Re is a Colombiana de Obs e icia y Ginecología Vol. 70 No. 1 •Ene o-Ma zo 2019 • (27-38)
In es IgacIón o IgInal
Re Colomb Obs e Ginecol ISSN 2463-0225 (On line) 2019;70:27-38
Abs ac
Objec i e: To de elop and assess an equa ion based
on ma e nal clinical pa ame e s and hi d imes-
e ul asound biome y (combined me hod), and
compa e i wi h ul asound-es ima ed oe al weigh
(EFW) calcula ed using he Hadlock 2 o mula.
Ma e ial and me hods: Coho s udy. A o al o
1,224 women wi h single on p egnancies who had
unde gone oe al ul asound scanning (USS) a 34
weeks we e ec ui ed. The s udy was conduc ed
a a e e ence cen e in Valencia (Spain) be ween
Janua y and Decembe 2016. A ges a ion-adjus ed
p ojec ion (GAP) me hod was applied o es ima ed
oe al-weigh - o -ges a ional-age by oe al gende
a deli e y (EFWa). A mul i a ia e eg ession was
THIRD TRIMESTER ULTRASOUND SCAN
COMBINED WITH A CLINICAL METHOD FOR
ACCURATE BIRTHWEIGHT PREDICTION AT
TERM: A COHORT STUDY IN SPAIN
Ecog a ía del e ce imes e combinada con un
mé odo clínico pa a mejo a la p edicción del
peso del ecién nacido a é mino: un es udio de
coho es en España
Ra ael Vila-Candel, RNM, PhD1; F ancisco Ja ie So iano-Vidal, RNM, MSc2;
En ique Cas o-Sánchez RN, PhD3
Recei ed: June 6/18 - Accep ed: Ma ch 13/19
1 La Ribe a Hospi al Heal h Depa men , Alzi a. Facul y o Nu sing,
Uni e sidad Ca ólica de Valencia “San Vicen e Má i ”. Valencia, Spain.
[email p o ec ed].
2 Facul y o Nu sing, Uni e sidad Ca ólica deValencia “San Vicen e
Má i ”. Valencia. Xà i a-On inyen Heal h Depa men . Xà i a,
Valencia, Spain.
3 Na ional Ins i u e o Heal h Resea ch Heal h P o ec ion Resea ch Uni
(NIHR HPRU) In Heal hca e Associa ed In ec ion and An imic obial
Resis ance a Impe ial College London. London, England.
c ea ed o es ima e oe al weigh a e m (EFWm )
using an h opome ic, demog aphic, ul asono-
g aphic and obs e ic-neona al a iables. EFWa
and EFWm we e calcula ed and compa ed wi h
ac ual bi hweigh .
Resul s: The p opo ion o EFWm wi hin <10%
o ac ual bi hweigh was g ea e han EFWa (82%
s. 65%, p<0.001). The mean ela i e e o in
oe al-weigh p edic ions by using EFWm was
educed om 6.7% o 0.9% (di e ence 5.7% 95%
CI: 5.4 o 6.0) pai ed - es p<0.001, signi ican ly
imp o ing he accu acy a ainable wi h USS. The
EFWm ou pe o med he GAP me hod in p edic -
ing bi hweigh , wi hin 1% ela i e e o . Fo new-
bo ns <2,500 g, he p opo ion o es ima es wi hin
<10% o he ac ual bi hweigh o he EFWm
was g ea e han ha o he EFWa (20.4 s. 16.3%,
p=0.005). Fo babies wi h no mal bi hweigh
(2,500-3,999 g), EFWm was a be e p edic o o
bi hweigh han EFWa (84.5 s. 65.7%, p<0.001).
Conclusions: Ma hema ical modelling o p edic
bi hweigh imp o es hi d imes e ou ine ul-
DOI: h p://dx.doi.o g/10.18597/ cog.3201
Re is a Colombiana de Obs e icia y Ginecología Vol. 70 No. 1 • 2019
28
asound measu emen o es ima e neona al weigh
a e m.
Key wo ds: P egnancy; bi h weigh ; ul asonog-
aphy; mul i a ia e analysis; s a is ics.
RESUMEN
Obje i os: desa olla y e alua un modelo p e-
dic i o de acue do con los pa áme os clínicos
ma e nos y la biome ía de la ecog a ía del e ce
imes e, que pueda mejo a el pode de p edic-
ción del peso al nace en el ecién nacido a é mino,
en compa ación con la es imación calculada po
ecog a ía del peso e al (PFE) usando la ó mula
de Hadlock II.
Ma e iales y mé odos: e isión de 1224 muje es
con emba azos únicos que se habían some ido a
una ecog a ía e al a las 34 semanas (EF). El es udio
se ealizó en un cen o de e e encia en Valencia
(España) en e ene o y diciemb e de 2016. Se aplicó
un mé odo de p oyección ajus ada de ges ación
(PAG) pa a es ima el peso al nace pa a la edad
ges acional y sexo e al en el pa o (PFEa). Se c eó
una eg esión mul i a ian e pa a es ima el peso
e al al nace (PFE m) median e a iables an o-
pomé icas, demog á icas, ecog á icas y obs é ico-
neona ales. Los modelos PFE m y PFEa ue on
calculados pa a compa a sus di e encias espec o
al peso eal al nace .
Resul ados: la p opo ción de PFE m den o de
< 10 % del peso eal al nace ue mayo que la de
PFEa (82 % s. 65 %, p < 0,001). El e o ela i o
medio en las p edicciones de peso e al median e
el uso PFE m pasó de 6, a 0,9 % (Di e encia de
p opo ciones: 5,7 %; IC 95 %: 5,4-6,0); medias
pa eadas: p < 0,001, siendo signi ica i amen e
mejo que la p ecisión que puede se ob enida con
el mé odo ecog á ico. El PFE m supe ó al mé odo
PAG y p edice el peso al nace con un e o ela i o
del 1 %. Pa a ecién nacidos con < 2500 g la p o-
po ción de es imaciones del peso eal < 10 % del
PFE m ue mayo que la del PFEa (20,4 % s. 16,3 %;
p = 0,005). En los ecién nacidos con peso no mal
al nace (2500-3999 g), la capacidad p edic i a pa a
es ima el peso al nace ealizada median e PFE m
ue mejo que la ealizada median e PFEa (84,5 %
s. 65,7%; p < 0,001).
Conclusiones: el modelo ma emá ico c eado
pa a p edeci el peso al nace mejo a la medición
u ina ia de la ecog a ía en el e ce imes e del
emba azo pa a es ima el peso del ecién nacido a
é mino.
Palab as cla e: emba azo; peso al nace ; ul a-
sonog a ía; análisis mul i a ian e; es adís ica.
INTRODUCTION
Accu a e p edic ion o oe al weigh has been o
g ea in e es in obs e ics due o i s signi ican
impac on he cou se and ou come o labou and
deli e y (1, 2). Inco ec es ima ion o oe al weigh
can esul in mul iple and o en dange ous com-
plica ions o he p egnan mo he and he oe us
(2). I has been sugges ed ha accu a e es ima ion
o oe al weigh may con ibu e o success ul ma-
nagemen du ing labou and ca e o he newbo n
in he neona al pe iod, and help a oid complica-
ions associa ed wi h oe al mac osomia o low-
bi hweigh newbo ns, he eby dec easing pe ina al
mo bidi y and mo ali y (3, 4). Un o una ely, bi h
weigh is unknown un il bi h akes place. (5) As
oe al weigh canno be measu ed di ec ly, i mus
be es ima ed om oe al and ma e nal ana omical
cha ac e is ics (3).
Me hods o accu a e p edic ion o bi hweigh
p io o deli e y a e equi ed o es ablish s a egies
designed o educe ad e se p egnancy ou comes
(5, 6). The ools cu en ly used o es ima e oe al
weigh include he e alua ion o oe al g ow h as-
sessmen , and can be b oadly classi ied as ma e nal
me hods, clinical me hods, and imaging me hods
like ul asonog aphy (7, 8). Ul asound es ima ion
(USS) is mo e expensi e and complica ed han
ma e nal o clinical es ima ion, bu i is cu en ly
expec ed o p o ide a mo e accu a e p edic ion
o bi hweigh (6). In p ac ice, he mos common
equa ions o calcula ing es ima ed oe al weigh
(EFW) by USS a e epo ed o be he Shepa d and
29
Thi d T imesTe ulT asound scan combined wiTh a clinical meThod o accu aTe bi ThweighT p edicTion aT Te m: a coho T sTudy in spain
Hadlock o mulae (9). The cu en ly used Hadlock
o mula o oe al weigh es ima ion has an e o
a e o 20%, which may luc ua e depending on he
skills o he examine , equipmen base, condi ions
o he examina ion, as well as he s age o p eg-
nancy o labou (10-12). Rega dless o he o mula
used, he accu acy o he sonog aphic es ima e o
he EFW is a ec ed by subop imal imaging and
biological a ia ion (13, 14). In addi ion, he ac-
cu acy o he sonog aphic es ima e dec eases wi h
inc easing bi hweigh (15), and ends o be o e -
es ima ed in p egnancies suspec ed o being la ge
o ges a ional age (LGA) and unde es ima ed in
p egnancies wi h p e e m p ema u e up u e o
memb anes (PPROM) and suspec ed oe al g ow h
es ic ion (FGR) (16). The sensi i i y and speci ic-
i y o he Hadlock o mula in he de ec ion o oe al
mac osomia a e 62% and 93%, espec i ely (4, 12,
16, 17). The le el o in a/in e -obse e a iabili y
in oe al measu emen , and he impac o e o s
on g ow h assessmen and disc epancies wi hin
s udy designs, exceed 14% wi h 95% con idence
in e als (18-20).
Two la ge s udies ha e compa ed oe al weigh
p edic ion (clinical me hod s. ul asonog aphy)
and ound ha USS was mo e accu a e han clini-
cal es ima ion o bi hweigh in he lowe ange
(<2,500 g) (21). Howe e , his was no he case in
he 2,500–4,000 g anges, whe e clinical es ima ion
was mo e p ecise (22). Finally, bo h me hods we e
equally adequa e in he highe ange o bi hweigh
(>4,000 g) (23). On he o he hand, Chauhan e
al. (4) ound ha ul asound es ima ion o oe al
weigh was mo e accu a e han clinical es ima-
ion in p e e m p egnancies, bu no so in e m
o pos - e m p egnancies. Due o such limi a ion,
esea che s ha e explo ed o he sonog aphic o
clinical pa ame e s o a mix o hem, co ela ing
wi h oe al weigh , wi h a highe p edic i e alue.
We he e o e aimed o de i e a eliable equa ion
based on ma e nal clinical pa ame e s and hi d
imes e ul asound biome y (combined me hod)
and compa e i wi h he ul asound-es ima ed
oe al weigh (EFW) calcula ed using he Hadlock
2 o mula.
MATERIAL AND METHODS
Design and popula ion. We pe o med a p ospec i e
coho s udy in women seen a La Ribe a Uni e -
si y Hospi al (LRUH) in Valencia (Spain) o p eg-
nancy ollow-up and deli e y be ween Janua y and
Decembe 2016. We included women wi h a i s
p ena al appoin men and USS be ween 5 and 12
weeks o p egnancy, single- oe us p egnancy wi h
no oe al abno mali ies, and bi h be ween 38 and
40 weeks. Da a on ma e nal p e-p egnancy weigh ,
symphysis- undal heigh measu emen (SFH), and
USS examina ion a hi d imes e (34 weeks) by
a gynaecologis had o be documen ed in he elec-
onic heal h eco d. P egnancies complica ed by
polyhyd amnios, hypo hy oidism, p eeclampsia,
ges a ional diabe es and oligohyd amnios we e ex-
cluded om he s udy. The LRUH is a public 300-
bed e ia y-le el heal hca e cen e which p o ides
heal h se ices o 250,000 people app oxima ely.
I is he sole hospi al p o iding ma e ni y se ices
o p egnan women in he a ea, wi h an a e age o
3,000 bi hs pe yea .
Sample size and sampling. Du ing he s udy pe iod,
2,017 women consen ed o be included in he s udy.
A ep esen a i e sample size was calcula ed. The
null hypo hesis was ha he e we e no s a is i-
cally signi ican di e ences in bi hweigh accu acy
be ween he clinical and sonog aphic me hods.
Assuming an expec ed di e ence be ween bo h
es ima es o weigh (EFWm and EFWa) g ea e
han o equal o 200 g o conside s a is ically
signi ican di e ences, accep ing an alpha isk o
0.05 and be a o 0.2 in a wo-sided es , a common
s anda d de ia ion o 450, and a d op-ou a e o
30%, hen he necessa y sample would be 39 pa -
icipan s wi hin each g oup. Howe e , all p egnan
women who ag eed o pa icipa e du ing he s udy
pe iod (one yea ) we e included.
P ocedu e. The oppo uni y o pa icipa e in
he s udy was o e ed be o e he assessmen . An
Re is a Colombiana de Obs e icia y Ginecología Vol. 70 No. 1 • 2019
30
in o med consen was eques ed by he gynecologis
in cha ge. A he ime o ec ui men in he hi d
imes e , demog aphic a iables such as ma e nal
age and coun y o o igin we e ga he ed om elec-
onic ma e ni y eco ds a discha ge. Sel - epo ed
obacco use in he hi d imes e was also ob ained
om he elec onic medical eco ds a he heal h
cen es by communi y midwi es du ing he p ena al
con ol pe iod. All o hese da a we e collec ed by
he esea che in cha ge. Measu emen s ob ained
and eco ded on he p ena al cha s a LRUH by he
communi y midwi e esponsible o he women’s
ou ine p ena al isi s we e also e alua ed by he
esea che s a 34 weeks and he deli e y da e, in
o de o educe biased measu emen s. Da a on
obs e ical a iables we e also collec ed. SFH was
measu ed in cen ime es wi h nonelas ic measu e-
men ape om he uppe bo de o he symphysis
pubis o he op o he u e ine undus, o e e sed
di ec ion (13). Ul asound examina ions a 34
weeks we e pe o med by sonog aphe s wi h ap-
p op ia e aining on he SONOLINE G60 model
(SIEMENS). Ul asound es ima ed oe al weigh
(EFW) was calcula ed using he Hadlock 2 o mula
(8,20), using ou oe al indices: Bipa ie al diame e
(BPD), emu leng h (FL) and abdominal ci cum-
e ence (AC), eco ded a 34 weeks a LRUH (Log
10 weigh = 1,326 - 0.00326 AC x FL + 0.0107
HC + 0.0438AC + 0.158 x FL). Bi hweigh was
eco ded in he deli e y oom by midwi es ollow-
ing clamping and umbilical co d sepa a ion, using
a digi al scale (SECA®, Vogel & Halke GmbH &
Co. Hambu g, Ge many) wi hin a 10 g accu acy.
Weigh was documen ed in he elec onic medical
eco d oge he wi h all o he bi h- ela ed da a.
The ges a ion-adjus ed p ojec ion (GAP) me h-
od was applied o each p egnancy by calcula ing he
a io be ween he EFW, a he ime o he emo e
ul asound, and he median oe al weigh o ha
ges a ional age by oe al gende (19, 24). This a io
was hen mul iplied by he median bi hweigh
o he ges a ional age by oe al gende a deli e y
esul ing in he GAP-p edic ed bi hweigh . The
ul asound es ima ions (EFWa) we e each com-
pa ed o he ac ual bi hweigh . Finally, me hods
we e compa ed: Es ima ed oe al weigh by GAP
me hod (EFWa) (USS a hi d imes e wi h he
Hadlock 2 o mula adjus ed by ges a ional age a
bi h and oe al gende ), and es ima ed oe al weigh
wi h he combined clinical and GAP me hod (EF-
Wm ) (mul i a ia e lineal eg ession me hod) we e
calcula ed o compa e wi h he ac ual bi hweigh .
Va iables o be measu ed: P edic i e a iables we e
age, ges a ional weigh gain and pa i y, oe al gende ,
ges a ional age a bi h (GA, in comple e weeks),
smoking in 3 d imes e (as dicho omous a iable).
Ges a ional weigh gain (GWG) was calcula ed a e
applying he di e ence be ween weigh on he day
o deli e y and weigh on he i s consul a ion
documen ed in he clinical eco d. P e-p egnancy
BMI was calcula ed aking in o accoun he ini ial
weigh (5-8 weeks o p egnancy) and he ma e nal
heigh squa ed (kg/m2). Resul a iables: Bi h-
weigh , ges a ion-adjus ed p ojec ion (GAP).
S a is ical analyses. Desc ip i e s a is ics da a
a e p esen ed as mean and s anda d de ia ion o
con inuous a iables, o median and in e qua ile
ange o non-no mally dis ibu ed and ca ego i-
cal a iables p o ided as ange and pe cen age.
No mali y o con inuous a iables was assessed
using he Kolmogo o -Smi no es . In he e en
ha a iables did no adjus o no mali y, a non-
pa ame ic es was used. In he bi a ia e analysis,
co ela ions be ween he dependen (bi hweigh )
and independen a iables collec ed we e s udied
using he S uden - es o compa e mean quan i a-
i e a iables.
All a iables wi h s a is ical signi icance (p<0.05)
and clinical alue we e included in a mul i a ia e
analysis o iden i y he mos accu a e bi hweigh
p edic ion equa ion (EFWm ). To analyse he ela-
ionship be ween bi hweigh and di e en co a i-
ables (clinic, demog aphic and obs e ical a iables),
an adjus ed mul iple linea eg ession model was ap-
plied using a s epwise me hod o a iables shown o
ha e an e ec on bi hweigh . In he linea eg ession
31
Thi d T imesTe ulT asound scan combined wiTh a clinical meThod o accu aTe bi ThweighT p edicTion aT Te m: a coho T sTudy in spain
model, he pa ial F was used o compa e he di -
e en models ob ained. The p inciple o pa simony
was es ablished in o de o selec he simples model
wi h he smalles numbe o a iables.
The accu acy o he di e en me hods (EFWa/
EFWm ) o es ima ing oe al weigh was hen
e alua ed by calcula ing he Pea son co ela ion R
coe icien be ween he es ima ed oe al weigh s
ob ained using each equa ion and ac ual bi hweigh .
In aclass Co ela ion Coe icien (ICC) was used o
e alua e he deg ee o ag eemen o bo h me hods
wi h he ac ual neona al weigh . Absolu e e o was
de ined as he absolu e alue o EFWa//EFWm
minus he ac ual bi hweigh , and he ela i e e o
alue as he absolu e EFWa//EFWm e o di ided
by he bi hweigh mul iplied by 100.
Mean e o di e ences be ween bo h me hods
we e assessed by he pai ed - es o Gaussian
con inuous da a. The mean e o ep esen s he
sum o he posi i e (o e es ima ion) and nega i e
(unde es ima ion) de ia ions om he ac ual bi h-
weigh , app oxima ing ze o in a me hod wi h e y
low o no sys ema ic e o . A h eshold o ela i e
e o wi hin ±10% o ac ual bi hweigh was cho-
sen as he cu -o alue o examina ion accu acy.
The pe cen ages o bi hweigh p edic ions wi hin
10% o he ac ual bi hweigh we e calcula ed and
compa ed using he McNema es . Each ou come
measu e was hen assessed o o e all oe al weigh
and o h ee ca ego ies o weigh <2,500 g, 2,500-
<4,000 g, and ≥4,000 g. The o e all co ela ion
coe icien s o ul asound-based, and clinically
de e mined es ima es we e also compa ed. All s a-
is ical es s we e pe o med using he SPSS Ve sion
23 so wa e package (IBM SPSS Inc., 2008 Chicago,
IL, USA; www.spss.com); p alues o <0.05 e lec
s a is ical signi icance.
E hical conside a ions: The s udy was conduc ed in
acco dance wi h he basic p inciples o all medical
esea ch (Decla a ion o Helsinki), espec ing he
applicable legal p ecep s ega ding he p o ec ion
o pe sonal da a. The s udy was app o ed by he
Resea ch E hics Commi ee o La Ribe a Uni e si y
Hospi al (Re e ence no. #441-14). Conside a ions
such as con iden iali y, olun a y pa icipa ion,
and ull in o ma ion on he na u e o he s udy
we e ex ended o all pa icipan s. The a ending
gynaecologis ec ui ed he women a e he hi d
imes e USS and ob ained hei in o med consen
o pa icipa e in he s udy.
RESULTS
O he ini ial 2,017 p egnan women deli e ed a
he LRUH du ing he s udy pe iod, 458 women
(22.7%) did no mee he inclusion c i e ia: 36
(7.9%) win p egnancies, 27 (5.9%) <35 weeks,
334 (72.9%) >40 weeks a bi h, 22 (4.8%) poly-
hyd amnios, 34 (7.4%) oligohyd amnios, and i e
(1.0%) oe al de o mi ies. Missing da a om an-
ena al ma e ni y eco ds esul ed in 335 women
(16.6%) subsequen ly excluded om he inal analy-
sis:108 (32.2%) wi hou documen ed hi d imes-
e US, and 227 (67.8%) wi hou SFH eco ded.
The e o e, a o al o 1,224 (60.7%) p egnancies
we e inally included o analysis. The mean age
o he pa icipan s was 31.0 ± 6.0 yea s (median
32.50, ange 18-42), mean ges a ional age a deli e y
was 39.14 ± 1.5 weeks ( ange 35-40), and 48.6%
(595/1224) women we e p imipa ous. The mean
ac ual bi hweigh o he s udy popula ion was
3,254 ± 448.4 g. Fo y-nine (4.0%) had a bi h-
weigh o <2,500 g, 1,118 (91.3%) weighed be ween
2,500-3,999 g, and 57 (4.7%) weighed >4,000g.
The demog aphic and clinical cha ac e is ics o he
s udy popula ion a e depic ed in Table 1.
The a iables ha showed s a is ical signi icance
wi h bi hweigh in he bi a ia e analysis we e: coun-
y o o igin (p=0.007), pa i y (p<0.001), ma e nal
age (p=0.007), p e-ges a ional BMI (p<0.001),
SFH (p<0.001), smoke s a us (p=0.0012), ges a-
ional weigh gain (p=0.005) and EFWa (p<0.001).
A mul i a iable model was pe o med. The p edic i e
a iables and coe icien s in he mul i a ia e analysis
a e shown in Table 2. The ollowing equa ion was
de i ed: EFWm (g) = -560.4 + (SFH x 51.6) –
(smoke [0=no, 1=yes) x 74.6) + (GAP x 0.59).
Re is a Colombiana de Obs e icia y Ginecología Vol. 70 No. 1 • 2019
32
Di e ences o he weigh o he newbo n be-
ween EFWm and EFWa a e shown in Table 3.
Ac ual bi hweigh had a s ong posi i e Pea son
wo- ail co ela ion wi h bo h he combined me h-
od (EFWm ) and ul asound (EFWa) es ima ed
oe al weigh s (R=0.91, p<0.001 s. R=0.87,
p<0.001, espec i ely), as shown in Figu e 1. The
di e ence in bi hweigh p edic ion be ween bo h
me hods (EFWa and EFWm ) and he ac ual bi h-
weigh was analysed. Fo he EFWa, he di e ence
Table 1.
Demog aphic and obs e ic cha ac e is ics o 1,224 women seen o p egnancy ollow-up
and deli e y a La Ribe a Uni e si y Hospi al (LRUH) in Valencia (Spain), 2016
Value
Ma e nal na ional o igin Spain 993 (81.1)
(n (%))
Eu opean coun ies 83 (6.8)
O he Eu opean coun ies 19 (1.6)
No h A ica 83 (6.8)
Cen al-Sou h Ame ica 32 (2.6)
Asia 14 (1.1)
Foe al gende (n (%)) Male 620 (50.7)
Female 604 (49.3)
Bi hweigh (n (%))
<2,500 g 49 (4.0)
2,500-3,999 g 1118 (91.3)
>4,000 g 57 (4.7)
P e-ges a ional BMI
<18.5
18.6-24.9
25.0-29.0
>30.0
36 (2.9)
765 (62.5)
273 (22.3)
150 (12.3)
Smoke in 3 d T imes e (n (%)) Yes 96 (7.8)
No 1128 (92.2)
Ma e nal age (mean ± SD) 31±6.0
Pa i y (mean ± SD) 0.65±0.8
Ges a ional age (mean ± SD) 275.6±7.6
Bi hweigh (g) (mean ± SD) 3254±448.4
GWG (kg) (mean ± SD) 11.9±5.3
BMI: body mass index; GWG: ges a ional weigh gain; SD: s anda d de ia ion
wi h he ac ual bi hweigh was 187.4g ± 361.4
(95% CI: 167.1-207.7), whe eas o he EFWm
he di e ence was -0.68g ± 315.6 (95% CI: -18.3-
17.1), di e ences being s a is ically signi ican ,
espec i ely (pai ed - es p<0.001; p<0.001).
The di e ence in weigh es ima ion be ween he
wo me hods, based on he di e ence be ween
he es ima ed weigh and he ac ual inal weigh ,
was 188.1g ± 361 (95% CI: 178.2-197.9; pai ed
- es p<0.001). The mean ela i e alue e o o
33
Thi d T imesTe ulT asound scan combined wiTh a clinical meThod o accu aTe bi ThweighT p edicTion aT Te m: a coho T sTudy in spain
Table 2.
Mul i a ia e linea eg ession analysis p edic ing bi hweigh by clinical me hod in 1,224 newbo ns
o p egnan women seen a La Ribe a Uni e si y Hospi al (LRUH) in Valencia (Spain), 2016
S anda dized coe icien 95% CI
B SE p- alue Lowe limi Uppe limi
(Cons an ) -560.417 125.336 <0.001 -806.314 -314.519
SFH 51.602 3.692 <0.001 44.360 58.845
Smoke 3 d T imes e -74.638 33.621 0.027 -140.599 -8.676
EFWa* 0.593 0.023 <0.001 0.547 0.639
R:710; R2 adjus ed: 503
95% CI: 95% con idence in e al; B: uns anda dized eg ession coe icien ; SE: s anda d e o o he es ima ed; SFH: symphysis- undal heigh a 35-40 weeks; EFWa:
es ima ed oe al weigh by ul asound scan a 33-35 weeks wi h he Hadlock 2, by GAP me hod (adjus ed by ges a ional age a bi h and oe al gende ).
Table 3.
Accu acy o combined me hod and ul asound es ima ed oe al weigh s o 1,224 newbo ns
o p egnan women seen a La Ribe a Uni e si y Hospi al (LRUH) in Valencia (Spain), 2016
(n=1,224).
ABW p edic ion
(95% CI)
Mean ela i e e o
(95% CI)
P edic ion
wi hin ±10% (%) ICC
EFWa 3442.17 6.67% 65.3 0.743
(3419.08-3465.27) (5.99-7.35) (p<0.001) *
EFWm 3254.08 0.97% 82.7 0.803
(3236.22-3271.93) (0.41-1.55) (p<0.001) *
ABW: Ac ual Bi h weigh ; EFWa: es ima ed oe al weigh by hi d- imes e ul asound scan using he Hadlock 2 o mula adjus ed by ges a ional age a bi h and
oe al gende ; EFWm : es ima ed oe al weigh by mul i a ia e linea eg ession; 95% CI: 95% Con idence in e al; ICC: in aclass co ela ion coe icien wi h he
ac ual bi hweigh .
*p- alue: ob ained by eliabili y analysis using a wo-way mixed model wi h absolu e ag eemen ype.
EFWm was lowe han o EFWa (0.97% ± 10.1
s. 6.67% ± 12.1), and he di e ences we e s a-
is ically signi ican by pai ed - es (5.7% ± 2.0,
95%CI: 5.4-6.0, p<0.001).
The p opo ion o ul asound es ima ed weigh s
(EFWa) wi hin <10% o he ac ual bi hweigh was
signi ican ly lowe han ha ob ained wi h he com-
bined me hod (EFWm ) (65.3% s. 82.7%) he di -
e ence being signi ican (McNema es , p<0.001).
The ICC was signi ican ly highe in he case o he
combined me hod es ima ion e sus ac ual weigh ,
when compa ed o ul asound es ima ion e sus
ac ual weigh (0.803 s. 0.743, p<0.001).
The e we e s a is ically signi ican di e ences
be ween weigh es ima ion me hods by bi h-
weigh ca ego ies. Fo newbo ns wi h <2,500g
bi hweigh , he p opo ion o es ima es wi hin
<10% o he ac ual bi hweigh o he EFWm
was signi ican ly g ea e han o EFWa (20.4% s.
16.3%, p=0.005). Fo babies wi h no mal bi h-
weigh (2,500-3,999g), he combined me hod was
signi ican ly g ea e han he ul asound me hod
(84.5 s. 65.7%, p<0.001). And inally, in mac-
osomic newbo ns (>4,000g) he p opo ion o
es ima es wi hin <10% o he ac ual bi hweigh
o he combined me hod we e lowe han he
Re is a Colombiana de Obs e icia y Ginecología Vol. 70 No. 1 • 2019
34
Table 4.
Compa ison p edic ions wi hin 10% be ween he accu acy o combined me hod and ul asound
es ima ed oe al weigh s o 1.224 newbo ns o p egnan women seen a La Ribe a Uni e si y Hospi al
(LRUH) in Valencia (Spain), 2016
Bi h weigh ca ego ies
EFWm
P edic ion wi hin
<10% n (%)
EFWa
P edic ion wi hin
<10% n (%)
p- alue*
<2,500 (n=49) 10 (20.4%) 8 (16.3%) 0.005
2,500-3,999 (n=1,118) 945 (84.5%) 735 (65.7%) <0.001
>4,000 (n=57) 57 (100.0%) 56 (98.2%) N/A
EFWa: es ima ed oe al weigh by hi d- imes e ul asound scan wi h he Hadlock 2 o mula adjus ed by ges a ional age a bi h and oe al gende ;
EFWm : es ima ed oe al weigh mul i a iable eg ession; N/A: insu icien cell numbe o pe o m analysis.
*p- alue: McNema es
Figu e 1.
Co ela ion be ween mul i a ia e linea eg ession and ul asound oe al weigh es ima ion o 1,224
newbo ns o p egnan women seen a La Ribe a Uni e si y Hospi al (LRUH) in Valencia (Spain), 2016
35
Thi d T imesTe ulT asound scan combined wiTh a clinical meThod o accu aTe bi ThweighT p edicTion aT Te m: a coho T sTudy in spain
ul asound me hod (100.0 s. 98.2%), howe e
he di e ences we e no s a is ically signi ican ly
(p=0.238). Table 4 p esen s he di e ences in e -
o o bo h es ima ions’ me hods and o di e en
bi hweigh ca ego ies.
DISCUSSION
In his pape , we epo a no el mul i a iable model
based on ma e nal cha ac e is ics capable o p e-
dic ing neona al weigh in a la ge popula ion o no -
mal p egnancies. Only a ew s udies ha e p e iously
compa ed he p edic i e capaci y o bi hweigh by
clinical and ul asonic measu emen s (3, 22, 23, 25,
26). Ou esul s sugges ha he EFWm model
ha akes in o accoun SFH, smoking and EFWa
can adequa ely es ima e neona al weigh a e m
(mean ela i e e o and p edic ion wi hin 10%).
Addi ionally, ou s udy shows a s a is ically signi i-
can di e ence in explaining bi hweigh when i
is compa ed wi h EFW by adjus ed USS (EFWa).
When he esul was compa ed wi h ac ual
bi hweigh , he EFWm p edic ion inc eased i s
accu acy o wi hin 5.7% o ac ual bi hweigh (±
188g), a esul imp o ing he indings epo ed by
Emechebe e al. (9.2%; 299 g), and esembling bes
epo ed alues o clinical me hods desc ibed by
o he s au ho s (3, 27, 28). Fo he ul asonog aphic
me hod alone, ou esul s (a ound 7%) a e also
consis en wi h o he s udies whe e he mean ela-
i e e o o p edic ed bi hweigh a ied om 6%
o 12% o ac ual bi hweigh (9, 15). Based on ou
esul s, EFWm yields a p edic ion wi hin 10% o
ac ual bi hweigh , a ound 83% in line wi h he
epo by Cu i e al., (84%) and ou pe o ming
he da a by Shi u o Kayem e al. (70% and 58%,
espec i ely) (13, 26).
Ou s udy has shown ha EFWm is as accu a e
as EFWa wi hin he no mal bi hweigh ange in
acco dance wi h o he esea ches ha ha e shown
accu acies (mean ela i e e o ) be ween 7–19% (3,
16). Rega ding he ca ego y unde 2,500 g, howe e ,
he accu acy o he clinical me hod is lowe . Only
20% o bi hweigh s below he 2,500 g h eshold a e
es ima ed p ope ly by he clinical me hod wi hin
10% o ac ual bi hweigh , in line wi h Scioscia o
Dudley e al. (8, 29) who epo ed a mean ela i e
e o anging be ween 13–19%. In he high bi h-
weigh (≥4,000 g) g oup, he di e ence in he
means was no s a is ically signi ican in p edic ing
mac osomic newbo ns, in line wi h o he s udies
(3, 19), sugges ing ha bo h ul asound and he
clinical me hod a e equally accu a e in p edic ing
oe al mac osomia (22).
O e all, he mean ela i e e o o he clinical
me hod was lowe han he ul asound me hod
(1%). This sugges s ha he clinical me hod o
oe al weigh es ima ion is gene ally mo e accu a e
han he ul asound me hod. Whils ou indings
a e consis en wi h hose epo ed elsewhe e (23,
24), he small p opo ion o low bi hweigh and
mac osomic newbo ns in ou sample wa an s
cau ion in he in e p e a ion o he esul s and
sugges s ha u he s udies wi h la ge samples
and cen ed on hese subpopula ions would be
necessa y. In e es ingly, he mean absolu e e o
can be misleading because i is he sum o posi i e
and nega i e de ia ions om ac ual bi hweigh ,
hus a i icially educing he di e ence be ween
ac ual bi hweigh and es ima ed bi hweigh . I
is a measu e o sys ema ic e o in each me hod
a he han a a ia ion om bi hweigh . On he
o he hand, he mean ela i e e o e lec s he
a iabili y no ed ega dless o i s di ec ion and, as
such, being a mo e accu a e p edic o o di e ences
in ela ion o ac ual bi hweigh . Thus, o p ac ical
clinical pu poses, he a ia ion be ween p edic ed
bi hweigh and ac ual bi hweigh is bes exp essed
as mean ela i e e o (3, 22).
This s udy has se e al limi a ions. Fi s , i migh
be a gued ha he knowledge o ul asound mea-
su emen by he clinician measu ing undal heigh
o he subop imal eco d comple ion iden i ied in
ou clinical heal h eco ds would in luence la e
measu emen s. Howe e , he co ela ion coe -
icien s ela ing USS and SFH measu emen s o
bi hweigh did no depend on he o de in which
