Professional Documents
Culture Documents
The major influence on birth weight is gestational age. At any given week of gestation, however, size
varies enormously. There is no specific cut-off that separates abnormally large or small babies for
gestational age from normal. Instead, function alters as a continuum across the weight distribution.
Small babies are prone to hypoxia, acidosis, and stillbirth. Large babies on the other hand are
associated with prolonged labor and mechanical problems. The optimum size for fetal survival is 1
to 1.5 standard deviations above the mean, whereas cesarean section rates are lowest when the fetal
weight is 0.5 to 1.5 standard deviations below the mean. Antenatal detection of both very small and
very large babies is difficult and imprecise. Expectant management is therefore preferable unless
there are very specific indications for emergency or operative delivery.
© 2004 Elsevier Inc. All rights reserved.
he management of babies that are large or cases in each step becomes small, approximately
T small for gestational age presents many
problems, not the least of which is that size at
100 or less.
Figure 2 illustrates that the lowest perinatal
birth is a continuum without precise cut-offs mortality occurs when the birth weight is 1.0 to
delineating normal from abnormal. This can be 1.5 standard deviations (SD) above the mean (1
illustrated using data from the North West SD above the mean is approximately the 84th
Thames regional database (containing data on percentile, 2 SD is approximately the 98th cen-
all pregnancies booked into 15 maternity units tile). It can also be seen that perinatal mortality
in the North West Thames Region of London increases steadily away from this central group,
between 1988-1998 inclusive). The database in- without any clear cut-off at a particular SDS. As
cludes 497,105 pregnancies, of which 439,689 babies are smaller for gestational age, their peri-
progressed far enough in pregnancy to have a natal mortality increases, reaching 100% at 5.5
birth weight recorded. Data were entered by standard deviations from the mean. Larger ba-
trained clerks or midwives with on-line valida- bies also have a higher perinatal mortality, but
tion and prompting, and standard definitions after a standard deviation of 2.5, the numbers of
for a range of clinical measurements. The qual- perinatal deaths are too small to be meaningful.
ity of data collected has been shown to be high.1
The overall perinatal mortality rate was 2.63 per
For the purposes of this paper, I have analyzed
thousand, of which the greater proportion was
data from 137,114 White European women in
made up of stillbirths. Antepartum stillbirths ac-
their first registrable pregnancy, and gestational
counted for 1.46 per thousand, intrapartum still-
age at delivery of 37 weeks or more. This avoids
births 0.39 per thousand, and indeterminate
the complication of ethnic differences, and the
stillbirths 0.15 per thousand. The pattern of still-
effect of preterm delivery. For each week of
births mirrored that of overall perinatal mortal-
gestational age, the mean and the standard de-
viation of birth weight was calculated for boys
and girls, and then the birth weight of each baby
From the Division of Paediatrics, Obstetrics, and Gynecology, De-
expressed as a positive or negative standard de-
partment of Obstetrics and Gynaecology, Imperial College, London,
viation from the mean according to gender UK.
(standard deviation score, SDS). Address reprint requests to Philip Steer, BSc, MB, BS, MD, FRCOG,
Figure 1 illustrates the distribution of birth Academic Department of Obstetrics and Gynaecology, Chelsea and
weight by standard deviation scores, in steps of Westminster Hospital, 369 Fulham Rd, London SW10 9 NH, UK;
e-mail: p.steer@imperial.ac.uk.
0.5 of a standard deviation. It can be seen that © 2004 Elsevier Inc. All rights reserved.
outside a standard deviation of 3 (which in- 0146-0005/04/2801-0007$30.00/0
cludes 99.73% of the population) the number of doi:10.1053/j.semperi.2003.10.013
Figure 5. Percentage of babies transferred to the Figure 7. Mean Apgar score (⫾ 95% confidence lim-
special care baby unit following birth according to its) at 5 minutes by standard deviation score of birth
standard deviation score of birth weight. weight.
of growth during gestation is a better predictor large babies irrespective of the aetiology of their
of peripartum and neonatal dysfunction than macrosomia. Measures of size that take into ac-
size per se.3 Many small babies are constitution- count maternal dimensions have advantages
ally small (“healthy small”4), while some babies (see Gardosi, this issue), but even then, correc-
that are failing to achieve their growth potential tion for severe maternal underweight for exam-
are above the 10 centile.5 These latter babies can ple may result in false reassurance.
be termed “normal weight growth restricted.”
Large babies on the other hand may be consti- Implications for Management
tutionally large, or large because their mother is
overweight or diabetic. The implications, such as Small Babies
neonatal hypoglycemia, are greater in the latter Antenatal detection of babies that are small for
group, although mechanical problems in labor gestational age is currently neither easy nor ef-
such as shoulder dystocia are common in all fective. Ideally, and to be most cost-effective,
screening on a population basis should be prac-
tical using history and clinical examination. The
single most effective indicator of a high risk of a
small for dates baby is a small baby in a previous
shows that in at least 50% of cases the fetal heart livery. There is no simple answer to this di-
rate pattern returns to normal once labor be- lemma, which has to be addressed by close con-
gins. This is presumably because of increased sultation with the parents and neonatologists.
sympathetic drive causing redistribution of A small but significant proportion of intra-
blood flow in the fetus toward the vital organs uterine growth restriction is caused by infection
such as the brain and the heart, resulting in with cytomegalovirus, which is very difficult to
improved oxygenation. However, if there is evi- detect antenatally; it is usually diagnosed by cul-
dence of marked hypoxaemia, suggestive of al- ture of the virus from the urine of the newborn
ready established acidosis, delivery by cesarean baby. Rubella is also a significant cause but is
section is probably safer because the tolerance of usually clinically more obvious because of the
the growth restricted fetus to further hypoxia maternal rash which occurs. Abnormalities of
during labor is limited.22,23 Some have suggested growth associated with maternal use of “recre-
that the addition of intrapartum hypoxemia to ational drugs” (such as cocaine, heroin, and al-
restricted antepartum cardiac output may signif- cohol) and medically prescribed drugs such as
icantly increase sequelae such as necrotizing en- anticonvulsants and Warfarin should also be
terocolitis,24 although others have found little or considered and managed appropriately.
no effect.25,26
It should also always be borne in mind that
the aetiology of intrauterine growth restriction Large Babies
may not be simply uteroplacental insufficiency.
Any small baby should be screened carefully for As shown in Figures 1 to 10, the risk to the large
congenital abnormalities, including aneuploidy. for gestational age baby is not so much antepar-
This involves detailed and systematic ultrasound tum hypoxaemia and acidosis as the problem of
examination, and chromosomal analysis from prolonged labor and intrapartum trauma. The
amniocentesis and cordocentesis may also be majority of macrosomic babies are not associ-
necessary. About 1 in 5 very small babies will ated with diabetes, but with large mothers. Thus,
prove to have congenital abnormalities, and this within the North West Thames database cohort
proportion is higher if growth failure is detected as defined above for this study, macrosomia (de-
before 26 weeks, or there is associated polyhy- fined arbitrarily as a birth weight ⬎ ⫽ 4 kg)
dramnios.4 This is because growth restriction occurred in 9.5% of 123,681 women with normal
associated with uteroplacental insufficiency in- carbohydrate tolerance during pregnancy,
creases in incidence as pregnancy progresses. It 15.8% of 766 women with gestational diabetes,
is uncommon before 26 weeks’ gestation, and is and 23.4% of 312 women with established dia-
usually associated with oligohydramnios because betes. However, macrosomia associated with di-
of the reduced renal perfusion in growth re- abetes accounted for only 1.6% of all babies
stricted babies. Management of growth re- weighing 4 kg or more. In contrast, the risk of
stricted babies with aneuploidy, particularly tri- macrosomia rose from 4.8% in underweight
somy 21, can pose ethical problems. Parents may women (body mass index [BMI] ⬍ 20, n ⫽
feel ambivalent about emergency delivery by ce- 8,040), to 8.1% in women with a normal BMI (20
sarean section for a Down’s fetus with failure of to 24.9, n ⫽ 61,260), 12% in overweight women
growth, or an abnormal fetal heart rate tracing. (BMI 25 to 29.9, n ⫽ 27,687) and 15.6% in obese
However, such babies may not necessarily die women (BMI 30 or more, n ⫽ 9,156). However,
without emergency delivery, and expectant man- unlike diabetes, overweight was associated with
agement can therefore result in a baby with 32.9% of macrosomia, and obesity with 14.1%.
impaired intellectual potential being further Thus, almost half of all macrosomia occurred in
compromised by hypoxic ischemic damage. An- women who were overweight or obese. More-
other area of ethical difficulty is the manage- over, these macrosomic babies were more than
ment of multiple pregnancy, where one or more twice as likely to be delivered by emergency ce-
of the fetuses is growth restricted, and the re- sarean section as babies weighing less than 4 kg
mainder are growing normally. To rescue the (21% v10.3%). Unfortunately, it is in obese
growth restricted baby will place the normally mothers that fetal weight assessment is most dif-
grown fetus at increased risk from preterm de- ficult, both clinically and by ultrasound, and
Large and Small for GA Fetuses 65
both large and small babies are likely to go accuracy deteriorates away from the mean value
undetected. with a tendency toward the mean. Thus, there is
One of the most feared causes of intrapartum a tendency to overestimate the weight of small
trauma is shoulder dystocia. Most papers have babies and underestimate the weight of large
reported that fetal macrosomia is the principal ones. For example, in one study, in diabetic
risk factor for shoulder dystocia.27-33 However, patients, 26.3 % of the birth weights were under-
our ability to detect macrosomia antenatally is estimated by more than 15 %, compared to 5.4%
limited and none of these studies recommended in a nondiabetic control group.38 In another
routine abdominal delivery for suspected mac- study, the mean error in diabetic patients was
rosomia. In one study of 4,480 births,30 in which 265g, with the error being over 500g in almost
the overall frequency of shoulder dystocia was 10%.39 Even in diabetics, maternal weight rather
2% (90 cases), the majority of cases (93%, 83 than glycemic control is one of the most useful
cases) occurred in infants weighing less than predictors.40 The topic of route of delivery of
4,500g. In a detailed study of cost effectiveness, macrosomic fetuses in diabetic pregnancies has
Rouse et al34 concluded that “For the 97% of recently been fully reviewed in this Journal by
pregnant women who are not diabetic, a policy Conway.41
of elective cesarean delivery for ultrasonographi-
cally diagnosed fetal macrosomia is medically
and economically unsound. In pregnancies com- Summary
plicated by diabetes, such a policy appears to be
The management of small for dates and large
more tenable, although the merits of such an
for dates fetuses is complicated by the fact that
approach are debatable.”
they exist as part of a continuum, with no clear
If elective cesarean section is not justified,
distinction between pathology and normality.
what about induction of labor before the due
Accordingly, management decisions must be
date? Attempting to deliver the baby vaginally
based not only on the estimated fetal weight for
before it gets too large is a logical approach.
gestational age, but also on functional criteria
However, even the most recent studies do not
such as changes in growth velocity, and mea-
show any advantage to induction before the due
sures of fetal wellbeing including components of
date.35 Moreover, a recent systematic review of 9
the biophysical profile. The mode of delivery will
observational studies and 2 randomized trials of
depend on a balance being struck between fetal
labor induction for suspected fetal macrosomia
and maternal risk, the facilities available for
has shown that such a policy results in an in-
monitoring labor, and the wishes of the parents.
creased cesarean delivery rate without improv-
ing perinatal outcomes.36 It appears that the
benefits of the baby being smaller are vitiated by References
the disadvantage of a less efficient labor.
An important aspect of the management is 1. Cleary R, Beard RW, Coles J, et al: The quality of rou-
our ability, or perhaps more correctly our lack of tinely collected maternity data. Br J Obstet Gynaecol
101:1042-1047, 1994
ability, to identify macrosomic fetuses correctly
2. Mahadevan N, Pearce M, Steer PJ: The proper measure
by ultrasound. As so often happens, there is a of intrauterine growth retardation is function, not size.
conflict between sensitivity and specificity. For Br J Obstet Gynaecol 101:1033-1035, 1994
example, in the study reported by Gilby et al37 if 3. Danielian PJ, Allman ACJ, Steer PJ: Is obstetric and
an abdominal circumference of 35 cm was used neonatal outcome worse in fetuses who fail to reach
their own growth potential? Br J Obstet Gynaecol 99:452-
as the cut-off, 99% of macrosomic babies were
454, 1992
identified, but only one in ten of babies labelled 4. Resnik R: Intrauterine growth restriction. Obstet Gy-
ultrasonically as macrosomic had a birth necol 99:490-496, 2002
weight ⬎ 4 kg. On the other hand, if a cut-off of 5. Chard T, Costeloe K, Leaf A: Evidence of growth retar-
abdominal circumference of 38 cm was used, dation in neonates of apparently normal weight. Eur J
Obstet Gynecol Reprod Biol 45:59-62, 1992
more than half the babies so identified weighed
6. Neilson JP: Symphysis-fundal height measurement in preg-
less than 4 kg and only just over 50% of macro- nancy. Cochrane.Database.Syst.Rev 2000;CD000944.
somic babies were identified. With ultrasonic 7. Lindhard A, Nielsen PV, Mouritsen LA, et al: The impli-
fetal weight estimation, as with most techniques, cations of introducing the symphyseal-fundal height-
66 Philip Steer
measurement. A prospective randomized controlled tery and necrotising enterocolitis. Arch Dis Child 66:805-
trial. Br J Obstet Gynaecol 97:675-680, 1990 807, 1991
8. Niknafs P, Sibbald J: Accuracy of single ultrasound pa- 26. Adiotomre PN, Johnstone FD, Laing IA: Effect of absent
rameters in detection of fetal growth restriction. Am J end diastolic flow velocity in the fetal umbilical artery on
Perinatol 18:325-334, 2001 subsequent outcome. Arch Dis Child 76:F35-F38, 1997
9. Bricker L, Neilson JP: Routine ultrasound in late preg- 27. Bahar AM: Risk factors and fetal outcome in cases of
nancy (after 24 weeks gestation). Cochrane.Database. shoulder dystocia compared with normal deliveries of a
Syst.Rev. 2000;CD001451 similar birth weight. Br J Obstet Gynaecol 103:868-872,
10. Gulmezoglu AM, Hofmeyr GJ: Bed rest in hospital for 1996
suspected impaired fetal growth. Cochrane.Database. 28. Baskett TF, Allen AC: Perinatal implications of shoulder
Syst.Rev. 2000;CD000034 dystocia. Obstet Gynecol 86:14-17, 1995
11. Say L, Gulmezoglu AM, Hofmeyr GJ: Maternal nutrient 29. Ecker JL, Greenberg JA, Norwitz ER, et al: Birth weight
supplementation for suspected impaired fetal growth. as a predictor of brachial plexus injury. Obstet Gynecol
Cochrane.Database.Syst.Rev. 2003;CD000148 89:643-647, 1997
12. Say L, Gulmezoglu AM, Hofmeyr GJ: Maternal oxygen 30. Gonen R, Spiegel D, Abend M: Is macrosomia predict-
administration for suspected impaired fetal growth. able, and are shoulder dystocia and birth trauma pre-
Cochrane.Database.Syst.Rev. 2003;CD000137 ventable? Obstet Gynecol 88:526-529, 1996
13. Gulmezoglu AM, Hofmeyr GJ: Plasma volume expansion 31. Graham EM, Forouzan I, Morgan MA: A retrospective
for suspected impaired fetal growth. Cochrane Database analysis of Erb’s palsy cases and their relation to birth
Syst Rev 2000;CD000167 weight and trauma at delivery. J Mat Fet Med 6:1-5, 1997
14. Gulmezoglu AM, Hofmeyr GJ: Betamimetics for sus- 32. Nocon JJ, McKenzie DK, Thomas LJ, et al: Shoulder
pected impaired fetal growth. Cochrane Database Syst dystocia: an analysis of risks and obstetric maneuvers.
Rev 2001;CD000036 Am J Obstet Gynecol 168:1732-1737; discussion 1737-
15. Gulmezoglu AM, Hofmeyr GJ: Calcium channel blockers
1739, 1993
for potential impaired fetal growth. Cochrane Database
33. Perlow JH, Wigton T, Hart J, et al: Birth trauma. A
Syst Rev 2000;CD000049
five-year review of incidence and associated perinatal
16. Brodszki J, Marsal K: Management of pregnancies with
factors. J Reprod Med 41:754-760, 1996
suspected intrauterine growth retardation in Sweden.
34. Rouse DJ, Owen J, Goldenberg RL, et al: The effective-
Results of a questionnaire. Acta Obstet Gynecol Scand
ness and costs of elective cesarean delivery for fetal
79:723-728, 2000
macrosomia diagnosed by ultrasound. JAMA 276:1480-
17. Galan HL, Ferrazzi E, Hobbins JC: Intrauterine growth
1486, 1996
restriction (IUGR): Biometric and Doppler assessment.
35. Mulik V, Usha Kiran TS, Bethal J, et al: The outcome of
Prenat Diagn 22:331-337, 2002
macrosomic fetuses in a low risk primigravid population.
18. Neilson JP, Alfirevic Z: Doppler ultrasound for fetal
Int J Gynaecol.Obstet 80:15-22, 2003
assessment in high risk pregnancies. The Cochrane Li-
brary (2). 2002. Oxford, Update Software. 36. Sanchez-Ramos L, Bernstein S, Kaunitz AM: Expectant
19. The GRIT Study Group: A randomised trial of timed management versus labor induction for suspected fetal
delivery for the compromised preterm fetus: Short term macrosomia: A systematic review. Obstet Gynecol 100:
outcomes and Bayesian interpretation. Br J Obstet 997-1002, 2002
Gynaecol 110:27-32, 2003 37. Gilby JR, Williams MC, Spellacy WN: Fetal abdominal
20. Dawes GS, Moulden M, Redman CW: System 8000: Com- circumference measurements of 35 and 38 cm as pre-
puterized antenatal FHR analysis. J Perinat Med 19:47- dictors of macrosomia. A risk factor for shoulder dysto-
51, 1991 cia. J Reprod Med 45:936-938, 2000
21. Dawes GS, Moulden M, Redman CW: Improvements in 38. Wong SF, Chan FY, Cincotta RB, et al: Sonographic
computerized fetal heart rate analysis antepartum. estimation of fetal weight in macrosomic fetuses: dia-
J Perinat Med 24:25-36, 1996 betic versus non-diabetic pregnancies. Aust NZJ Obstet
22. Low JA, Pancham SR, Worthington D: Fetal heart decel- Gynaecol 41:429-432, 2001
eration patterns in relation to asphyxia and weight-ges- 39. Best G, Pressman EK: Ultrasonographic prediction of
tational age percentile of the fetus. Obstet Gynecol 47: birth weight in diabetic pregnancies. Obstet Gynecol
14-20, 1976 99:740-744, 2002
23. Low JA: Metabolic acidosis and fetal reserve. Baillieres 40. Schaefer-Graf UM, Heuer R, Kilavuz O, et al: Maternal
Clin Obstet Gynaecol 10:211-224, 1996 obesity not maternal glucose values correlates best with
24. Kempley ST, Gamsu HR, Vyas S, et al: Effects of intra- high rates of fetal macrosomia in pregnancies compli-
uterine growth retardation on postnatal visceral and cated by gestational diabetes. J Perinat Med 30:313-321,
cerebral blood flow velocity. Arch Dis Child 66:1115- 2002
1118, 1991 41. Conway DL: Delivery of the macrosomic infant: Cesar-
25. Malcolm G, Ellwood D, Devonald K, et al: Absent or ean section versus vaginal delivery. Semin Perinatol 26:
reversed end diastolic flow velocity in the umbilical ar- 225-231, 2002