The Journal of Maternal-Fetal and Neonatal Medicine, February 2011; 24(2): 271279

Which is the most accurate formula to estimate fetal weight in women with severe
preterm preeclampsia?
LUT GEERTS & TANIA WIDMER
Department of Obstetrics and Gynaecology, Tygerberg Hospital, University of Stellenbosch, South Africa
(Received 13 January 2010; revised 31 March 2010; accepted 9 April 2010)
Abstract
Objective. To identify the most accurate formula to estimate fetal weight (EFW) from ultrasound parameters in severe
preterm preeclampsia.
Methods. In a prospective study, serial ultrasound assessments were performed in 123 women with severe preterm
preeclampsia. The EFW, calculated for 111 live born, normal, singleton fetuses within 7 days of delivery using 38 published
formulae, was compared to the actual birth weight (ABW). Accuracy was assessed by correlations, mean (absolute and
signed) (%) errors, % correct predictions within 520% of ABW and limits of agreement.
Results. Accuracy was highly variable. Most formulae systematically overestimated ABW. Five Hadlock formulae utilizing
three or four variables and Woo 3 formula had the highest accuracy and did not differ significantly (mean absolute % errors
6.87.2%, SDs 5.35.8%, 4 75% of estimations within 10% of ABW and 95% limits of agreement between -18/20% and
14/15%). They were not negatively affected by clinical variables but had some inconsistency in bias over the ABW range.
All other formulae, including those targeted for small, preterm or growth restricted fetuses, were inferior and/or affected by
multiple clinical variables.
Conclusion. In this GA window, Hadlock formulae using three or four variables or Woo 3 formula can be recommended.
Keywords: Fetal weight estimation, ultrasound, preterm preeclampsia

Introduction
In patients with severe preterm preeclampsia, important
obstetrical decisions, such as timing and method of
delivery, are influenced by the expected postnatal prognosis. This prognosis is significantly related to the
gestational age (GA) [13] but in developing countries,
where late initiation of antenatal care, lack of early
ultrasound assessment and inaccurate recollection of last
menstrual period date (LMP) are common, the true GA is
often unknown. For this reason fetal weight is often used as
a surrogate prognostic indicator to assist in obstetric and
neonatal management decisions especially in small and/or
growth restricted fetuses who have the highest risk of fetal
compromise and perinatal morbidity and mortality [1,4,5].
In the preterm fetus, ultrasonographic estimated fetal
weight (EFW) is more accurate than clinical estimation
[6] and a good correlation exists between survival rates in
extremely preterm infants based on ultrasonographic EFW
compared to actual birth weight (ABW) [7].
Most formulae used to calculate EFW from ultrasound
parameters have been developed 2030 years ago in often
healthy term fetuses, mostly in developed countries and in
ethnic groups not representative of our local population,
and many formulae have poorer precision in smaller fetuses
[810]. Some formulae were developed for preterm and/or
small fetuses but not validated specifically in the preterm
asymmetrically growth-restricted fetus frequently found in
severe preterm preeclampsia. The ideal formula with high

precision and little bias in all circumstances has not yet

been found and opinions vary on whether inaccuracies
in EFW are primarily due to the equation or due to
inaccuracies in the measurements [8,11,12]. Currently, the
Hadlock 4b [13] formula based on biparietal diameter
(BPD), head circumference (HC), abdominal circumference (AC) and femur length (FL) is used to calculate EFW
in patients examined at Tygerberg Hospital. This study
was undertaken to ascertain whether alternative formulae
estimate fetal weight more accurately in patients with
severe preterm preeclampsia, to quantify the degree of
inaccuracy and to determine which clinical parameters
influence this.

Methods
This prospective study took place at Tygerberg Hospital,
a secondary and tertiary referral centre in the Metro
East region of Cape Town, South Africa, serving a community of mixed ethnicity and low socioeconomic status.
The study was approved by the Committee for Human
Research of the University of Stellenbosch and written
informed consent was obtained. Consecutive women with
severe preterm (2434 weeks) preeclampsia were recruited
from the labor ward, antenatal ward or ultrasound unit.
Severe preeclampsia was diagnosed in women with
proteinuric hypertension if there was severe hypertension
(diastolic blood pressure (BP) of 120 mmHg on one

Correspondence: Prof. Lut Geerts, Room 2056, Second floor, Clinical building, Tygerberg Campus, Francie Van Zyl Drive, Parow, 7505, South Africa, Po Box
19081, Parow, 7505, South Africa. Tel: 27-21-938-9524. Fax: 27-21-931-6595. E-mail: lgeerts@sun.ac.za
ISSN 1476-7058 print/ISSN 1476-4954 online 2011 Informa UK, Ltd.
DOI: 10.3109/14767058.2010.485232

272

L. Geerts & T. Widmer

occasion or diastolic BP 110 or systolic BP 160 mmHg

on more than one occasion 4 h apart) and/or severe
proteniuria (repeated 3 on semiquantitative assay or 5 g
per 24 h) and/or signs of threatening eclampsia and/or
evidence of hemolysis, thrombocytopaenia, impaired liver
or renal function tests or abnormal umbilical artery
resistance index (above the 95th centile for the local
population). GA was determined by the best estimation
available, such as LMP, confirmed by first trimester
ultrasound where possible, or late ultrasound. Three
experienced sonologists assessed the women serially with
a 3.5-MHz curvilinear abdominal probe with color and
duplex Doppler facilities (Siemens Sonoline Antares1,
Germany). While in routine clinical practice ultrasound
assessments are performed on a two-weekly basis, for the
purpose of this specific study consent was obtained to
perform Doppler assessments two to three times weekly
and biometry once a week in order to maximize the
availability of results close to delivery. Measurements of
BPD, HC, AC, and FL were performed on standard
planes: BPD was measured from the outer-to-inner edge of
the skull bones; HC and AC were obtained by the ellipse
method. The amniotic fluid index (AFI) was measured
according to the four quadrant method [14] and the
pulsatility index (PI systolic diastolic/mean velocity)
was measured for the umbilical (UA), uterine (UtA) and
middle cerebral artery (MCA) and ductus venosus by auto
tracing at least three consecutive waveforms of adequate
quality and the presence, absence or reversal of enddiastolic blood flow (EDF) in the UA was noted. HC/AC
ratio 4 95th percentage of indicated asymmetric growth
[15]. Birth weight (BW) deficit was expressed as percent of
the median BW for the given best estimate of GA. ABW
and date of birth were obtained from delivery records.
Multiple gestations, intra-uterine deaths, fetuses with
abnormalities, and patients delivering more than 7 days
after the last EFW were excluded.
EFW was calculated with 38 formulae utilizing various
numbers and combinations of the four ultrasound parameters, obtained from the literature or available on
the ultrasound equipment (Table I). The formulae were
compared in terms of mean (systematic) percentage error
(mean of [ABW7EFW] 6 100/ABW), mean absolute
(random) percentage error (mean of absolute value of
[ABW7EFW] 6 100/ABW) and Pearson correlation coefficient between EFW and ABW. Further analysis was
carried out with SPSS statistical package (version 17, SPSS
Inc, Chicago, IL) on formulae with mean random errors
510%. Weight predictions falling within 5, 10, 15, and
20% of ABW were compared by chi-square. Bias inconsistency was assessed by determining the correlation
between the mean (signed) % error and GA or ABW.
Paired samples t-test was used to compare random errors
and the effect of dichotomous variables (GA 5 430
weeks, ABW 5 41000 g, AFI 5 4 8 cm, UA PI 5 4
P95, MCA PI 4 5P5, UA/MCA PI 5 4 1, asymmetric
growth or EDF in UA (present or absent)). Correlations
with continuous variables were determined by Pearson
correlation coefficient. p-values of 50.05 (two-tailed) were
considered statistically significant. Bland and Altmans
method was used to determine the 95% limits of
agreement referring to the random error [16].

Results
During 10 months starting in February 2003, 4292 women
delivered at the hospital and of the 123 women with severe

preterm preeclampsia who were assessed, 111 were

suitable for this analysis (nine fetuses demised during
expectant management and three had their last biometry
assessed 8 or 9 days before delivery). Mean maternal age
was 27.2 + 5.9 years and median gravidity 2 (17) (45
primigravida). The mean BW was 1402.1 + 422.2 g
(51000 g in 22, 10001499 g in 44, 15001999 g in 33,
and 2000 g in 12). GA was determined by ultrasound
prior to 24 weeks in 42 women and by LMP being
compatible with later ultrasound biometry in another 24
women. In 9 women GA estimation was based purely on
menstrual history and in 36 purely on a first ultrasound
examination after 24 weeks. Mean GA at delivery according to this best estimate was 221.7 + 14.9 days (range 191
252) with 20 women being delivered before 30 weeks
estimated GA, 30 between 30 and 32 weeks, 34 between
32 and 34 weeks, and 20 after 34 weeks. When compared
to the median expected BW for the estimated GA in our
local population, 40 infants weighed less than 75% of the
expected BW, 56 between 75% and 100% and 15%
weighed more than the expected median. Abnormal
ultrasound findings were common: raised UA PI in 62
cases (55.9%) (16 had absent or reversed EDF); low MCA
PI in 54 (49.1%); increased UA/MCA PI ratio (4 1) in 40
(36.4%); raised DV PI in 11 (9.9%); AFI 55 cm in 18
(16.2%); and asymmetric growth in 30 (27%). The median
interval between EFW-assessment and delivery was 2 (07)
days and it was within 2 days in 74 women (66.7%).
EFW for all formulae correlated significantly with ABW
(correlation coefficients 4 0.95 for 18 formulae). Thirtytwo formulae (84.2%) systematically overestimated the
ABW and Hadlock 3Ha had the smallest systematic
(71.2 + 8.6%) and random error (6.8 + 5.3%). The
mean % (systematic) errors in EFW are shown in
ascending order in Figure 1 and mean absolute %
(random) errors in Figure 2. Compared to Hadlock 3Ha,
there was no significant difference in systematic error for
Woo 3 and in random error for Hadlock 3Hb, 4a and b, 3B
and Woo 3 formulae.
Eighteen formulae had a mean random error 510% and
the % of EFW falling within 5, 10, 15, and 20% of ABW
are shown in Figure 3. Predictions were within 5% of the
ABW in 4 44% of cases for Hadlock (2a, 2b, 3Ha and 4b)
and Sabbagha SGA and within 10, 15, or 20% in 4 76, 90,
and 97% of cases, respectively for Hadlock 4a, 4b, 3Ha
and b. Compared to Hadlock 3Ha none of the percentages
was significantly different for any of the Hadlock formulae
using three or four parameters, as well as both Woo
formulae, Scott and Warsof 1. Figure 4 graphically displays
the relative position of the 18 formulae on a plot of mean %
error against its SD. The best performing formulae are
within the shaded box, with systematic errors 53% and
SDs 510%. Compared to Hadlock 3Hb or 4a or b, the
difference in systematic error was not significant for both
Hadlock 2 formulae, Shepard 2, Scott, Weinberger and
Sabbagha SGA. Table II shows the 95% levels of agreement for these 18 formulae, with smallest ranges for
Weiner 1 and the 5 Hadlock formulae using three or four
parameters. Table III shows the factors affecting the
accuracy of the different formulae. Woo 3 and Hadlock
2a, 2b, 3Ha and 4a were not influenced by any of the
factors assessed in this study but most other formulae were
significantly affected by more than one of the variables.
Random errors improved with increasing ABW, advancing
estimated GA or increasing BW deficit for some formulae
while they worsened for others. Accuracy was better in
asymmetrically grown fetuses for Hadlock 4b, 3Hb and 3B,
Schild and Weiner 1 formulae but even in fetuses with a

Campbella
Combsb
Ferreroc
Hadlock 2ad
Hadlock 2be
Hadlock 3Be
Hadlock 3Hae
Hadlock 3Hbe
Hadlock 4ad
Hadlock 4be
Halaskaf
Higginbottomg
Honarvarh
Hsieh 1i
Hsieh 2i
Jordaan 1j
Jordaan 2j
JSUM
Merzk
Ottl
Sabbagha AGAm
Sabbagha SGAm
Schildn
Scott8
Shepard 1p
Shepard 2p
Shinozukaq
Siemerr
Thurnaus
Vintzileost
Warsof 1u
Warsof 2v
Weinberger w
Weiner 1x
Weiner 2x

Author
(AC)
(HC,AC,FL)
(AC,FL)
(AC,BPD)
(AC,FL)
(BPD,AC,FL)
(HC,AC,FL)
(HC,AC,FL)
(BPD,HC,AC,FL)
(BPD,HC,AC,FL)
(BPD,AC,FL)
(AC)
(FL)
(BPD,AC)
(BPD,AC,FL)
(BPD,AC)
(BPD,HC,AC)
(BPD,AC)
(HC,AC,FL)
(GA,HC,AC,FL)
(GA,HC,AC,FL)
(HC,AC,FL)
(HC,AC,FL)
(BPD,AC)
(BPD,AC)
(BPD,AC,FL)
(BPD,FL,AC)
(BPD,AC)
(BPD,AC)
(BPD,AC)
(AC,FL)
(BPD,AC)
(HC,AC,FL)
(HC,AC)

1975
1993
1994
1984
1985
1985
1985
1985
1984
1985
2006
1975
2001
1987
1987
1983
1983

1988
1986
1989
1989
2004
1996
1982
1982
1987
2009
1983
1987
1977
1986
1984
1985
1985

(continued)

Ln EFW/1000 74.564 0.282(AC) 0.00331 (AC)2

EFW 0.23718(AC)2(FL) 0.03312(HC)3
Log10EFW 0.77125 0.13244(AC) 7 0.12996(FL) 0.00173588(AC)2 0.00309212(FL)(AC) 0.18984(FL/AC)
Log10EFW 1.1134 0.05845(AC) 0.000604(AC)2 0.007365(BPD)2 0.000595(BPD)(AC) 0.1694(BPD)
Log10EFW 1.304 0.05281(AC) 0.1938(FL) 7 0.004(AC)(FL)
Log10EFW 1.335 0.0034(AC)(FL) 0.0316(BPD) 0.0457(AC) 0.1623(FL)
Log10EFW 1.5662 0.0108(HC) 0.0468(AC) 0.171(FL) 0.00034(HC)2 7 0.003685(AC)(FL)
Log10EFW 1.326 7 0.00326(AC)(FL) 0.0107(HC) 0.0438(AC) 0.158(FL)
Log10EFW 1.5115 0.0436(AC) 0.1517(FL) 0.00321(AC)(FL) 0.0006923(BPD)(HC)
Log10EFW 1.3596 0.0064(HC) 0.0424(AC) 0.174(FL) 0.00061(BPD)(AC) 7 0.00386(AC)(FL)
Log10 EFW 0.64041(BPD) 7 0.03257(BPD)2 0.00154(AC)(FL)
EFW 0.0816(AC)3
EFW 0.042(FL)2 0.32(FL) 7 1.36
Log10EFW 2.135 0.0056541(AC)(BPD) 7 0.00015515(BPD)(AC)2 0.000019782(AC)3 0.052594(BPD)
Log10EFW 2.7193 0.0094962(AC)(BPD) 7 0.1432(FL) 7 0.00076742(AC)(BPD)2 0.001745(FL)(BPD)2
Log10EFW 71.1683 0.0377(AC) 0.095(BPD) 0.0015(BPD)(AC)
Log10EFW 2.3231 0.02904(AC) 0.0079(HC) 7 0.0058(BPD)
Provided by Siemens (Germany) on Sonoline Antares
EFW 73200.40479 157.07186(AC) 15.90391(BPD)2
Log10EFW 72.0661 0.04355(HC) 0.0008582(HC)(AC) 1.2594(FL/AC) 0.5394(AC)
EFW (755.3 7 16.35(SUM) 0.25838(SUM)2
EFW 1849.4 7 47.13(SUM) 0.37721(SUM)2
EFW 5381.193 150.324(HC) 2.069(FL)3 0.0232(AC)3 7 6235.478(log10(HC))
Ln EFW 0.66(ln(HC)) 1.04(ln(AC)) 0.985(ln(FL))
Log10EFW 71.7492 0.166(BPD) 0.046(AC) 0.002646(AC)(BPD)
Log10EFW 1.2508 0.166(BPD) 0.046(AC) 7 0.002646(AC)(BPD)
EFW 0.23966(AC)2(FL) 1.6230(BPD)3
EFW 75948.336 2101.261(lnAC) 15.613(FL)2 0.577(BPD)3
EFW 9.337(BPD)(AC) 299.076
Log10EFW 1.879 0.084(BPD) 0.026(AC)
Log10(EFW) 0.144(BPD) 0.032(AC) 70.000111(AC)(BPD)2 7 1.599
Ln EFW 2.792 1.08(FL) 0.0036(AC)2 0.027(FL)(AC)
EFW 10.1(AC)(BPD) 481
Log10EFW 1.6961 0.02253(HC) 0.01645(AC) 0.06439(FL)
Log10(BW) 1.6575 0.04035(HC) 0.01285(AC)

Equation

Table I. Formulae used to calculate the estimated fetal weight (in alphabetical order).

Estimated fetal weight

273

1985 (AC,FL)
1985 (BPD,AC)
1985 (BPD,AC,FL)

Log10EFW 0.59 0.08(AC) 0.28(FL) 7 0.00716(AC)(FL)

Log10EFW 1.63 0.16(BPD) 0.00111(AC)2 7 0.0000859(BPD)(AC)2
Log10EFW 1.54 0.15(BPD) 0.00111(AC)2 7 0.0000764(BPD)(AC)2 0.05(FL) 7 0.000992(FL)(AC)

Equation

EFW, estimated fetal weight; BPD, Biparietal diameter; HC, Head circumference; AC, abdominal circumference; FL, femur length; SUM GA HC 2(AC) FL; GA, gestational age. References
for equations used:
a. Campbell S, Wilkin D. Ultrasonic measurement of fetal abdomen circumference in the estimation of fetal weight. Br J Obstet Gynaecol 1975;82:689697.
b. Combs CA, Jaekle RK, Rosenn B, Pope M, Miodovnik M, Siddiqi TA. Sonographic estimation of fetal weight based on a model of fetal volume. Obstet Gynecol 1993;82:365370.
c. Ferrero A, Maggi E, Giancotti A, Torcia F, Pach` A. Regression formula for estimation of fetal weight with use of abdominal circumference and femur length: a prospective study. J Ultrasound Med
1994;13:823833.
d. Hadlock FP, Harrist RB, Carpenter RJ, Deter RL, Park SK. Sonographic estimation of fetal weight. The value of femur length in addition to head and abdomen measurements. Radiology
1984;150:535540.
e. Hadlock FP, Harrist RB, Sharman RS, Deter RL, Park SK. Estimation of fetal weight with the use of head, body, and femur measurements a prospective study. Am J Obstet Gynecol.
1985;151:333337.
f. Halaska MG, Vlk R, Feldmar P, Hrehorcak M, Krcmar M, Mlcochova H, Mala I, Rob L. Predicting term birth weight using ultrasound and maternal characteristics. Eur J Obstet Gynecol Reprod
Biol 2006;128:231235.
g. Higginbottom J, Slater J, Porter G, Whitfield CR. Estimation of fetal weight from ultrasonic measurement of trunk circumference. Br J Obstet Gynaecol 1975;82:698701.
h. Honarvar M, Allahyari M, Dehbashi S. Assessment of fetal weight based on ultrasonic femur length after the second trimester. Int J Gynecol Obstet 2001;73:1520.
i. Hsieh FJ, Chang FM, Huang HC, Lu CC, Ko TM, Chen HY. Computer-assisted analysis for prediction of fetal weight by ultrasound-comparison of biparietal diameter (BPD), abdominal
circumference (AC) and femur length (FL). Taiwan Yi Xue Hui Za Zhi 1987;86:957964.
j. Jordaan HV. Estimation of fetal weight by ultrasound. J Clin Ultrasound 1983;11:5966.
k. Merz E, Lieser H, Schicketanz KH, Harle J. Intrauterine fetal weight assessment using ultrasound. A comparison of several weight assessment methods and development of a new formula for the
determination of fetal weight. Ultraschall Med 1988;9:1524.
l. Ott WJ, Doyle S, Flamm S, Wittman J. Accurate ultrasonic estimation of fetal weight. Prospective analysis of new ultrasonic formulas. Am J Perinatol 1986;3:307310.
m. Sabbagha RE, Minogue J, Tamura RK, Hungerford SA. Estimation of birth weight by use of ultrasonographic formulas targeted to large-, appropriate-, and small-for-gestational-age fetuses. Am J
Obstet Gynecol 1989;160:85460.
n. Schild RL, Fell K, Fimmers R, Gembruch U, Hansmann M. A new formula for calculating weight in the fetus of 5or 1600 g. Ultrasound Obstet Gynecol 2004;24:775780.
o. Scott F, Beeby P, Abbott J, Edelman D, Boogert A. New formula for estimating fetal weight below 1000 g: comparison with existing formulas. J Ultrasound Med 1996;15:669672.
p. Shepard MJ, Richards VA, Berkowitz RL, Warsof SL, Hobbins JC. An evaluation of two equations for predicting fetal weight by ultrasound. Am J Obstet Gynecol 1982;142:4754.
q. Shinozuka N, Okai T, Kohzuma S, Mukubo M, Shih CT, Maeda T, Kuwabara Y, Mizuno M. Formulas for fetal weight estimation by ultrasound measurements based on neonatal specific gravities
and volumes. Am J Obstet Gynecol 1987;157:11401145.
r. Siemer J, Hilbert A, Hart N, Meurer B, Goecke T, Schild RL. A new sonographic weight formula for fetuses 5or 2500 g. Ultraschall Med 2009;30:4751.
s. Thurnau GR, Tamura RK, Sabbagha R: A simple estimated fetal weight equation based on real-time ultrasound measurements of fetuses less than thirty-four weeks of gestation. Am J Obstet
Gynecol 1983;145:557561.
t. Vintzileos AM, Campbell WA, Rodis JF, Bors-Koefoed R, Nochimson DJ. Fetal weight estimation formulas with head, abdominal, femur, and thigh circumference measurements. Am J Obstet
Gynecol 1987;157:410414.
u. Warsof SL, Gohari P, Berkowitz RL, Hobbins JC. The estimation of fetal weight by computer-assisted analysis. Am J Obstet Gynecol 1977;128:881892.
v. Warsof SL, Wolf P, Coulehan J, Queenan JT. Comparison of fetal weight estimation formulas with and without head measurements. Obstet Gynecol 1986;67:569573.
w. Weinberger E, Cyr DR, Hirsch JH, Richardson T, Hanson JA, Mack LA. Estimating fetal weight less than 2000 g: An accurate simple method. Am J Roentgenol 1984;142:973977.
x. Weiner CP, Sabbagha RE, Vaisrub N, Socol ML. Ultrasonic fetal weight prediction: role of head circumference and femur length. Obstet Gynecol 1985;65:812817.
y. Woo JS, Wan CW, Cho KM. Computer-assisted evaluation of ultrasonic fetal weight prediction using multiple regression equations with and without the fetal femur length. J Ultrasound Med
1985;4:6567.

Woo1y
Woo2y
Woo3y

Author

Table I. (Continued).

274
L. Geerts & T. Widmer

Estimated fetal weight

275

Figure 1. Mean (systematic) percentage error of the estimated fetal weight (EFW) compared to the actual birth weight (ABW)
[100 6 (ABW-EFW)/ABW] for the different formulae in ascending order, together with the standard deviation (SD).

Figure 2. Mean absolute (random) percentage error of the estimated fetal weight (EFW) compared to the actual birth weight (ABW) [100
xjABW-EFWj/ABW] for the different formulae in ascending order, together with standard deviation (SD).

normal HC/AC ratio, these three Hadlock formulae were

amongst the most accurate and hardly different from
Hadlock 3Ha. None of the systematic or random errors
was affected by the method of GA estimation. The
Hadlock formulae and both Sabbagha formulae had
consistent bias over the entire estimated GA range while
all other formulae had greater systematic errors in earlier
gestations. With the exception of Hadlock 2a and b, all
formulae were inconsistent across the ABW range with
greater systematic overestimation of weight in smaller
fetuses.

Discussion
This study has compared the accuracy of the largest
number of EFW formulae, and is the first study

investigating the accuracy of EFW in the indigenous

population of South Africa and in the specific clinical
context of severe preterm preeclampsia with its associated
high incidence of low BW, growth restriction, abnormal
placental function, and oligohydramnios. Only formulae
based on BPD, HC, AC, and/or FL with circumferences
measured by ellipse or two diameter methods were
included and the findings can not be extrapolated to
formulae incorporating other parameters or using tracing
[17]. We have not corrected for the interval between EFW
and ABW (by gestation-adjusted projection or adding a set
weight per day [18,19]) because the premise that the EFW
centile would not change over time was not fulfilled due to
the high incidence of placental insufficiency. To limit the
effect of unpredictable fetal growth over a longer interval,
only cases measured within 7 days of delivery were
included in this study.

276

L. Geerts & T. Widmer

Figure 3. Percentage of correct fetal weight (EFW) predictions within 5, 10, 15 and 20% of the actual birth weight (ABW) for the 18
formulae with random error of less than 10%. * p 5 0.05 when compared to Hadlock 3Ha formula.

Figure 4. Plot of the mean (systematic) percentage error of the different formulae against their standard deviation. The best formulae are
within the shaded area.

The studys main finding is the highly variable accuracy

of the published formulae. Although some authors did not
find one formula superior to another [11,20,21], others
have also noted considerable variation in bias and precision
[9,22,23].
The precision of the best 18 formulae compared
favorably with the literature, often with systematic errors
53%, random errors 68% and correct predictions within
5, 10, or 15% of the ABW in the mid-40s, high-70s and
high-80s %, respectively [11,24,25]. The Hadlock 3Ha
formula had the smallest errors but all five Hadlock

formulae incorporating three or four parameters and

Woo 3 were comparable in terms of random error and %
correct predictions and can therefore be recommended
with equal confidence. Other studies have also noted the
high accuracy of Hadlocks formulae [8,12,21,22,24] even
in small or growth restricted fetuses [23,25,26] and in a
literature review [11] Hadlocks formulae had the smallest
mean errors across many studies. Hadlock and Woo3
formulae were not easily affected by clinical parameters,
including ABW. This supports the large study of Kurmanavicius et al. [12] showing Hadlock 2b and 3Hbs

Estimated fetal weight

systematic and random errors to be the most stable across
most of the ABW range. Hadlocks formulae were also not
affected by the BW deficit which may seem in contradiction to Simon et al. [10] who found larger errors in SGA
fetuses but others have previously reported high accuracy
in SGA fetuses [20,27]. Only the fetal growth pattern
affected three of Hadlocks formulaes random error but
the accuracy was actually better in asymmetric growth and
these equations performed best even in symmetrically
grown fetuses. This suggests that these formulae may be

Table II. The 95% limits of agreement between estimated fetal

weight and actual weight (ABW) at delivery for the best 18
formulae, expressed as percentage of ABW.

Hadlock 3Ha
Hadlock 3Hb
Hadlock 4a
Hadlock 4b
Hadlock 3b
Scott
Woo 3
Warsof 1
Hadlock 2a
Hadlock 2b
Woo 2
Sabbagha SGA
Sabbagha AGA
Schild
Shepard 2
Weinberger
Weiner 1
Vintzileos

Mean %
error

SD

Upper
limit

Lower
limit

71.20
72.50
72.60
72.44
72.86
72.32
72.31
2.79
72.94
72.97
2.92
73.14
76.18
4.99
73.05
72.03
7.73
76.10

8.59
8.47
8.59
8.53
8.84
9.71
9.75
10.34
9.63
9.63
10.12
10.31
9.96
9.25
10.18
11.11
8.10
10.97

15.97
14.45
14.58
14.61
14.81
17.11
17.20
23.48
16.33
16.28
23.15
17.49
13.75
23.50
17.31
20.20
23.93
15.83

718.38
718.44
719.77
719.49
720.54
721.75
721.82
717.89
722.21
722.22
717.31
723.77
726.10
713.51
723.40
724.26
78.47
728.03

SD standard deviation.

277

equally useful in settings other than severe preeclampsia,

for example in preterm labor or in idiopathic placental
insufficiency.
The random errors for most of the other top 18 formulae
in our study were clearly affected by a variety of (often
multiple) clinical parameters and the effects were sometimes in opposite directions. The Schild and Weiner 1
formulae, both specifically designed for small and preterm
fetuses, had less significant random errors in smaller,
younger, asymmetrically grown fetuses with larger BW
deficit but the opposite was found for Scott (designed for
fetuses 51000 g), Weinberger (developed in fetuses
52000 g) and Sabbagha SGA (designed for growth
restricted fetuses) as well as the general Shepard 2 and
Vintzileos equations. Oligohydramnios had no negative
effect on any of the tested formulae, which is in contrast
to some previous reports but in keeping with others
[2830].
Most formulae systematically overestimated weight in
this study, as seen by others, especially in fetuses that were
preterm, of low BW or growth restricted [10,21,23,25,26].
Other studies however have demonstrated systematic
underestimations of the ABW with many formulae in
advanced as well as preterm pregnancies or small fetuses
[8,10,12,18,24]. This apparent contradiction can be a
chance finding for formulae with a small systematic error
and is in keeping with studies showing both under- and
overestimations with Hadlocks formulae but always with
small errors and often not significantly different from
zero [11,21]. An alternative explanation for this apparent
contradiction is inconsistency of bias with significant
differences in systematic errors across the BW range
[8,20] leading to differing overall results depending on
the BW mix of the cohort in question. In the current study,
bias inconsistency existed for 16 of the 18 best formulae
and was always in the same direction, tending to overestimate the smaller fetus and underestimate the larger
one. The overall overestimation may therefore be partially
explained by the high number of small fetuses in this

Table III. Effect of clinical and ultrasound variables on the absolute percentage errors of the estimated fetal weight for the different formulae.
Formula
Hadlock 3Ha
Hadlock 3Hb
Hadlock 4a
Hadlock 4b
Hadlock 3B
Hadlock 2a
Hadlock 2b
Woo 2
Woo 3
Warsof 1
Sabbagha AGA
Sabbagha SGA
Schild
Scott
Shepard 2
Weinberger
Weiner 1
Vintzileos

ABW

BW deficit

GA

AFI

Asymmetry

Other

7
7
7

7
7

7
7
7

UA/MCA, Interval

Age, UA, UA/MCA

Interval

MCA

ABW, actual birth weight; BW deficit, % difference between ABW and expected median BW for that estimated gestation; GA, estimated
gestational age; Interval, days between estimation and delivery; AFI, amniotic fluid index; Asymmetry, head/abdominal circumference 4
95th centile for GA; UA, umbilical artery pulsatility index (PI); MCA, Middle cerebral artery PI.
, larger error with increasing value of the variable; 7, smaller error with increasing value of the variable.

278

L. Geerts & T. Widmer

cohort. Significant bias inconsistency was found for

Hadlocks formulae and has previously been reported
[8,9,11,20] but Simon et al. [10] found the more marked
overestimations in smaller and SGA fetuses not different
from zero.
Of the eight formulae specifically designed for small and
preterm fetuses (references in Table I), only five ranked
amongst the 18 best formulae but they were inferior to
Hadlock in many respects. The Schild formula, designed
in a lower BW range but including 41% SGA fetuses,
almost exclusively overestimated weights 51000 g in spite
of a 5% overall underestimation. In contrast to the original
study, its precision was significantly affected by multiple
clinical factors, thereby limiting its clinical value considerably. The Weiner 1 formula was developed in preterm
fetuses with a mean BW close to the current study but the
main pathology was preterm rupture of the membranes.
The formula cannot be recommended because of marked
underestimations as also previously reported [20,21,23].
The Weinberger formula, developed for weights 52000 g,
has not been evaluated extensively [21,23] and had the
largest SD of the top 18 formulae and highly inconsistent
bias with severe overestimations in small fetuses. Sabbagha
developed formulae targeted for SGA, AGA (appropriate
for GA) and LGA (large for GA) fetuses and showed
smaller systematic and random errors compared to
Hadlock. Shamley and Landon [24] however reported
significant inaccuracy at term with somewhat better
performance 52500 g and Jouannic et al. [23] found the
AGA formula to be among the best 51250 g (but not
superior to Hadlock 3Hb). In the current study, both
SGA and AGA formulae had larger errors in smaller
fetuses and with larger BW deficits and while predictions
within 5, 10, and 15% of ABW were similar to Hadlocks
formulae, systematic and random errors were significantly
inferior. Scott developed a formula for weights 51000 g
but the incidence of placental insufficiency in the original
cohort was not documented and Hadlock was the most
accurate of 11 formulae used for comparison. The
systematic error and SD were quite favorable in the
current cohort, in line with Schild reporting the highest
% correct predictions and almost no systematic bias for
Scotts formula in a cohort including 41% SGA fetuses.
We did however find larger errors in smaller and younger
fetuses and with larger BW deficits.
In summary, this study indicates that the performance of
the different formulae is highly variable and that in this GA
and ABW range the most accurate results for EFW, both
for symmetrically and asymmetrically grown fetuses, were
obtained with Hadlocks formulae using three or four
variables. In spite of a small systematic overestimation,
they had high precision and were more accurate than
formulae specifically developed for small, preterm, and/or
growth restricted fetuses. Their accuracy is not negatively
influenced by abnormal clinical parameters which are
common in conditions like preterm preeclampsia and their
independence of estimated GA is an additional benefit in
populations where GA is often not accurately known. It
needs to be kept in mind though that, as seen with all other
formulae tested, there was a significant tendency to
overestimate the EFW of the smallest fetuses.

Acknowledgments
The authors thank Mrs. D. Grove for assistance with data
management and analysis and Mrs. E. Carstens for
recruitment and data collection.

References
1. Geerts L, Odendaal HJ. Severe early onset pre-eclampsia:
prognostic value of ultrasound and Doppler assessment.
J Perinatol 2007;27:335342.
2. Wiltin AG, Saade GR, Mattar F, Sibai BM. Predictors of
neonatal outcome in women with severe preeclampsia or
eclampsia between 24 and 33 weeks gestation. Am J Obstet
Gynecol 2000;182:607611.
3. Hall DR, Odendaal HJ, Kirsten GF, Smith J, Grove D.
Expectant management of early onset, severe pre-eclampsia:
perinatal outcome. Br J Obstet Gynaecol 2000;107:1258
1264.
4. Figueras F, Figueras J, Meler E, Eixarch E, Coll O, Gratacos
E, Gardosi J, Carbonell X. Customised birthweight standards
accurately predict perinatal morbidity. Arch Dis Child Fetal
Neonatal Ed 2007;92:F277F280.
5. Chammas MF, Nguyen TM, Li MA, Nuwayhid BS,
Castro LC. Expectant management of severe preterm
preeclampsia: is intrauterine growth restriction an indication
for immediate delivery? Am J Obstet Gynecol 2000;183:853
858.
6. Chauhan SP, Hendrix NW, Magann EF, Morrison JC,
Kenney SP, Devoe LD. Limitations of clinical and sonographic estimates of birth weight: experience with 1034
parturients. Obstet Gynecol 1998;91:7277.
7. Hovick TJ Jr, Vintzileos AM, Campbell WA, Rodis JF,
Nochimson DJ. Neonatal survival rates based on estimated
fetal weights in extremely premature infants. Am J Perinatol
1989;6:329330.
8. Anderson NG, Jolley IJ, Wells JE. Sonographic estimation of
fetal weight: comparison of bias, precision and consistency
using 12 different formulae. Ultrasound Obstet Gynecol
2007;30:173179.
9. Scott F, Beeby P, Abbott J. Accuracy of estimated fetal weight
below 1,000 g. Aust N Z J Obstet Gynaecol 1996;36:129132.
10. Simon NV, Levisky JS, Shearer DM, OLear MS, Flood JT.
Influence of fetal growth patterns on sonographic estimation
of fetal weight. J Clin Ultrasound 1987;15:376383.
11. Dudley NJ. A systematic review of the ultrasound estimation
of fetal weight. Ultrasound Obstet Gynecol 2005;25:8089.
12. Kurmanavicius J, Burkhardt T, Wisser J, Huch R. Ultrasonographic fetal weight estimation: accuracy of formulas and
accuracy of examiners by birth weight from 500 to 5000 g. J
Perinat Med 2004;32:155161.
13. Hadlock FP, Harrist RB, Carpenter RJ, Deter RL, Park SK.
Sonographic estimation of fetal weight. The value of femur
length in addition to head and abdomen measurements.
Radiology 1984;150:535540.
14. Phelan JP, Smith CV, Broussard P, Small M. Amniotic fluid
volume assessment with the four-quadrant technique at 3642
weeks gestation. J Reprod Med 1987;32:540542.
15. Snijders RJ, Nicolaides KH. Fetal biometry at 1440 weeks
gestation. Ultrasound Obstet Gynecol 1994;4:3448.
16. Bland JM, Altman DG. Statistical methods for assessing
agreement between two methods of clinical measurement.
Lancet 1986;1:307310.
17. Smulian JC, Ranzini AC, Ananth CV, Rosenberg JC,
Vintzileos AM. Comparison of three sonographic circumference measurement techniques to predict birth weight. Obstet
Gynecol 1999;93:692696.
18. Chien PF, Owen P, Khan KS. Validity of ultrasound
estimation of fetal weight. Obstet Gynecol 2000;95:856860.
19. Mongelli M, Gardosi J. Gestation-adjusted projection of
estimated fetal weight. Acta Obstet Gynecol Scand 1996;75:
2831.
20. Robson SC, Gallivan S, Walkinshaw SA, Vaughan J, Rodeck
CH. Ultrasonic estimation of fetal weight: use of targeted

Estimated fetal weight

21.

22.

23.

24.

25.

formulas in small for gestational age fetuses. Obstet Gynecol

1993;82:359364.
Medchill MT, Peterson CM, Kreinick C, Garbaciak J.
Prediction of estimated fetal weight in extremely low birth
weight neonates (5001000 g). Obstet Gynecol 1991;78:286
290.
Burd I, Srinivas S, Pare E, Dharan V, Wang E. Is sonographic
estimation of fetal weight influenced by formula selection? J
Ultrasound Med 2009;28:10191024.
Jouannic JM, Grange G, Goffinet F, Benachi A, Carbrol D.
Validity of sonographic formulas for estimating fetal weight
below 1,250 g: a series of 119 cases. Fetal Diagn Ther
2001;16:254258.
Shamley KT, Landon MB. Accuracy and modifying factors
for ultrasonographic determination of fetal weight at term.
Obstet Gynecol 1994;84:926930.
Siemer J, Egger N, Hart N, Meurer B, Muller A, Dathe O,
Goecke T, Schild RL. Fetal weight estimation by ultrasound:
comparison of 11 different formulae and examiners with
differing skill levels. Ultraschall Med 2008;29:159164.

279

26. Townsend RR, Filly RA, Callen PW, Laros RK. Factors
affecting prenatal sonographic estimation of weight in
extremely low birthweight infants. J Ultrasound Med
1988;7:183187.
27. Guidetti DA, Divon MY, Braverman JJ, Langer O,
Merkatz IR. Sonographic estimates of fetal weight in the
intrauterine growth retardation population. Am J Perinatol
1990;7:57.
28. Edwards A, Goff J, Baker L. Accuracy and modifying
factors of the sonographic estimation of fetal weight in a
high-risk population. Aust N Z J Obstet Gynaecol 2001;41:
187190.
29. Ott WJ, Doyle S, Flamm S. Accurate ultrasonic estimation of fetal weight. Effect of head shape, growth
patterns and amniotic fluid volume. Am J Perinatol
1986;3:193197.
30. Chauhan SP, Scardo JA, Hendrix NW, Magann EF,
Morrison JC. Accuracy of sonographically estimated fetal
weight with and without oligohydramnios. A casecontrol
study. J Reprod Med 1999;44:969973.

Copyright of Journal of Maternal-Fetal & Neonatal Medicine is the property of Taylor & Francis Ltd and its
content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for individual use.