population and stated on the chart that:

“The wide normal range of BPD in late pregnancy must be appreciated. It is not

expected that the BPD be used to assess gestation late in pregnancy. The values from

thirty three weeks are intended to predict the growth in fetal head size from a known


       Researchers have speculated on the reasons for the differences seen in the

various fetal measurement charts. Deter et al (1981) believed that the significant

variations in the BPD growth curves are seen in both longitudinal and cross sectional

studies but that to identify any small subgroups with a different growth pattern,

longitudinal studies are necessary. Sabbagha and Hughey (1978) analysed four BPD

studies from different authors and found that the combined mean of the groups showed

no variance from each individual group. Kurtz and Wapner (1980) hypothesized that

variances in populations could account for the differences but after comparing

seventeen charts found no significant variation due to population. Deter (1981) also

compared the average values of seventeen BPD tables and found similar values in

Fetal Growth Assessment                                                    Background

thirteen of the studies. The table values conceived by Campbell et al (1977), Hobbins

(1979), Schneider et al and Weiner (1977) in the other four studies were consistently

lower, but no reasons for these differences were given. Whereas most charts showed a

general relationship of BPD to gestational age (GA) Graham and Sanders (1991) found

a discrepancy in the Yale BPD chart by Hobbins (1979) where the GA assigned to a

BPD of 80mm was 33 weeks, the chart of Weiner (1977) gave 31 weeks.

       Head shape also affects the BPD with flattened, elongated heads, as seen in

dolichocephaly, decreasing the BPD (Benson, 1991) who also claimed that fetuses in a

breech position often had a smaller BPD than the BPD taken from a cephalic lie fetus

of the same gestation. A dolichocephalic head shape is often seen in patients with

premature rupture of membranes (PROM) and as such the BPD is an unreliable

measurement (Wolfson and Zador 1983). This problem could be overcome, according

to O‟Keefe et al (1985) by using long bone lengths and/or head circumference to assess

fetal size, as long bones cannot be pushed out of shape with fetal position and a head

circumference based on an OFD/BPD measurement will compensate for the head

shape. Doubilet and Greenes (1984) also suggested the use of the head circumference,

as a means of correcting the BPD for shape. This measurement was called an area

corrected BPD (BPDa), and could be derived from the BPD and the OFD with the

resulting measurement read off the selected BPD chart to determine gestational age.

Another method to overcome the problem of head shape was derived by Hadlock et al

(1984) who assessed gestational age based on the head perimeter or circumference

which equals the sum of the BPD and OFD multiplied by 1.57. Jeanty et al (1991) used

a similar method but allowed for the ovoid shape by multiplying by 1.62 instead of

Fetal Growth Assessment                                                        Background

1.57.   Post measuring computations have largely been eliminated by the advanced

measuring packages available on modern ultrasound systems.

        Published studies involving the OFD are not as numerous as the BPD. The

OFD results obtained by Levi (1973) were significantly different when compared with

the figures of Jordaan (1983) and Deter (1981). Campbell (1975) thought that the head

images used by Levi et al (1973) to take these measurements were not the accepted

BPD plane and as a result, the measurements done would not have been a true OFD.

Deter (1981), Hansmann (1986), Merz (1988) and Romero (1988) showed the

importance of the OFD, as an adjunct to other parameters particularly in the third

trimester when head shape can affect size assignment.

1.11.2 Abdominal Circumference

The abdominal circumference measurement has the widest range of error of any of the

fetal parameters used for assessing age/growth and yet is one of the most important

measurements to use when assessing fetal weight due to the abdominal fat layer. It is

this fat layer that is decreased in the growth restricted fetus and increased in the fetus

with overgrowth. The accepted imaging plane for the abdomen according to Deter

(1984) should be a true transverse cut at the level of the fetal liver and stomach,

including the left portal vein at the umbilical region. Campbell and Thoms (1977) were

the first investigators to realise the importance of the AC measurement for fetal growth

by incorporating it in to a ratio with the fetal head. In 1990 the British Institute of

Radiology (BIR) analysed available fetal measurement charts for suitability for use.

Methodology and mathematical modelling played a critical role in the evaluation with

Fetal Growth Assessment                                                       Background

the final recommendation that the chart of Deter, Harrist and Hadlock (1982) be used

for abdominal circumference based on the average normal range of +/-13% of the mean

throughout pregnancy. Hadlock, Harrist and Deter produced another AC chart in 1982

but this was not recommended by the BIR due to the use of a constant standard

deviation for the entire gestation.    Hadlock (1985) found the intra/inter observer

variability of the abdominal circumference to be the greatest of all the fetal parameters

as it is the most important fetal parameter for inclusion in a weight formula and yet is

the most difficult to obtain the correct imaging plane for accurate measurement in late

pregnancy due to fetal lie. Dudley (1995) agreed with this finding, as did Benson and

Doubilet (1995).     Hadlock compared the percentile range of three groups of

investigators, Hadlock et al (1984), Jeanty et al (1984) and Tamura et al (1980) and

noticed that the AC varied by up to 8cms in the third trimester for the 90 th percentile.

Jeanty‟s 95th percentile was consistent with Hadlocks 90th percentile for the third

trimester.    The Jeanty and Hadlock studies consisted of middle class Caucasian

pregnant women whilst Tamura had a cross section of racial groups.             Although

population difference may have been an issue the most likely cause for the

discrepancies would be imaging plane, measurement methodology or mathematical

modelling of the charts. Dudley and Chapman (2002) and Gull et al (2002) found the

imaging plane of the abdomen as well as the measuring of the abdominal circumference

to be major causes for error, particularly in late pregnancy. Even acknowledging the

difficulties of consistent abdominal circumference measuring Campbell and Wilkin

(1975) were the first to use only this measurement in a fetal weight formula. All other

fetal weight formulae combine the abdominal circumference with other fetal


Fetal Growth Assessment                                                      Background

1.11.3 The Femur and Humerus

Long bones are relatively easy measurements to obtain throughout pregnancy and are

reliable parameters of fetal growth (O‟Brien and Queenan, 1985) especially when the

lie of the fetus precludes an accurate head measurement. Merz et al (1987) compared

prenatal measurements of ossified femoral diaphyses on aborted fetuses with postnatal

x-rays and reported only 0.7mm or 3% difference. Hill and Guzick (1992) found the

femur to be the most accurate of all of the parameters, whilst the results of Ott (1985)

found the femur to be the least accurate. The femur length has a linear relationship to

the crown-heel length and is therefore an important parameter to add to any fetal weight

formula as it makes allowances for population differences. Jeanty et al (1981), O‟Brien

et al (1981) and Hohler et al (1982) showed a linear relationship between the femur

length and the BPD throughout pregnancy. Hohler et al (1982) derived a FL/BPD ratio

as another means of assessing fetal size/growth. The possible causes for ratios being

either too high or too low (Graham and Sanders 1985) include images incorrectly

measured, inaccurate image planes and fetal abnormalities. Hadlock et al (1984) also

looked at incorporating the femur in a ratio with the head circumference where, if the

ratio was high, the head measurement was not used for assessing age/growth, and if the

ratio was low, the femur length was rejected.       Hadlock (1990) found significant

differences in variability in femur lengths, measured from mid second trimester to term,

seen between the various investigators who studied femur lengths. Jeanty and Rodesch

(1984) calculated a variability of plus or minus 2 weeks from 20 weeks to term but

Hadlock (1990) believed this result was due to a statistical error although gave no

reason to back the claim. The observed variability, backed up by the study of Benson

et al (1991), was around plus or minus 3.5 weeks at term. O‟Brien et al (1981) studied

Fetal Growth Assessment                                                     Background

over 1000 femur lengths and found that the confidence limits increased from 6 days at

20 weeks gestation to 18 days from 31 weeks to term. Varying statistical methods

could also be the cause for opposing views on the accuracy of the femur length

measurements in determining gestational age.

        Fetal humeral length measurement charts are not as common as femur charts

(Seeds, Cefalo, 1982) and the published data have been from small series, with only a

few using a cross sectional investigation. The most popular is that produced by Jeanty

and Rodesch (1984) and Merz (1988).             In 1992 Benacerraf and Neuberg linked

shortened long bones, in particular the humerus, with chromosomal abnormalities with

a short humerus was present in fifty-three percent of fetuses with Down syndrome, with

a false positive rate of just under six- percent.

1.11.4 Fetal Weight Assessment

Ultrasonic assessment of fetal weight relies on accurate measuring of the fetal

parameters, described previously, that are to be used in the formula. A study by

Townsend et al (1988) showed that correct measurements of parameters at the correct

imaging plane gave a more accurate weight than using measurements from suboptimal

images. This was in agreement with Dudley (1995). The accuracy of any ultrasonic

fetal weight formula is dependent on its 95% confidence limit. For example, one of the

more popular and accurate formulas is that of Hadlock (1983), which incorporates

head, abdomen and femur measurements. The 95% confidence limit for this formula is

+/- 15% and so it is expected that 95% of all births will be within 15% of the actual

birth weight. Benson (1998) claims that the more parameters included in the weight

Fetal Growth Assessment                                                         Background

formula, the more accurate they are, so the formula by Hadlock et al (1990) which used

only the abdomen has a 95% confidence limit of +/-22%, which is too great for clinical

use. Hadlock then combined the head and abdomen resulting in +/-18%, abdomen and

femur +/- 16% and head, abdomen and femur (1985) +/-15%. Doubilet (1995) claims

there is a distinction between weight prediction in diabetic and non diabetic

pregnancies with the difference in the 95% confidence limit for formulas that utilise

head, abdomen and femur being +/-24% for diabetic compared with +/-15% in normal


       One of the earliest fetal weight charts was that of Campbell and Wilkins (1975),

which utilised only the fetal abdominal circumference measurement in the equation.

The error factor was the reliance of a digital map measurer to trace the circumference

from a polaroid image similar to the method used by Hadlock and Deter for their

respective charts in 1982 and 1984. Although many other charts are readily available

this weight formula is still available as a pull down chart on most of the state of the art

ultrasound systems being installed around the world. Many authors have questioned

the accuracy of fetal weight estimations, particularly for prediction of macrosomia. A

retrospective study by Parry et al (2000) showed a 42% false positive rate for predicting

macrosomia with ultrasound as for birth-weights between 3500g and 4000g only 24%

of pregnancies were correctly identified as being non-macrosomic.                 Another

retrospective study by Rouse and Owen (1999) showed that ultrasound of 100

pregnancies identified 16 macrosomic fetuses of which only 7 were actually greater

than 4000 grams at birth. The variations seen in fetal weight estimations on the same

fetus but performed on different ultrasound systems could be due to inter-observer error

or, more likely, different weight formulae in the ultrasound systems. Gull et al (2002)

Fetal Growth Assessment                                                         Background

sited intra and inter-observer errors as being the main cause of differences seen in fetal

weight estimations. Most fetal weight formulae are based on the presumption of

normal fetal growth and do not take in to consideration the factors attributing to either a

growth restricted or macrosomic fetus. The differing abdominal circumference of a

normal fetus and one affected by gestational diabetes is a prime example.

       It is important to use the fetal weight formula with the lowest random error

through the range of weights. De Jong et al (1985) thought that the Hadlock formula,

which used the abdominal circumference and femur length, was the best fetal weight

formula to use. Hadlock however found the mean deviation from actual birth weight to

be 0.3% for all his formulas but the standard deviations varied with the AC/FL being

the least accurate at +/- 8.2%, the BPD/AC/FL +/- 7.7%, the HC/AC/FL +/- 7.6% and

BPD/HC/AC/FL +/- 7.5%. When divided into weight ranges the accuracy for each

formula within each group changed and this led Hadlock to advocate the use of the

HC/AC/FL formula.

1.12   Importance of Ultrasound Assessment in the Third Trimester of Pregnancy

Third trimester ultrasound is used clinically in selected at risk pregnancies which are

scanned to assess fetal size and well being.        Of particular importance are those

pregnancies complicated with gestational diabetes, maternal hypertension or intra

uterine growth restriction/overgrowth.      Fetal size is determined ultrasonically by

measuring fetal parameters, such as the head and abdominal circumference and long

bones, then comparing the measurements with charts, which relate the observed value

of a fetal parameter to gestational age. By utilising these figures in a formula an

Fetal Growth Assessment                                                       Background

estimation of fetal weight can be obtained which can then be compared with the

obstetricians‟ estimation and may help determine the time and mode of delivery.

Obtaining the correct imaging planes in the third trimester is more difficult than in the

second trimester (Dudley and Chapman 2002). Benson and Doubilet (1998) noted the

difficulty of accurate measuring when they created third trimester fetal weight

percentiles. Hadlock et al (1985) was also of this opinion, claiming the abdominal

circumference was the most difficult parameter to image in the correct measuring plane.

In 1991 Hadlock observed that babies born to diabetic mothers had asymmetric growth

in the third trimester with the abdomen being larger due to fat deposition than the rest

of the body. This discrepancy can be identified by observing the ultrasonic head /

abdominal ratio.   Accuracy of ultrasonic fetal measuring is essential in the third

trimester (Gull 2002, Dudley and Chapman 2002) if errors are to be avoided that may

affect the decision making for mode and timing of delivery. Inter and intra sonographer

measuring accuracy has been investigated by Deter and Hadlock (1982), Chang et al

(1993), Dudley and Chapman (2002) and Gull et al (2002) all concluding that the

correct image plane is the most important factor when estimating fetal growth in the

third trimester.

1.12.1 The Placenta and Umbilical Cord Doppler

Doppler ultrasound has become a useful adjunct to the obstetric examination with the

ability to assess the utero-placental circulation in late pregnancy. Ultrasound is the

only non-invasive method of tracking placental size, shape, echotexture and

retroplacental area (Spirt and Gordon 1998) and in the detection of placental creta and

abruption (Nyberg et al 1987).      From 33 weeks gestation placental calcification

Fetal Growth Assessment                                                          Background

increases exponentially with gestational age, and can be readily identified with

ultrasound, but this calcification does not appear to be of any clinical significance (Spirt

1984, Fox 1987).      Post partum, ultrasound can also detect retained products of

conception. Prior to the development of Doppler ultrasound it was not possible to

safely evaluate uteroplacental blood flow.         Umbilical artery Doppler measures

resistance to flow within the placenta. A progressive increase in diastolic flow velocity

with increasing gestation is considered normal and is believed to reflect the progressive

decreasing placental resistance. Diastolic flow velocity is determined by analysis of the

Doppler spectrum with the most commonly used indices being the systolic/diastolic

ratio (S/D), resistive index (RI) and pulsatility index (PI), all of which correlate well

with normal pregnancy.       The S/D and PI show similar predictive values when

correlated against adverse perinatal events with an S/D of 3 after 30 weeks gestation

considered abnormal. Factors affecting flow waveform include maternal position, fetal

heart rate, fetal breathing and fetal compression of the cord, all of which can be

overcome by repositioning the mother and waiting for ceasation of extreme fetal

movement. Concalves et al (2001) found that raised Doppler values in uterine arteries

indicated resistance to blood flow in the uteroplacental circulation.         This was in

agreement with Campbell et al (1983), Trudinger (1985), Adamson et al (1989) and

Olofsson et al (1993). Trudinger (1991) said that flow velocity waveforms (FVW‟s)

could be used in the management of women with GDM where waveforms remained

within normal limits for the fetus that was continuing to grow but that a high resistance

pattern in the FVW could be an indicator of growth cessation. Bracero et al (1988) also

showed a relationship between high FVW‟s and adverse outcomes in diabetic


Fetal Growth Assessment                                                       Background

1.13   Mathematical Modelling

The techniques used to analyse collected data are diverse with some methods more

accurate than others. (Evans 1988). Deter et al (1981) believed that the construction of

true fetal growth charts is best done with data collected in a longitudinal study. Jeanty

(1991) concurred with Deter et al (1981), saying that statistical analysis of a

longitudinal study gives a stronger curve fit. Hadlock (1990) however, preferred a

cross sectional study when collecting a large amount of data, and defined a properly

designed study as being one that measures a large number of fetuses evenly distributed

over the entire range of gestational age. Deter (1992) said that cross sectional analysis

assumes there is a common growth process subject only to random variation and the

main advantage is that adequate data can be collected in a shorter period of time.

Jeanty (1991) thought the main advantage of cross sectional fetal growth studies was

that a large number of measurements can be collected in a relatively short time.

       It is interesting to note the application of regression analysis by the various

researchers. The majority of fetal measurement charts were constructed by medical

imaging physicians, and yet statistical analysis in itself is a specialised field. Jeanty

(2001) gave the most thorough description of chart formation and mathematical

modelling but it was Altman (1991) who should be credited for highlighting the

mistakes made by the physicians in the preceding years. Merz (1987) and Hadlock et al

(1984) both stressed the importance of correct adaptation of regression analysis but

Merz and Grupner (1989) were the two who showed how, by using polynomials of an

appropriate degree, an adequately smooth curve fit can be obtained. The higher the

order polynomial the more the series will reproduce the original data. Hadlock (1984)

Fetal Growth Assessment                                                        Background

considered it inappropriate to use a polynomial any higher than third order as he found

no improvement to the R2 value. Linear regression tends to average raw data and so

Jeanty (1991) suggested the use of the lowest order of polynomial regression that

summarizes the data.     Rossevik and Deter (1984) cited the main disadvantage of

polynomials is that the coefficient was likely to change if a different set of independent

variables were used if the investigation was repeated.

       Jeanty (1991) described regression analysis, polynomials, coefficients of

correlation and determination, and prediction from equations and suggested that the

chosen equations:

“should be based on a large sample of data covering a wide range of values for the

independent variable. It should adequately represent the extremes of the curve.” and

“the same table should not be used to assess gestational age from a parameter, and to

determine the normality of this parameter against the gestational age.”

Hadlock et al (1984) developed regression analysis equations for various fetal

parameters but demonstrated that the simple averaging technique and the complex

regression equations give equivalent results. This fact justified the decision of the

Australian investigators de Crespigny et al to formulate the 1989 BPD Australian chart

using simple averaging techniques. Doubilet and Greenes (1991) found that many

published formulae are systematically biased towards underestimation of GA in late

pregnancy believing that this underestimation was due not from the method of

obtaining the measurements but to how the regression formula was derived.

Fetal Growth Assessment                                                         Background

       Another prominent set of work is that of Altman and Royston (1995) who

described how to overcome the bias of the estimates of normal variability for individual

measurements seen with correlations between measurements collected longitudinally.

There have been excellent critical reviews of all of the Altman, Chitty and Royston

papers on fetal growth curves as they have attempted to illustrate simple but effective

methods of analysis of any authors‟ graphs. One such paper was Royston and Wright

(1998) devising a method for estimating reference intervals for various fetal age/growth

charts where the mean and standard deviations are defined as simple formulas plotted

as smooth curves, which are first order polynomials – or quadratic curves. These new

approaches to statistical analysis have been made possible by the rapid advancements in

computer statistical packages.

1.14   Summary of the Literature

Buchanan, Kjos, Bochner and Langer were the authors reviewed in this work for fetal

growth assessment with ultrasound of pregnancies complicated with gestational

diabetes mellitus. It was disappointing that there is very little published research on the

relationship between fetal growth, gestational diabetes and maternal ethnicity. Hadlock

has been one the most prolific researcher on the use of ultrasound to assess fetal size

and growth and for fetal measurement chart formation for the various fetal parameters

produced during the nineteen eighties. The studies published by Hadlock and his co-

workers include BPD charts (Hadlock, Deter, Harrist and Park, 1981), intra and inter-

observer error figures (Hadlock 1982), femur length curves (Hadlock Deter, Harrist and

Park 1984), and articles on fetal growth in racially mixed populations (Hadlock and

Shah 1987) and (Hadlock, Harrist and Yogesh 1990). Statistical analysis has been

Fetal Growth Assessment                                                          Background

another of Hadlock‟s specialities with reports produced on statistical errors (1990), the

duplication of techniques (1991), variability in predicting fetal age (1992) and

ultrasonic reflection of maternal age (1992).       Hadlock‟s works on gestational age

assessment became very repetitive, with many published works identical except for the

fetal parameter being assessed. There appeared to be fierce competition between some

of the larger ultrasound institutes in North America to publish works.             Hadlock

criticises other authors, particularly the main opponent in fetal growth charts, Jeanty,

without justifying the reasons. Jeanty produced growth curves for the same extensive

range of fetal parameters as Hadlock and in addition developed charts for the radius,

ulna, tibia and fibula. Merz was also responsible for a variety of fetal growth charts

such as BPD (Merz, 1987), OFD (Merz, 1988), femur length (Merz, 1987) and (Merz,

Grupner and Kern, 1989) and humerus length (Merz, Grupner and Kern, 1989) and

(Merz, Kim-Kern and Pehl, 1987). The difference between the two Merz charts was

that in 1988 regression analysis was applied to the 1983 version. The latest quality

overseas growth charts available are the 1988 humerus and femur charts produced by

Merz et al from 530 middle class German patients.

       It is not acceptable that most overseas researchers, using relatively small sample

populations, have found no statistically significant differences between ethnic groups,

or between fetal growth charts produced in other countries and therefore conclude that

there is no point in repeating the research.        Westerway‟s (1997) study of racial

differences in ultrasonic fetal measurements questions the notion from some overseas

studies (Spencer et al 1995, Meire et al 1981), that measurements between different

racial groups are not statistically significantly different. Fetal growth curves, regardless

of origin, appear to be consistent until the mid third trimester when a variation of up to

Fetal Growth Assessment                                                           Background

three to four weeks on either side of the mean value is accepted. As a result of this

latitude, many of the authors in the preceding review have concluded that ethnicity has

no statistically significant effect on fetal growth charts for the various fetal parameters.

       Regardless of the fetal measurement charts used it has become apparent that the

charts are only as reliable as the sonographer correctly measuring an image taken at the

correct imaging plane. Dudley (2002) commendably recommended the implementation

of standardised measurement charts and imaging planes to enable introduction of

quality control in ultrasound practices. Similarly in Australia ASUM has suggested

fetal measuring protocols and imaging planes and recommended fetal measurement

charts for the BPD, occipito-frontal diameter, head and abdominal circumference and

femur and humerus lengths.

1.15   Justification for the Study

Australia‟s attempt at chart formation is poor.        In 2000 the only ultrasonic fetal

measurement chart based on an Australian population was the BPD chart of de

Crespigny and Spiers (1989) and the OFD chart from the Royal Hospital for Women in

Sydney. Apart from the BPD chart mentioned above, all other fetal measurement

charts recommended for use in Australia by ASUM were based on middle class

Caucasian populations from America, Germany and Great Britain. This raises the

question of complacency in fetal size/growth assessment in the Australian ultrasound

community and it is time these charts were replaced by relevant data from a relevant


Fetal Growth Assessment                                                        Background

       As discussed earlier, a trend was identified at Hornsby Ku-Ring-Gai Hospital in

Northern Sydney Health Area regarding an increase in macrosomic births, post partum

haemorrhage and birth interventions in both Caucasian and Chinese pregnancies. One

sub-study, the incidence of fetal macrosomia and birth complications in Chinese

immigrant women aimed to investigate these trends, particularly to clarify any

relationship between ethnicity, fetal size and birth problems.

       Obstetricians can utilise ultrasound at all stages of gestation to confirm their

clinical findings and there were many ultrasound referrals from the Royal North Shore

Hospital and the Hornsby Ku-Ring-Gai Hospital Antenatal Clinics that questioned

small for gestational age in Chinese women in late pregnancy based on the fundal

height and maternal weight gain. It was thought that these small for dates pregnancies

may in fact be the normal for Chinese pregnancies and those Chinese pregnancies

assumed to be normal, when compared with Caucasian pregnancies, were those

affected by gestational diabetes. Large babies are a problem and ethnicity should

influence the definition of large so it was also hoped to identify those pregnancies

deemed to be large for gestational age and so at a greater risk from the problems

associated with fetal overgrowth. By assessing fetal size and growth with ultrasound in

late pregnancy in the at risk groups such as large and small for gestational age or

gestational diabetics, many potential problems could be averted. To correctly assess

fetal size it is necessary to accurately measure the various fetal parameters at different

stages of pregnancy. Of particular interest was accurate dating in the first trimester,

third trimester measurements for weight estimation and inter and intra sonographer

measuring accuracy.

Fetal Growth Assessment                                                     Background

1.16   Objectives of the Study

Based on the preceding background and examination of the literature, this study had

four major objectives:

       1.     To construct ultrasonic fetal biometry charts based on an Australian

              population for the crown rump length, head circumference and

              abdominal circumference.      Fetal biometry charts are assessed and

              comparisons made with the graphs in the literature that are relevant to

              this new work.

       2.     The    second    objective   aimed   to   examine   fetal   measurement

              discrepancies. Initially in the first trimester CRL measuring accuracy

              was evaluated, along with the BPD, head and abdominal circumference

              and long bone length searching for the stage at which dates could be

              conflicting, leading to problems with timing for invasive procedures

              such as CVS, amniocentesis and maternal blood analysis. In the third

              trimester inter and intra sonographer fetal measurement reproducibility

              was evaluated in the final 6 weeks of pregnancy to detect any common

              pitfalls of measuring that could affect the report on fetal size. The

              hypothesis being that sonographers entrusted with the measuring of the

              fetus in the third trimester could do so in a competent and reproducible

              manner.     It is also hoped to detect any broad based problems of

              measuring that could be addressed to improve outcomes.

Fetal Growth Assessment                                                      Background

      3.     The third objective was to compare birth weights and rates of fetal

             macrosomia (birth weight > 4000g) and birth complications in both

             Chinese and Caucasian women for two time periods: 1992 and

             1999/2000 and within each of these time periods, to compare obstetric

             complication and intervention rates for these two groups of women.

             This was done by accessing and downloading statistics from the data

             base for all births of Caucasian and Chinese women at the Royal North

             Shore and Hornsby Hospital during the time periods in question.

      4.     The fourth and final study was to compare fetal size in the third trimester

             of pregnancy in both Chinese and Caucasian women with gestational

             diabetes mellitus and to determine whether the mean fetal size in

             Chinese women with GDM varied from either Caucasian pregnancies

             affected by GDM or normal Chinese pregnancies. This involved the

             serial measurement of the fetuses of three groups of women in the third

             trimester of pregnancy – Caucasians affected by GDM and Chinese both

             with and without GDM. The hypothesis was that the normal Chinese

             fetus was smaller than the normal Caucasian fetus and Caucasian and

             Chinese fetuses affected by GDM, in the last five weeks of gestation. It

             was also aimed to determine if any ultrasonically measured fetal

             parameter predicts excessive birth weight of Chinese or Caucasian

             babies and to calculate the fetal growth per day in grams in Caucasian

             and Chinese pregnancies.

Northern Sydney Health ethics approval - study number 0201-003 (Appendix 4).


To top