Communicating hunger and satiation in the first 2 years of life: a...

Communicating hunger and satiation in the first 2years oflife: a systematic review

Janet McNally*, Siobhan Hugh-Jones*, Samantha Caton†, Carel Vereijken‡,Hugo Weenen‡ and Marion Hetherington**Institute of Psychological Sciences, University of Leeds, Leeds, UK, †School of Health and Related Research (ScHARR), University of Sheffield, Sheffield, UK, and‡Danone Nutricia Research, Utrecht, Netherlands

Abstract

Responsive feeding has been identified as important in preventing overconsumption by infants. However, thisis predicated on an assumption that parents recognise and respond to infant feeding cues. Despite this,relatively little is understood about how infants engage parental feeding responses. Therefore, the aim of thissystematic review was to identify what is known about infant communication of hunger and satiation andwhat issues impact on the expression and perception of these states. A search of Medline, CINAHL, Webof Science, PsycINFO, Science Direct and Maternal and Infant care produced 27 papers. Eligibility criteriaincluded peer reviewed qualitative and/or quantitative publications on feeding behaviours, hunger, andsatiation/satiety cues of typically developing children in the first 2 years of life. Papers published between1966 and 2013 were included in the review. The review revealed that feeding cues and behaviours are shapedby numerous issues, such as infants’ physical attributes, individual psychological factors and environmentalfactors. Meanwhile, infant characteristics, external cues and mothers’ own characteristics affect how feedingcues are perceived. The existing literature provides insights into many aspects of hunger and satiation ininfancy; however, there are significant gaps in our knowledge. There is a lack of validated tools for measuringhunger and satiation, a need to understand how different infant characteristics impact on feeding behaviourand a need to extricate the respective contributions of infant and maternal characteristics to perceptions ofhunger and satiation. Further research is also recommended to differentiate between feeding driven by likingand that driven by hunger.

Keywords: infant feeding, hunger, satiation, behaviour, cue, communication.

Correspondence: Janet McNally, Institute of Psychological Sciences, University of Leeds, Leeds LS2 9JT, UK. E-mail: [email protected]

Introduction

Childhood obesity is prevalent in developed countries(Ogden et al. 2012; Wang & Lobstein 2006), andresearch has focussed on factors that might increaseobesity risk in children. Some of the factors identifiedthus far include parental body mass index (BMI),birthweight, early adiposity, weight gain during the firstyear of life and maternal feeding practices (Dev et al.2013; Reilly et al. 2005).

Several reviews indicate that maternal feedingpractices may increase obesity risk by influencing theearly entrainment of appetite control (De Lauzon-Guillain et al. 2012; Disantis et al. 2011; Hurley et al.2011). However, the precise mechanisms linking feeding

practices and childhood obesity remain unclear.DiSantis et al. (2011) proposed a theoretical role formaternal feeding ‘responsiveness’ in infant and childoverweight. ‘Responsive’ mothers are sensitive to hun-ger and satiation cues and respond to these appropri-ately, while discordant maternal responses are aproposed risk factor for obesity. Worobey et al. (2009)found lower maternal sensitivity to feeding cues at6months predicted infant weight gain between 6 and12months. Hurley et al. (2011) also found two typesof discordant response, restrictive feeding and indulgentfeeding, to be associated with a high BMI in infantsand young children. Meanwhile, DiSantis et al. (2011)suggested that a third kind of discordant response,maternal pressure to eat, may also increase obesity risk.

© 2015 The Authors. Maternal & Child Nutrition published by John Wiley & Sons Ltd. Maternal & Child Nutrition (2016), 12, pp. 205–228 205

DOI: 10.1111/mcn.12230

Review Article

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduc-tion in any medium, provided the original work is properly cited.

bs_bs_banner

http://creativecommons.org/licenses/by/4.0/

Evidence for the latter view is mixed as Farrow &Blissett (2008) found pressure to eat at 1year to beassociated with lower weight at 2years. However,Farrow & Blissett (2006) found infants with high weightgain in the first 6months whose mothers exhibited pres-sure to eat continued on this trajectory between 6 and12months. In addition, Lumeng et al. (2012) found as-sertive prompts to eat and maternal intrusiveness to beassociated with higher adiposity in toddlers. Poorresponsiveness to satiation cues in the form ofpressure to eat may therefore also affect obesity risk.

Notwithstanding reported associations betweenmaternal responsiveness and infant adiposity, thedirection of causality between these remains unclear.Overfeeding may arise from insensitivity to fullnesscues or the use of food to settle fractious infants(Worobey et al. 2009; Redsell et al. 2010). Restric-tive feeding practices may also play a role by in-creasing the desirability, and consequentconsumption, of restricted foods (Dev et al. 2013).Importantly though, mothers may simply beresponding to their child’s appetite (Webber et al.2010a) as some infants have a more avid appetitethan others (Agras et al. 1990). In turn, mothersmay restrict intake for children they perceive toover-eat or may pressure children with small appe-tites to eat more (Webber et al. 2010b). There istherefore a need to better understand the issues thataffect interpretations of and responses to infantfeeding cues in order to develop interventions toprevent overfeeding. The aim of the current reviewwas to consider the evidence regarding what infantscommunicate during meals and what parents

respond to. Specifically, the review aimed to iden-tify the following:

1. How hunger and satiation are expressed in infantsand toddlers.

2. The issues that impact on the expression andperception of infant feeding cues.

3. How hunger and satiation behaviour can bedifferentiated fromeating driven by the hedonic fea-tures of food.

Method

Search strategy

An initial scoping exercise was conducted to establishwhether reviews had been completed previously on in-fant feeding cues. The Cochrane Systematic ReviewDatabase was searched followed by Medline,CINAHL, Web of Science, PsycINFO, Science Directand Maternal and Infant care. The scoping exercisewas also used to generate search terms and synonymsand to establish the utility of the databases for thesearch. Final keyword search terms appear in Table 1.

Keyword searches were conducted up to January2014. Where databases offered combined keywordand subject heading search options (Medline, Maternaland Infant Health and PsycINFO), search terms (in-fant and feeding) were mapped to subject headings.Following keyword and combined keyword and sub-ject heading searches, results were refined by apply-ing initial limiters: English language, full text, peerreviewed, human and child.

Key messages• Hunger cues are easier to perceive by mothers than satiation cues, and feeding cues are easier to interpret as chil-dren grow older.

• Infant feeding cues are diverse and highly variable across and within individuals being influenced by many issuesincluding age, sex, genotype, developmental maturity and feeding method.

• Feeding interactions are dyadic in nature, and both infant characteristics (age, sex and temperament) and maternalcharacteristics (e.g. body mass index) may affect how feeding cues are perceived.

• There is a need to develop methods for measuring infant hunger and satiation and for discriminating feeding drivenby hunger from that driven by liking. Additional research is also recommended regarding the impact of differentinfant characteristics on feeding behaviour and of different maternal and infant characteristics on perceptions ofhunger and satiation.

206 J. McNally et al.

© 2015 The Authors. Maternal & Child Nutrition published by John Wiley & Sons Ltd. Maternal & Child Nutrition (2016), 12, pp. 205–228

Results

The study selection process is outlined in Fig. 1; 5635articles were returned. Their titles were screened by thefirst author according to the inclusion and exclusioncriteria (Table 2), and irrelevant papers were discarded(n=5515). Duplicate and review articles were thenremoved (n=35). The abstracts of the remaining articles(n=85) were screened for relevance, and exclusion andinclusion criteria were applied resulting in 50 articlesbeing discarded.

Of the remaining 35 articles, only 34 were fully accessi-ble. These were read in full, and those not fulfilling

inclusion and exclusion criteria were discarded (n=17).A reference list search and a citation search wereconducted for the remaining 17 articles. From these, afurther 13 articles were obtained that met theinclusion/exclusion criteria (Table 3). The remaining 30articles were subjected to quality assessment. Thosescoring below 11 on the 22 point scale were removed,and a final 27 articles were selected.

Quality assessment of studies

In the final stage of selection, articles were rated forquality using a tool developed by Moore (2012). Thetool was selected on its suitability for assessing both qual-itative and quantitative papers and non-interventionstudies. Quality ratings were subjected to inter-rater reli-ability analysis using a non-fully crossed design; themainauthor rated all papers, while second authors each rateda different subset of papers. A random sample of 14papers was selected for the intraclass correlation analy-sis. A high level of inter-rater agreement was found

Fig. 1. Systematic review selection process.

Table 1. Final search terms

(Infan* OR baby OR babies OR toddler* or newborn* or neonate*)AND

(Feed*OReat

*ORhungerOR satietyOR satiationOR fullnessORmeal

*)

AND(cue* OR behavio?r or behavio?rs OR sign* OR communication)

Hunger and satiation in the first 2 years of life 207


(single measures intraclass correlations by use of a one-way random effects model), r=0.82 (P< 0.001).

Overview of selected papers

Terminology

Several selected studies use the terms satiety andsatiation synonymously (e.g. Hodges et al. 2008;Llewellyn et al. 2012). This review distinguishesbetween these with ‘satiation’ referring to the processleading to the cessation of eating and ‘satiety’referring to the feeling of fullness after eating thatdetermines the interval before the next meal (Blundell& Bellisle, 2013).

Summary of selected studies

Themainmethodological features of the selected studiesare reported inTable 3.Most studieswere cross-sectional(n=11); others had longitudinal/repeated measures

components (n=8) or were experimental/quasi-experi-mental (n=6). Two studies were cohort studies. Theseinvolved questionnaires and one used modelling of her-itability of eating traits (Llewellyn et al. 2012). Most ofthe cross-sectional studies employed surveys and struc-tured observational methods. The exceptions to thiswere Hodges et al. (2008) and Anderson et al. (2001)who used semi-structured interviews and focus groups,respectively.

Areas of investigation

Six main research areas were identified in the retrievedstudies, as indicated below.A summary of findings fromthe selected papers appears in Table 3.1. Maternal perceptions of infants’ and toddlers’

hunger and satiation communications.2. Movement and sucking behaviours associated with

hunger and satiation.3. Impact of infant characteristics on the expression

and perception of hunger and satiation.4. Feeding behaviour norms in infancy.5. Feeding method, composition of milk, hunger, satia-

tion and satiety.6. Infant food preferences, how these are expressed and

implications for understanding hunger and satiation.

Findings

Maternal reports of feeding cues

Several studies have investigated mothers’ perceptionsof infants’ feeding cues. Anderson et al. (2001) usedfocus groups to examine maternal beliefs regardingreadiness for weaning. In this context, perceptions ofhunger related both to babies’ characteristics (e.g. ageand weight) and their behaviour (e.g. increased rateof milk consumption, agitation and changed sleepingpatterns). Mothers also reported being able to identifya ‘hungry cry’; however, this was differentiated fromother cries by time of day rather than the characteristicsof the cry itself. Reported satiation cues included thebaby seeming more ‘content’ and them wishing to eatless often.

Gross et al. (2010) also examined mothers’ percep-tions of infant hunger and satiation. In a survey relatingto general feeding rather than weaning, they found

Table 2. Inclusion and exclusion criteria

Inclusion criteria Exclusion criteria

– Qualitative or quantitative – Studies with a non-humanpopulation

– Peer reviewed – Studies with a primary focus onchildren over the age of 2 years

– Studies of feeding behavioursin typically developing childrenaged 0–2 years

– Studies with a primary focuson maternal (rather than infant)feeding behaviours

– Studies of hunger and satiationcues in typically developingchildren in the first 2 years of life

– Studies with a primary focuson feeding in premature infants

– Studies relating to infant ortoddler feeding in populationswith specific medical conditions(e.g. cystic fibrosis, developmentaldisorder, exposure to maternalsubstanceabuse and cleft lip)– Studies with a primary focus oninfant or toddler feeding inpopulations with maternaldisorder (e.g. depressive illnessand eating disorder)– Review articles/books– Papers not written in English– Papers where the full textversion is unavailable



Tab

le3.

Selected

papers

Autho

rsan

dtitle

Participan

tsan

dsample

Designan

dmetho

dsMainfind

ings

Implications

forun

derstand

ing

hung

eran

dsatia

tionin

infancy

Qua

lityratin

gsfirsta

ndsecond

raters

And

ersonet

al.(2001)

N=29

Cross-sectio

nal

Introd

uctio

nof

solid

swas

basedon

infant

age,size,w

eigh

tand

avarietyof

increased

infant

hung

ercues.

Bothinfant

beha

viou

rs(che

wingha

nds

andcrying

)an

dinfant

characteristics

(age

andsize)areused

bymothe

rsto

determ

inefeed

ingstatealon

gwith

external

cues

such

astim

e.

20/22

‘Rattling

theplate–

reason

san

dratio

nales

forearlyweaning

’

Multip

arou

san

dprim

iparou

smothe

rs,

meanage27

years,

ofba

bies

aged

betw

een

8an

d18

weeks,m

ean

age13

weeks.

Focus

grou

pdiscussion

sexploring

belie

fsan

dattitud

esregardingthe

introd

uctio

nof

solid

food

.Qua

litativeconten

tana

lysis.

19/22

Blossfieldet

al.(2007)

N=70

Qua

si-exp

erim

ental

Variabilityin

consum

ptionof

chop

pedcarrots

relatedto

familiarity

with

differen

ttextures,

high

erdietaryvariety,

food

fussinessan

dthenu

mbe

rof

teeth

infantspo

ssessed.

Amou

nteatenwas

associated

with

level

ofen

joym

ent.

Amou

ntof

food

consum

edvaries

accordingto

likingas

wella

swith

infant

characteristics(e.g.p

ickine

ssor

numbe

rof

teeth).

22/22

‘Texture

preferen

ces

of12-m

onth-old

infantsan

dtherole

ofearlyexpe

rien

ces’

39malean

d31

infants

aged

betw

een48

and

57weeks,m

eanage

52.7weeks.

Infantsfedchop

pedor

pureed

carrots.Measures–am

ount

offood

consum

ed,m

aterna

lratings

ofen

joym

enta

ndqu

estio

nnaire

measures,e.g.CEBQ*an

dFFQ.†

22/22

Darlin

gton

&Wrigh

t(2006)

N=75

Short-term

long

itudina

lSlow

weigh

tgainwas

sign

ificantly

associated

with

fearfultem

peramen

t.Fastw

eigh

tgainwas

associated

with

irritablebe

haviou

r.

Infant

tempe

ramen

tmay

affect

appe

tite

orthecommun

icationof

hung

er,

althou

ghmothe

rsmay

feed

irritable

babies

morein

orde

rto

soothe

them

.

20/22

‘The

influe

nceof

tempe

ramen

tonweigh

tgain

inearlyinfancy’

43malean

d32

female

infantsbe

tween8an

d12

weeks

ofage,mean

age10

weeks.

Infants’birthw

eigh

tsan

dweigh

tstake

nat

8–12

weeks.C

ompletion

ofIB

Q‡an

dBab

y’sDay

record

bymothe

rs.

20/22

Forestell&

Men

nella

(2012)

N=92

Exp

erim

ental

Infantswith

high

scores

onthe

approa

chdimen

sion

oftempe

ramen

tatemoreof

atest

vegetableforlong

eran

dwith

fewer

negativ

eexpression

s.

Infant

tempe

ramen

tmay

play

apa

rtin

food

acceptan

cean

dam

ount

consum

ed.

Con

sumptionisthereforeno

tpurely

determ

ined

byhu

nger.

22/22

‘Morethan

justapretty

face.

The

relatio

nshipbe

tween

infant’stempe

ramen

t,food

acceptan

ce,and

mothe

rs’pe

rcep

tions

oftheiren

joym

ento

ffoo

d’

48malean

d44

female

infants,mean

age52

weeks.

Infantsvide

o-recorded

whe

nfed

testvegetablein

labo

ratory

cond

ition

s.Measures:facial

expression

coding

;Infan

tTem

peramen

tScale

andmaterna

lratin

gsof

infants’en

joym

ent.

21/22

Gross

etal.(2010)

N=368

Cross-sectio

nal

Han

dsuckingwas

view

edas

ahu

nger

Com

mon

cues

areused

bymothe

rsto

iden

tifyhu

nger

andsatia

tion.

Low

er19/22

17/22

(Con

tinues)



Tab

le3.

(Con

tinue

d)

Autho

rsan

dtitle

Participan

tsan

dsample

Designan

dmetho

dsMainfind

ings

Implications

forun

derstand

ing

hung

eran

dsatia

tionin

infancy

Qua

lityratin

gsfirsta

ndsecond

raters

cuean

dhe

adturningas

asatia

tion

cue.Mostm

othe

rs(72%

)be

lievedcrying

mustb

ean

indicatio

nof

hung

er.M

ost(93%)

also

believedtheirba

bies

couldsensetheirow

nsatie

ty.M

othe

rswith

high

BMIs

andlow

educationa

llevels

appe

ared

less

sensitive

tosatia

tioncues.

materna

ledu

catio

nallevel

andhigh

erBMImay

beassociated

with

lower

awaren

essof

infant

satia

tion.

‘Materna

lperceptions

ofinfant

hung

er,satiety,

andpressuring

feed

ing

styles

inan

urba

nLatina

WIC

popu

latio

n’

Mothe

rs,m

eanage

28yearswith

infants

aged

<20

weeks,m

ean

infant

age18.8weeks.

Second

aryan

alysisof

survey

data

regardingmaterna

lperceptions

ofhu

nger,satiatio

nan

dpressuring

feed

ingstyle.

Hod

geset

al.(2008)

N=71

Cross-sectio

nal

Mothe

rs’respon

sivene

ssto

feed

ingcues

was

variab

le.S

omefocused

onam

ount

consum

ed,

while

othe

rsfocused

oninfant

stateor

oral

beha

viou

rs.S

pecificity

ofcues

increasedwith

infant

age.

Arang

eof

overta

ndsubtle

hung

eran

dsatia

tioncues

arerepo

rted

bymothe

rs,

e.g.crying

,licking

thelip

s,spittingfood

outa

ndstop

ping

themeal.Differen

tmothe

rsfocusedon

differen

tcue

s.

19/22

‘Materna

ldecisions

abou

tthe

initiation

andterm

inationof

infant

feed

ing’

Mothe

rsof

fullterm

infantsat

12,26or

52weeks

ofage,

35males

and36

females.M

eanmaterna

lage,28.9years.

Structured

interviewingan

dqu

alita

tiveconten

tana

lysis.

17/22

Hod

geset

al.(2013)

N=144

Cross-sectio

nal

Mothe

rsrespon

dedmore

tohu

nger

than

fulln

ess

cues.R

espo

nsiven

essto

cues

was

associated

with

materna

lcha

racteristics

(edu

catio

n,BMIan

dbreastfeed

ingdu

ratio

n).

Mothe

rsweremore

respon

sive

tohu

nger

cues

inolde

rself-feed

ing

child

ren.

Hun

gercues

may

bemoresalient

for

mothe

rsthan

satia

tioncues.M

othe

rsap

pear

morerespon

sive

tothecues

ofolde

rchild

ren.

Respo

nsiven

essto

satia

tionap

pearsto

beassociated

with

high

ered

ucationa

llevel,low

erBMI

andlong

erbreastfeed

ingdu

ratio

n.

22/22

‘Develop

men

tofthe

respon

sivene

ssto

child

feed

ingcues

scale’

Mothe

rsof

28to

104-

week-oldinfantsan

dtodd

lers,m

eanmaterna

lan

dinfant

agean

dM

:Fratio

unkn

own.

Develop

men

tand

testingof

anob

servationa

lmeasure

ofcaregiverrespon

sivene

ssto

child

feed

ingcues

usingstructured

observationof

infant/to

ddler

feed

ing.

21/22

Hwan

g(1978)

N=58

Short-term

long

itudina

lOnda

yfour,m

ean

numbe

rof

feed

ing

New

born

malean

dfemaleinfants

appe

arto

show

differen

tfeeding

14/22

15/22

(Con

tinues)



Tab

le3.

(Con

tinue

d)

Autho

rsan

dtitle

Participan

tsan

dsample

Designan

dmetho

dsMainfind

ings

Implications

forun

derstand

ing

hung

eran

dsatia

tionin

infancy

Qua

lityratin

gsfirsta

ndsecond

raters

period

swas

sign

ificantly

high

erformalethan

femaleinfants.The

first

feed

ingpe

riod

onda

yfour

was

sign

ificantly

long

erforfemales

than

males.D

uringfeed

ing

onbo

thda

ys,m

ale

infantscriedmore

than

females.

beha

viou

rs,w

ithpo

ssible

implications

formaterna

lperceptions

ofhu

nger

andsatia

tion.

‘Mothe

r–infant

interaction–effects

ofsexon

infant

feed

ingbe

havior’

Prim

iparou

smothe

rs23

malean

d35

female

newbo

rninfantsob

served

at<1week(2

and4da

ys).

Materna

lage

unkn

own.

Tim

esampled

observationof

two

sing

lebreastfeed

ingsessions

onda

ystw

oan

dfour

oflifein

hospita

lsettin

g.

Lew

&Butterw

orth

(1995)

N=18

Cross-sectio

nal

Nodifferen

cefoun

dbe

tweenthedistribu

tion

ofha

nd–facean

dha

nd–

mou

thcontactspre-feed

.Propo

rtionof

hand

–mou

thcontactswas

notg

reater

before

feed

ingthan

after

feed

ing.Ope

nmou

thpo

stures

before

hand

–

mou

thcontactson

lyoccurred

before

feed

ing.

Ope

nmou

thpo

stures

priorto

hand

mou

thcontactsmay

bean

indicatio

nof

hung

erin

newbo

rninfants.

21/22

‘The

effectsof

hung

eron

hand

–mou

thcoordina

tion

inne

wbo

rninfants’

New

born

term

infants

born

betw

een38

and

42weeks

gestationa

lageob

served

at1

weekor

youn

ger.

Structured

observations

ofinfants

before

andaftermilk

feed

ingby

form

ulaor

breast.A

nalysisof

differen

cesbe

tweenha

nd–face

andha

nd–m

outh

contacts.

20/22

Llewellynet

al.(2011)

N=2402

Coh

ortstudy

Four

appe

titeconstructs

wereiden

tified

–food

respon

sivene

ss,enjoy

men

tof

food

,satiety

respon

sivene

ssan

dslow

ness

ineatin

g.Allconstructs

hadgo

odinternal

relia

bilityan

dcorrelated

with

‘gen

eral’ap

petite.

Group

differen

ceswere

observed

inrelatio

nto

appe

titivebe

haviou

rs.

Differen

tgroup

sof

infantsha

vedifferen

tapp

etitive

beha

viou

rs;e.g.,

males

appe

arto

have

larger

appe

tites

andto

beless

satie

tyrespon

sive

than

females;p

rematureinfantsha

vesm

aller

appe

tites

andhigh

ersatie

tysensitivity

than

term

infants;an

dbreastfedinfants

appe

arless

satie

tyrespon

sive

than

form

ula-fedinfants.

21/22

‘Develop

men

tand

factor

structureof

theBab

yEating

Beh

aviour

Que

stionn

aire

intheGem

inib

irth

coho

rt’

1194

malean

d1208

femaleinfants,mean

age32.8weeks,ran

ge16–80weeks.

BEBQ

§formilk

-fed

infants.

Que

stionn

aire

itemsrefine

dvia

interviewswith

asampleof

mothe

rs(n

=10).

21/22

Llewellynet

al.(2012)

N=4634

Coh

ortstudy

Infant

weigh

twas

correlated

with

21/22

21/22

(Con

tinues)



Tab

le3.

(Con

tinue

d)

Autho

rsan

dtitle

Participan

tsan

dsample

Designan

dmetho

dsMainfind

ings

Implications

forun

derstand

ing

hung

eran

dsatia

tionin

infancy

Qua

lityratin

gsfirsta

ndsecond

raters

BEBQ

appe

titetraits.

Gen

eticinflue

ncewas

show

nforsatie

tyrespon

sivene

ss,slowne

ssin

eatin

gan

dap

petite.

Eatingtraitsof

infantsarehe

ritable.

Exp

ressionof

appe

titeistherefore

influe

nced

bygeno

type

.

‘Inh

erite

dbe

havioral

suscep

tibility

toad

iposity

ininfancy:amultiv

ariate

gene

tican

alysisof

appe

tite

andweigh

tintheGem

ini

birthcoho

rt’

2289

males

and2345

femaleinfants,

meanage32.8weeks,

rang

e16–80weeks.

BEBQ

§an

dinfant

weigh

tmeasurestake

nat

12weeks

+multiv

ariate

gene

ticmod

ellin

g.

McM

eekinet

al.(2013)

N=698

Cross-sectio

nal

Mothe

rsof

infants

with

difficult

tempe

ramen

tsrepo

rted

alower

awaren

essof

hung

eran

dsatia

tioncues

andweremorelik

ely

tousefood

tosoothe

.

Itmay

bedifficultformothe

rsof

infantswith

difficulttempe

ramen

tsto

distingu

ishhu

nger

andsatia

tioncues

from

othe

rkind

sof

distress.M

aterna

lde

pression

also

appe

arsto

beassociated

with

lower

awaren

essof

infant

feed

ingcues

andgreateruseof

food

tocalm

babies.

21/22

‘Associatio

nsbe

tweeninfant

tempe

ramen

tand

early

feed

ingpractices.A

cross-

sectiona

lstudy

ofAustralian

mothe

r–infant

dyad

sfrom

theno

urishrand

omised

controlle

dtrial’

342malean

d356

femaleinfantsbe

tween

8an

d28

weeks

ofprim

iparou

smothe

rs.

Meaninfant

age17.2

weeks.M

eanmaterna

lage30.1years.

Materna

lself-repo

rton

STSI

¶

andIFQ.∥

20/22

Men

nella

etal.(2001)

N=46

Exp

erim

ental

Infantsexpo

sedto

carrot

flavou

rsin

uteroor

during

lactationexhibited

fewer

negativ

efacial

expression

sto

carrot-

flavou

redcereal

than

plaincereal.Infan

tsexpo

sedto

carrot

flavou

rin

uterowere

perceivedby

mothe

rsto

enjoycarrot-

flavou

redcereal

morethan

plaincereal.

Previou

sexpo

sure

toflavou

rlead

sto

greateracceptan

ce,g

reater

enjoym

ent

andgreaterconsum

ption.

Amou

nteatenisno

tpurelyde

term

ined

byhu

nger.F

aciale

xpressionmay

beon

eway

ofdifferen

tiatin

gbe

tweencessation

ofeatin

gdu

eto

dislikean

dthat

arising

from

satia

tion.

21/22

‘Prena

tala

ndpo

stna

talfl

avor

learning

byhu

man

infants’

28malean

d18

female

infants.Meaninfant

age22.6weeks.

Infantsassign

edto

oneof

threegrou

psinvo

lving

drinking

carrot

juiceor

water

during

pregna

ncyan

dbreastfeed

ing.Respo

nses

tocerealscontaining

water

orcarrot

juiceweremeasuredvia

facial

expression

coding

,materna

lratings

ofen

joym

ent

andam

ount

consum

ed.

20/22

Men

nella

etal.(2009)

N=97

Exp

erim

ental

Typ

eof

form

ulafed

toinfantsim

pacted

onrespon

sesto

differen

ttastin

gcereals.Formula-fed

infantseatin

gcomplem

entary

food

s

Prior

expo

sure

lead

sto

greater

consum

ptionof

food

with

familiar

taste

compo

unds.N

egativefacial

expression

may

prov

ideaba

sisfordistingu

ishing

betw

eensatia

tionan

ddislike.

21/22

‘Early

milk

feed

inginflue

nces

tasteacceptan

cean

dlik

ing

during

infancy’

Fullterm

infantsbe

tween

16an

d36

weeks,m

ean

age25

weeks,w

hoha

dbe

enspoo

nfedba

bycereal

forat

least2

weeks.

Subg

roup

sof

breastfedan

dtw

otype

sof

form

ula-fed

babies

wereob

served

ondifferen

toccasions

tomeasure

acceptan

ceof

sweet,salty,b

itter,

savo

ury,sour

andplaincereals.

21/22

(Con

tinues)



Tab

le3.

(Con

tinue

d)

Autho

rsan

dtitle

Participan

tsan

dsample

Designan

dmetho

dsMainfind

ings

Implications

forun

derstand

ing

hung

eran

dsatia

tionin

infancy

Qua

lityratin

gsfirsta

ndsecond

raters

show

edpreferen

cesfor

thetastes

offood

sto

which

they

had

alread

ybe

enexpo

sed.

Nisbe

tt&

Gurwitz

(1970)

(Exp

erim

ent1

)Exp

erim

ental

Heavy

infantswere

morerespon

sive

than

med

ium

andlight

weigh

tinfan

tsto

sweetene

dform

ula.

Fem

aleinfants

respon

dedmoreto

sweetene

dform

ula

than

males.H

eavier

andfemaleinfants

consum

edsign

ificantly

less

inthesm

allh

ole

cond

ition

.Med

ium

weigh

t,lig

hter

weigh

tan

dmaleinfants’

consum

ptionwas

not

sign

ificantly

affected

bythiscond

ition

.

Sexan

dweigh

tmay

impa

cton

satie

tyrespon

sivene

ssto

sweetene

dmilk

.Se

xan

dweigh

tmay

impa

cton

effort

expe

nded

infeed

ingan

dconseq

uent

amou

ntconsum

ed.

18/22

‘Weigh

t,sex,an

dtheeatin

gbe

havior

ofhu

man

newbo

rns’

N=42

Infantsin

threeweigh

tgroup

swerealternatelyfedasw

eet

andstan

dard

form

ulaof

the

sameen

ergy

density

daily

atthesametim

e.Intake

perfeed

was

recorded

.Infan

tswere

form

ulafedov

er2da

yswith

ano

rmal

orsm

allh

oleteat.

Mothe

rsrecorded

consum

ption

andtim

eat

which

feed

sbe

gan

anden

ded.

17/22

22malean

d20

femalene

wbo

rninfants.

(Exp

erim

ent2

)N=34

18malean

d16

female

newbo

rninfants.

Parkinson

&Drewett(2001)

N=100

Cross-sectio

nal

Despite

similarity

intheageof

the

todd

lers

self-feed

ing

andbe

ingfedvaried

high

ly.Intak

ewas

correlated

with

numbe

rof

bitesrather

than

meald

uration.

Self-feed

ing

ledto

along

ermealtim

eon

average,while

long

ermealswereassociated

with

lower

food

intake

.

Num

berof

bitesmay

beabe

tter

indicatio

nof

hung

erlevelsthan

meal

duratio

n,althou

ghaccoun

tneeds

tobe

take

nof

whe

ther

thechild

selffeed

sor

isfedby

themothe

r.Se

lf-feed

ingtend

sto

lead

tolong

ermeald

urationan

dlower

intake

intodd

lers.

20/22

‘Feeding

beha

viou

rin

the

weaning

period

’

Mothe

rinfant

dyad

s.51

malean

d49

female.Infants/todd

lers

observed

betw

een52

and61

weeks,m

eanage

55weeks.M

aterna

lage

rang

e≤2

4to

≥35years.

Naturalistic

observationof

two

mealtim

esan

alysed

usingall

occurren

cesampling.Cod

esde

velope

dregardingmothe

rs’

feed

ingof

child

renan

dchild

self-feed

ingan

drelatedchild

beha

viou

rs.

17/22

(Con

tinues)



Tab

le3.

(Con

tinue

d)

Autho

rsan

dtitle

Participan

tsan

dsample

Designan

dmetho

dsMainfind

ings

Implications

forun

derstand

ing

hung

eran

dsatia

tionin

infancy

Qua

lityratin

gsfirsta

ndsecond

raters

Pau

leta

l.(1996)

N=20

Short-term

long

itudina

lTwo-week-oldinfants

werevisually

attentive

whe

nfeed

ing.Motor

activ

ityan

dalertness

shifted

from

pre-feed

ing

topo

st-feeding

time

during

thefirst6

mon

ths.

Motor

beha

viou

rsdiffer

with

feed

ing

statean

dat

differen

tpointsin

the

feed

ingcycleaccordingto

infant

age.

Differen

cesalso

appe

arto

existinthe

suckingbe

haviou

rsan

dconsum

ption

patterns

ofform

ula-fedan

dmilk

-fed

babies.

15/22

‘Infan

tfeeding

beha

vior:

developm

entinpa

tterns

andmotivation’

Fullterm

infantsat

2,10,18an

d26

weeks.

Sexun

know

n.

Structured

observationpriorto

during

andaftermilk

feed

ing.

Observatio

nssupp

ortedby

vide

oan

dpo

lygrap

hicrecording

ofbe

haviou

rssuch

assucking,

breathingan

dsw

allowing.

13/22

Reauet

al.(1996)

N=281

Cross-sectio

nal

Nodifferen

ceswere

repo

rted

betw

eenfeed

ing

timein

term

sof

birth

weigh

torbirthorde

r.90%

ofinfantsan

dtodd

lers

took

fewer

than

30min

toeata

meal.Rep

ortsof

feed

ing

prob

lemswereespe

cially

common

intodd

lers.

Feeding

prob

lemsarecommon

ininfantsan

despe

cially

todd

lers.

Variabilityin

hung

erisno

rmal.

Meald

urations

beyo

nd30

min

may

indicate

feed

ingprob

lems.

17/22

‘Infan

tand

todd

lerfeed

ing

patterns

andprob

lems:

norm

ativeda

taan

da

newdirection’

157malean

d124female

infantsan

dtodd

lers,

agerang

e12–108

weeks.

Meanageun

know

n.

Survey

research

usingan

unvalid

ated

self-repo

rtqu

estio

nnaire.Q

uestionn

aire

itemsinclud

edinfant

and

todd

lerhu

nger

atthestarto

fameal,feed

ingbe

haviou

rs,

feed

ingprob

lemsan

dfeed

ing

duratio

n.

14/22

Skinne

ret

al.(1998)

N=98

Lon

gitudina

lHun

gercommun

ication

appe

ared

before

satia

tion

(4.4–5.7an

d5.8–7.5mon

ths,

respectiv

ely).E

xtreme

variab

ility

was

iden

tified

incommun

icative

beha

viou

rsat

mealtim

es.F

oodlik

esan

ddislikes

increased

with

ageas

didverbal

commun

icationrelatin

gto

eatin

g.

Hun

geran

dsatia

tioncommun

ication

ishigh

lyvariab

le.L

ikes/dislik

esare

easier

todiscernin

olde

rinfants

than

youn

geron

es,alth

ough

liking

was

exhibitedless

than

dislike

throug

hfacial

expression

.

19/22

‘Mealtimecommun

ication

patterns

ofinfantsfrom

2to

24mon

thsof

age’

Infant

mothe

rdy

ads.

Typ

ically

developing

infantsfrom

8to

96weeks.Infan

tsex,

meaninfant

agean

dmean

materna

lage

unkn

own.

Structured

interviewsan

dresearcher

administered

questio

nnaire

at10

timepo

ints

from

2to

24mon

ths.

Participan

tswererand

omly

assign

edto

sixinterviews.Data

werecollected

regardinginfant

andtodd

lermealtime

commun

icationat

each

timepo

int.

15/22

Steven

sonet

al.(1990)

N=34

Cross-sectio

nal

Feeding

outcom

eswere

similarforbo

thgrou

ps.

Pre-term

infantsfussed

moredu

ring

feed

ingthan

term

infants.Vocalisations

didno

tdiffer

betw

een

Exp

ressionof

hung

ermay

differ

subtly

inprem

atureba

bies.M

othe

rsof

theseba

bies

offerfood

morein

respon

seto

vocalisationthan

those

offullterm

babies.

17/22

‘Rhy

thmsin

thedialog

ueof

infant

feed

ing:preterm

andterm

infants’

Mothe

rinfant

dyad

swith

17pre-term

infantsan

d17

full

term

infantsat

32weeks

(age

gestationa

lly

Solid

food

feed

inginteractions

vide

o-recorded

throug

haon

e-way

mirror.Cod

ingof

materna

lan

dinfant

beha

viou

rssuch

asgaze,v

ocalisationan

dself-feed

.

15/22

(Con

tinues)



Tab

le3.

(Con

tinue

d)

Autho

rsan

dtitle

Participan

tsan

dsample

Designan

dmetho

dsMainfind

ings

Implications

forun

derstand

ing

hung

eran

dsatia

tionin

infancy

Qua

lityratin

gsfirsta

ndsecond

raters

grou

ps.H

owever,m

othe

rsof

prem

atureba

bies

respon

dedto

vocalisations

with

offers

offood

morethan

mothe

rsof

term

infants.

adjusted

).Materna

lage

andinfant

sexun

know

n.

Turke

witz

etal.(1966)

N=35

Cross-sectio

nal

The

prop

ortio

nof

hand

flexionto

extension

mov

emen

tswas

greater

priorto

feed

ingthan

post-feeding

,regardless

ofwhe

ther

infantswere

awak

eor

asleep

.

Han

dflexionap

pearsto

beassociated

with

hung

erin

newbo

rninfants.

17/22

‘Relationshipbe

tweenfeed

ing

cond

ition

andorganizatio

nof

flexor–exten

sormov

emen

tsin

thehu

man

neon

ate’

New

born

femaleinfants,

aged

<1to

1.5weeks,

meanage≏1week

Observatio

nalstudy

offlexion

andextensionmov

emen

tsof

infants’ha

ndsdu

ring

two5-min

period

spriorto

andpo

st-feeding

.

19/22

vanDijk

etal.(2009)

N=20

Short-term

long

itudina

lAmou

ntconsum

edpe

rmealincreased

over

time.

Mealtimedu

ratio

nwas

stab

leacross

time

(average

8to

10min).

Freq

uencyof

food

refusalsde

creased

over

time.Variability

was

foun

din

feed

ing

beha

viou

rsbo

thacross

andwith

ininfants

particularly

during

the

period

afterthe

introd

uctio

nof

solid

s.

Infant

feed

ingbe

haviou

rishigh

lyvariab

ledu

ring

theweaning

period

;ho

wever,m

eald

urationincreases

over

time.Foo

drefusalisalso

common

during

weaning

.

19/22

‘Variabilityin

eatin

gbe

havior

throug

hout

the

weaning

period

’

12malean

d8female

fullterm

infants

aged

betw

een16

and24

weeks,

meanage22

weeks.

Naturalistic

observationof

infantsan

dcaregivers

across

a12-w

eekpe

riod

follo

wingthe

introd

uctio

nof

solid

s.Feeding

vide

o-recorded

andcode

d.

17/22

Ven

tura

etal.(2012)

N=30

Exp

erim

ental

Infantsconsum

edsign

ificantly

less

cows’

milk

form

ulaan

dshow

edhigh

ersatie

tyratio

saftertheen

hanced

cows’milk

form

ulaan

dtheproteinhy

drolysate

Formulacompo

sitio

nim

pactson

both

satia

tionan

dsatie

tyregardless

ofen

ergy

conten

t.The

stud

ypo

tentially

offers

five

means

ofiden

tifying

hung

eran

dsatie

tyin

aresearch

context:am

ount

ofmilk

initially

consum

ed,rateof

consum

ption,

respon

seto

additio

nalo

ffersof

milk

,

21/22

‘Infan

tregulationof

intake

:theeffect

offree

glutam

ate

conten

tininfant

form

ulas’

14malean

d16

female

infants,mean

age8.5weeks.

Infantswerefedon

eof

three

differen

tformulas

over

3da

ys:

cows’milk

form

ula,aprotein

hydrolysateform

ulaan

dcows’

milk

form

ulawith

adde

dfree

glutam

ate.Sa

tiety

ratio

swere

calculated

foreach

form

ula.

20/22

(Con

tinues)



Tab

le3.

(Con

tinue

d)

Autho

rsan

dtitle

Participan

tsan

dsample

Designan

dmetho

dsMainfind

ings

Implications

forun

derstand

ing

hung

eran

dsatia

tionin

infancy

Qua

lityratin

gsfirsta

ndsecond

raters

than

stan

dard

cows’

milk

form

ula.

compe

nsationfeed

ingat

subseq

uent

meala

ndsatie

tyratio

.Wasseret

al.(2011)

N=217

Cross-sectio

nal

Infantswith

high

distress

tolim

itatio

nsweremore

likelyto

receivesolid

food

searly.Materna

lob

esity

was

associated

with

earlyintrod

uctio

nof

solid

s,an

dmaterna

lde

pression

was

associated

with

theearly

introd

uctio

nof

juice.

Infantswith

difficulttempe

ramen

tsmay

bepe

rceivedto

behu

ngrier

ormay

befedto

soothe

them

.Infan

tswith

difficulttempe

ramen

tsmay

also

begivenjuiceto

soothe

them

oras

acoping

respon

seby

depressedmothe

rs.

Obe

semothe

rsmay

misinterpret

difficulttempe

ramen

tfor

hung

eror

may

have

larger,h

ungrierba

bies.

22/22

‘Infan

tspe

rceivedas

“fussy”aremorelik

ely

toreceivecomplem

entary

food

sbe

fore

4mon

ths’

Infant

mothe

rdy

ads

visitedat

12,24,

36,48an

d72

weeks

ofinfant

age.101males

and116females.M

ean

materna

lage

22.7years.

Infant

feed

ingpa

tterns

assessed

thorou

ghdietaryhistoryan

d24-h

dietaryrecall.

Infant

tempe

ramen

ttraits

measuredby

Infant

Beh

aviour

Que

stionn

aire

–revised

22/22

Wrigh

teta

l.(1980)

N=190

Short-term

long

itudina

lWhe

relong

intervals

occurred

betw

een

feed

s,breastfedinfants

consum

edalarger

meal

than

form

ula-fedinfants.

Differen

ceswereno

ted

inthesuckingpa

ttern

ofbreastfedan

dform

ula-fedinfants.

Overthefirst2

mon

ths,

diurna

ldifferen

ces

appe

ared

inthesize

offeed

consum

edin

breastfedbu

tnot

form

ula-fedinfants.

Breastfed

andform

ula-fedba

bies

show

differen

tpatternsof

feed

ing

beha

viou

rin

term

sof

sucking

beha

viou

ran

dvariab

ility

ofconsum

ption.

15/22

‘Dobreast-feeding

mothe

rskn

owho

whu

ngry

theirba

bies

are?’

132form

ula-fedan

d58

breastfedinfantsin

thefirst8

weeks

oflife.

Infant

sexun

know

n.

Video

recordingof

feed

ing

sessions

atmon

thly

intervalsfrom

justafterbirthforform

ulaan

dbreastfedinfants.Mothe

rsalso

kept

diariesof

infants’feed

s.

15/22

Wrigh

t(1986)

N=30

Short-term

long

itudina

l77%

ofmothe

rsrepo

rted

theirinfants’

hung

ervaried

across

theda

y,moreso

forbo

ysthan

girls.

Milk

intake

didno

tvary

sign

ificantly

across

theda

y,an

dno

Mostb

reastfeeding

mothe

rswereab

leto

assess

accurately

theirinfant’shu

nger.H

owever,

infant

sexmay

exerta

ninflue

nceon

mothe

rs’

interpretatio

nof

hung

ercues.M

othe

rsof

boys

may

misinterpreth

ighactiv

ityan

darou

sallevels

ashu

nger.

15/22

‘The

developm

ento

fdifferen

cesin

thefeed

ing

beha

viou

rof

form

ulaan

dbreastfedhu

man

infants

from

birthto

2mon

ths’

Mothe

rsof

14male

and16

female

breastfedinfants,

meaninfant

age4

weeks.M

eanmaterna

lageun

know

n.

Mothe

rsaske

dthreequ

estio

nsregardinginfant

hung

er.14

mothe

rsalso

kept

a4-da

ydiary

offeed

s,prov

ided

hung

erratin

gsan

dweigh

edinfants

before

andafterfeed

s.

16/22

(Con

tinues)



Tab

le3.

(Con

tinue

d)

Autho

rsan

dtitle

Participan

tsan

dsample

Designan

dmetho

dsMainfind

ings

Implications

forun

derstand

ing

hung

eran

dsatia

tionin

infancy

Qua

lityratin

gsfirsta

ndsecond

raters

sign

ificant

differen

cewas

repo

rted

betw

een

boys’an

dgirls’milk

consum

ption.

Mothe

rs’

ratin

gsof

hung

ercorrelated

with

those

forintake

for9of

the14

mothe

rs.

You

ng&

Drewett(2000)

N=30

Short-term

long

itudina

lMed

ianmeald

uration

was

17min.T

here

was

high

variab

ility

betw

eeninfants

infeed

ingbe

haviou

rsan

dacross

meals.

Refusal

was

acommon

buth

ighly

variab

lebe

haviou

r–

med

ian11,ran

ge0–101.

At5

2weeks

ofage,todd

lers’eatingbe

haviou

ris

variab

leacross

meals.F

oodrefusaliscommon

inthisagegrou

p.Tod

dlersalso

consum

ede

sserts

faster

andwith

fewer

refusalsthan

maincourses.

19/22

‘Eatingbe

haviou

ran

dits

variab

ility

in1-year-old

child

ren’

13femalean

d17

male

infantsaged

50–57weeks

old.

Meanageun

know

n.

Video

-recorde

dob

servations

ofmealsin

theho

meov

er2

consecutiveda

yscode

dwith

asche

mede

velope

dfrom

two

othe

rstud

ies.

18/22

*Child

EatingBeh

aviour

Que

stionn

aire.

†Foo

dFrequ

ency

Que

stionn

aire.

‡Infant

Beh

aviour

Que

stionn

aire.

§ Bab

yEatingBeh

aviour

Que

stionn

aire.

¶Sh

ortT

empe

ramen

tScale

forInfants.

∥ Infan

tFeeding

Que

stionn

aire.



mothers were attentive to four hunger and satiationbehaviours: hand sucking, head turning, crying andbabies ‘knowing’ they were full. The list of cues wasgenerated by the authors, although participants agreedthey used them to identify hunger and satiation. Grosset al. (2010) also found associations between certainmaternal characteristics and perceptions of feedingcues: obese mothers were less likely to agree thatbabies could sense their own satiation, and maternalobesity and longer breastfeeding history were associ-ated with perceiving hand sucking as indicating hunger.

In a study involving semi-structured interviews,Hodges et al. (2008) investigated cues that promptedmothers to initiate and end feeding. Like Andersonet al. (2001), the authors foundmothers used both infantbehaviours and external cues (e.g. time) to identify hun-ger. Commonly identified hunger cues in this studywerecrying, fussing and licking the lips, and these werereported across several age groups (3, 6 and 12months).Commonly reported satiation cues included pullingaway, spitting food out and stopping feeding. Theauthors also found that the prominence, intensity andspecificity of infant cues guided decisions about initiat-ing and ending feeds and that mothers found cues easierto interpret with increasing infant age.

In a later study, Hodges et al. (2013) described thedevelopment of the Responsiveness to Child FeedingCues Scale (RCFCS). In devising this, the authors iden-tified 20 types of hunger cue and 28 types of satiationcue. Hunger and satiation cues were further categorisedas ‘early’ (e.g. increased alertness), ‘active’ (e.g. excit-atory movements) and ‘late’ (e.g. fussing and crying)in order to reflect changes in cue intensity. Satiationcues were not described directly in the study, althoughthe authors found mothers’ responsiveness to satiationto be predicted by certain maternal characteristics(lower BMI, longer breastfeeding duration and highereducational level). They also found mothers to be moreresponsive to hunger than satiation cues.

The only longitudinal study retrieved in the searchwas conducted by Skinner et al. (1998). They examinedmealtime communication behaviours in infants andtoddlers using structured interviews with mothers.The authors found that hunger behaviours, e.g. open-ing the mouth for the spoon, appeared at a youngerage than satiation behaviours, e.g. closing the mouth

to reject food (4.4 to 5.7months vs. 5.8 to 7.5months,respectively). They also noted that overall hunger andsatiation behaviours were highly variable acrossinfants. The study also examined infants’ communica-tion of food likes and dislikes. Findings relating to thisare discussed alongside research relating to foodpreferences.

Wright (1986) also observed variability in the expres-sion of hunger by infants although this time by infantsex.Mothers of breastfed babies were asked when theirinfants were most hungry, how they identified hungerand also about the variability of their breastmilk supply.All mothers ofmale babies agreed hunger varied acrossthe day, but only around half the mothers of femalesreported this. Mothers identified increased frequencyof feeding as a hunger cue for males, whereas agitationwas cited for females. Late afternoon and early eveningwere identified as hungry times for males, whilemothers of females associated hungry times with feelingthey had less breast milk, rather than time of day.Despite such differences, recordings of infant weighttaken from before and after feeding indicated that rela-tively constant volumes of milk were consumed by girlsand boys across the day. It appears then that mothers ofmale and female infants may interpret different behav-iours as hunger depending on the sex of their child(Wright 1986).

Movement and sucking behaviours associated with hunger andsatiation

A few studies have involved observations of infantsunder controlled conditions before, during and afterfeeding. Lew & Butterworth (1995) observed hand tomouth contacts in newborns pre-prandially and post-prandially. They found that hunger did not affect wherehand contacts were made on the face, and there was nodifference between the proportion of hand–mouth con-tacts before and after feeding. However, hand–mouthcontacts preceded by open mouth postures were onlyobserved before feeding. This coordination of openmouth postures with hand–mouth contacts may there-fore be associated with hunger in newborns.

Similarly, Turkewitz et al. (1966) examined handmovements before and after feeding. The researchersobserved the flexion and extension movements of



newborns’ hands and found that regardless of whetherinfants were awake or asleep, the proportion of flexionmovements was significantly greater before feedingthan after. Flexed hand postures may therefore beanother behavioural indication of hunger in younginfants.

While Turkewitz et al. (1966) and Lew andButterworth (1995) investigated infant hand move-ments before and after feeding, Paul et al. (1996)examined several aspects of pre-prandial and post-prandial behaviour. They video-recorded milk feedsin infants at 8-week intervals in infants between 2 and26weeks of age. The researchers found sucking behav-iours increased in rate with infant age, while thenumber and length of pauses in sucking decreased.Breast and formula feeding behaviours were comparedat 2weeks of age but not beyond this; breastfed infantsconsumed milk at less than a third of the rate offormula-fed babies, and breast feeds took around fourtimes longer than formula feeds. The authors alsoexamined motor activity during feeding. This was lowfor all age groups (Paul et al., 1996). Following feeding,motor activity andmuscle tone decreased in 2-week-oldinfants. However, post-feedingmotor activity increasedin older infants. The study therefore indicates thatinfant sucking and motor activity vary with hungerand satiation, although the precise pattern of behaviourdiffers with infant age and feeding method.

Effect of infant characteristics on hunger, satiation and feedingbehaviour

Several studies have examined associations betweeninfant characteristics and feeding behaviour. Using theInfant Temperament Scale (Carey & McDevitt, 1978),Forestell andMennella (2012) investigated associationsbetween temperament and liking of a novel vegetable.They found that infants with higher ratings on‘approach’ traits (those more willing to approach novelsituations) ate more green beans, and for longer, andshowed fewer negative facial expressions (assessed bymothers) than those with lower approach ratings.

Darlington & Wright (2006) also investigated theimpact of temperament on feeding although in relationto weight gain in the first 2months of life. Using theInfant Behaviour Questionnaire (IBQ) (Rothbart,

1981), they found that infants with high fearfulnessscores exhibited slowweight gain, while those with highscores on ‘distress to limitations’ showed faster weightgain. The IBQ was also used to investigate infant tem-perament and the early introduction of complementaryfeeds by low-income mothers by Wasser et al. (2011).They noted that both ‘distress to limitations’ and infant‘activity level’ were significantly associated with theintroduction of solids before 4months of age. In addi-tion, Wasser et al. found maternal obesity to be signifi-cantly associated with the early introduction of solids,suggesting again that maternal characteristics mayinfluence perceptions of infant hunger.

Research by McMeekin et al. (2013) further sup-ports the contention that both infant and maternalcharacteristics influence perceptions of feeding cues.In a study using the Short Temperament Scale(STSI) (Sanson et al. 1987), they found thatmothers of babies with ‘difficult temperaments’ weresignificantly more likely to feed their babies to calmthem. Meanwhile, regarding maternal characteristics,mothers with higher scores on the Edinburgh PostNatal Depression Scale (Cox et al. 1987) werefound to be significantly less aware of infant feedingcues and more likely to feed their babies to calmthem.

Llewellyn et al. (2011) also explored the impactof infant characteristics on feeding behaviour. Indeveloping the Baby Eating Questionnaire (BEBQ),they examined associations between individual char-acteristics and feeding traits. Male babies werefound to have larger appetites, to respond more tofood cues and to be less satiety responsive (sensitiveto feeling full and fullness between meals) thanfemales. Premature infants were reported to havesmaller appetites, lower enjoyment of food, slowerfeeding and higher satiety responsiveness than fullterm infants. Breastfed babies had larger overallappetites, were more responsive to food cues andwere less sensitive to satiety cues than mixed-fedor formula-fed babies. Finally, infants with higherbirthweights had larger appetites, fed more quickly,enjoyed food more and were less responsive to sati-ety than lower birthweight babies. Thus, sex,birthweight and gestational age at birth may allinfluence infant appetite and feeding cues.



The BEBQ was also used by Llewellyn et al.(2012) to investigate relationships between genotypeand eating traits. In this large scale twin study,details of zygosity, infant age, gestational age andsex were collected alongside appetite data at theage of 3months. Significant shared genetic effectswere found in twins regarding weight, slowness ineating traits, satiety responsiveness and appetitesize. The findings therefore suggest that appetiteand behaviours associated with this are shaped inpart by individual genetic make-up.

Additional evidence that gestational age at birthinfluences feeding behaviour comes from researchby Stevenson et al. (1990). They observed feedingbehaviour in term and pre-term infants. No signifi-cant differences were found between groups regard-ing amount eaten, infant vocalisations or infants’gaze at mothers during feeding. However, pre-terminfants were significantly fussier during feeding thanterm infants, and mothers of pre-term infantsresponded to vocalisations with offers of food, whilemothers of term infants did not.

In relation to sex and feeding cues, an observa-tional study of newborns by Hwang (1978) foundthat on the fourth day of life, boys suckled signifi-cantly more frequently and for shorter periods thangirls. In addition, Hwang noted that, during feeding,males were significantly more likely to fuss thanfemales, both on the second and fourth days afterbirth. Nisbett & Gurwitz (1970) also reported sexdifferences in feeding behaviour although withinthe context of formula feeding. They increased thesweetness of formula fed to newborns and foundfemale and heavier infants consumed significantlymore sweetened formula than male or lighterinfants. In a second experiment, the researchersmanipulated the size of the hole in the bottle teat,alternating feeds of standard formula with a regularand a small hole. Consumption by boys was notaffected by the small hole, although that of femaleand heavier babies was reduced. The findingssuggest that female and heavier infants may bemore responsive to sweetness or, possibly, are moreable to detect this. Female and heavier infants mayalso be less willing to expend energy on feedingwhen this is made more difficult.

Infant feeding behaviour norms

A number of studies have examined normative aspectsof infant feeding such as intake and duration of feeding.These provide contextual information that is helpful inunderstanding feeding behaviour and the expression offeeding cues in infancy. In an observational study oftoddlers, Parkinson &Drewett (2001) found that meanmeal duration across two observed meals was approxi-mately 19min with a mean intake of 165g. However,within these parameters, the authors found a highdegree of variability across individuals and meals. Theyfound meal duration and intake were not significantlycorrelated, but instead, intake increased significantlywith number of bites. Number of bites therefore maybe a better indicator of level of hunger in toddlers thanmeal duration.

Infant and toddler feeding norms were also investi-gated by Reau et al. (1996). They asked mothers aboutduration and enjoyment of eating, food refusal and eat-ing speed. Mean reported feeding duration did notdiffer significantly across age, birthweight or birthorder; 90% of infants and toddlers were reported tofinish a meal in less than 30min. Food refusal, however,was commonly reported in toddlers, indicating that thisis not necessarily a satiation cue but rather a develop-mentally typical eating behaviour in toddlers.

Young & Drewett (2000) conducted observationalresearch into toddlers’ eating behaviour. Their workprovides particular insights into feeding behaviour inthe contexts of savoury and sweet courses. Medianintake for desserts and mains was similar (71 and 82 g,respectively), although median durations were 5 and10min, respectively. Furthermore, median number offood refusals for sweet courses was around half thatfor the main course, indicating that the childrenconsumed desserts more quickly and with fewerrefusals than in main courses despite already beingpartly satiated.

Other observational research into infant eating pat-terns was carried out by Van Dijk et al. (2009), in thiscase in the specific context of weaning. They foundconsiderable variability within individuals in terms offood refusal, intake and meal duration. As might beexpected, this variability was greatest in the earliestspoon feeding sessions. The average duration of mealswas relatively constant (8 to 10min across the 3-month



period observed). Consumption, however, increasedduring the first 12weeks of weaning, while refusaldecreased. This study provides further evidence thenthat developmental stage impacts on behaviours associ-ated with hunger and satiation.

The impact of milk composition and feeding method on infantfeeding behaviour

While Paul et al. (1996) and Llewellyn et al. (2011)reported incidental differences in feeding behaviouraccording to feeding method, two studies have exam-ined relationships between feedingmethod and feedingbehaviour more directly.

It has been proposed that differences betweenbreastfed and formula-fed infants in growth velocityand in the experience of hunger and satiation may beattributable in part to milk composition. Breast milkdiffers from formula in amount and form of amino acidcontent, and this may play a role in the faster weightgain recorded in infants fed cow’s milk formula com-pared with breast milk. Because free amino acids suchas glutamate are implicated in satiation in both animaland human studies, Ventura et al. (2012) manipulatedmilk composition to examine its effects on intake andsatiety. They fed infants a standard cows’milk formula,a high free glutamate formula or a cow’s milk formulafortified with free glutamate. Infants consumed signifi-cantly less of the high free glutamate formula and thefortified cows’milk formula than the regular cow’s milkformula. The authors also examined the effect of for-mula composition on satiety (determined by the effectof the first meal on later consumption). They found sig-nificantly higher levels of satiety after consumption ofthe high free glutamate formula and the fortified cows’milk formula than standard cow’s milk formula.

The impact of milk composition on feeding behav-iour is also evident in research by Wright et al. (1980)in relation to breast milk and formula milk. Theseauthors video-recorded three feeding sessions atmonthly intervals from birth to 2months in bothformula and breastfed infants. Mothers also kept feed-ing diaries. Breastfed babies exhibited pauses insucking while feeding, whereas formula-fed infantsfed almost continuously. The authors also identified di-urnal variations in the size of feed consumed bybreastfed infants, with early morning feeds being the

largest of the day. This may represent a diurnal varia-tion in breastmilk composition or in the hunger or thirstof breastfed babies; however, it was not observed informula-fed babies. Feeding method (breastfeeding orformula) therefore appears to impact both on feedingbehaviours and patterns of hunger.

Food preferences and infant feeding behaviour

Several studies suggest that hunger and satiation arenot the only drivers for infant consumption but thathedonic responses to food also play a role. Mennellaet al. (2009) investigated acceptance of cereal flavouredwith breast milk, cow’s milk formula and hydrolysedcasein formulas (HCFs) in 4- to 9-month-old infants.HCFs have stronger savoury, bitter and sour tastes thanbreastmilk or cow’s milk formula, and the investigatorsfound that infants previously fed on these ate signifi-cantlymore savoury, sour and bitter tasting cereals thanthose breastfed or fed cow’s milk formula. Mennellaet al. (2009) also assessed liking of the cereals via theFacial Action Coding System (Ekman & Friesen,2002). Infants in this study showed fewer negative facialexpressions (e.g. brow lowering, nose wrinkling andsquinting) than the other groups. Thus, enjoyment oftaste (shown through facial expression) was signifi-cantly associated with amount consumed.

Food preference research has examined flavour aswell as taste preference. Mennella et al. (2001) exam-ined liking of carrot flavour in breastfed infants ofmothers who drank carrot juice or water during preg-nancy and lactation. Infants with previous exposure tocarrot flavour in utero or through breastfeeding showedfewer negative facial expressions and greater enjoy-ment of carrot juice-flavoured cereal (rated bymothers) than those without prior exposure. Amountconsumed and meal duration showed a similar trend,but these were not significant. In terms of food flavour(rather than taste), infants may communicate liking ordislike through facial expression more than intake oreating duration.

While several studies have examined taste andflavour preferences in infancy, the impact of texturepreference on eating has been reported by Blossfieldet al. (2007). They used mothers’ ratings to assessenjoyment of chopped or pureed carrots in toddlers.



Previous experience with different textures was thestrongest predictor of enjoyment of the choppedcarrots and was also associated with amount consumed.This again suggests that amount consumed and eatingduration are driven by enjoyment as well as hunger.

As noted previously, Skinner et al. (1998) examinedthe expression of infant and toddler food preferencesalongside other mealtime communication behaviours.They did so using open-ended questions to explorehow mothers identified food preferences in theirinfants. Mothers identified behaviours such as openingthe mouth readily as the spoon approached andconsuming a large amount, as indications of liking,while dislike was judged through facial expression andbody movements (throwing food and head turning).

Discussion

The purpose of this review was to consider the evi-dence regarding infants’ feeding cues, along withfactors that affect the expression and perceptionof these. The review revealed that feeding cuesand behaviours are shaped by numerous issues. Thesecan be conceptualised in terms of individualpsychological factors, infants' physical attributes and

environmental factors (Fig. 2). It is important to note,however, that many of these factors are inter-connected; e.g., psychological factors such as foodpreference influence consumption but are influencedthemselves by environmental factors such as exposure(Mennella et al. 2001; Blossfield et al. 2005) and otherindividual psychological factors such as temperament(Forestell & Mennella 2012).

Infant feeding cues and feeding behaviour

Maternal reports indicate that mothers use many cuesto assess hunger and satiation (Skinner et al. 1998;Anderson et al. 2001; Hodges et al. 2008). Both generaland specific cues may indicate hunger and fullness(Hodges et al. 2008), and there are indications that cuesvary in form and intensity and with developmentalstage (Hodges et al. 2008; Hodges et al. 2013; Skinneret al. 1998). Notwithstanding insights from maternalreports, this literature is relatively small. In addition,the heterogeneous aims and contexts of different stud-ies confound attempts to draw simple conclusions:Anderson et al. (2001) examined the specific contextof weaning, while Hodges et al. (2013) investigatedresponsive feeding. Studies differ also in the amountof detail provided regarding feeding cues and in the

Fig. 2. Main influencing factors on feeding behaviour in the first 2 years of life (established connections in solid lines, impact of individual factors on appetiteand interactions between factors in broken lines).



methods used to investigate feeding.Only Skinner et al.(1998) and Hodges et al. (2008) provided detailedinformation regarding feeding cues and details ofdevelopmental aspects of hunger and satiationbehaviours. Meanwhile, only Skinner at al. (1998)employed a longitudinal approach, and only Hodgeset al. (2013) developed a validated tool for observinghunger and fullness cues (the RCFCS). This, however,is not primarily concerned with tracking cues so muchas measuring responsive feeding.

Observational studies of infants in controlled con-ditions suggest that different motor and suckingbehaviours are indicative of hunger and satiation(Lew & Butterworth, 1995; Turkewitz et al. 1966;Paul et al. 1996) and vary with infant age(Paul et al. 1996). Such research provides insightsregarding fine details of hunger and satiation behav-iours that are less apparent in maternal reports.However, the research in this area has limitations.Again, the literature is small. There are also methodo-logical questions with some papers failing to reportissues that may bias results, e.g. observer blindness toexperimental condition (Lew & Butterworth 1995;Turkewitz et al. 1966). Only Paul et al. (1996)observed the same infants over an extended period(i.e. months rather than days), and only they com-pared behaviours before and after feeding with thoseduring feeding.

Alongside the observational work conducted in con-trolled conditions, Young &Drewett (2000), Parkinson&Drewett (2001) and Van Dijk et al. (2009) conductedobservations of normative infant eating behaviour innaturalistic settings. Meanwhile, Reau et al. (1996)investigated normative feeding behaviour using surveymethods. It is a relative strength that the feeding normsliterature includes both observational and longitudinalenquiry. Furthermore, evidence from these studies isgenerally consistent regarding ‘gross’ aspects of feedingbehaviour such as meal duration, intake and the impactof developmental changes on feeding. They also indi-cate that behaviours such as food refusal (which mightbe perceived as satiation) are common, particularly attransition points such as weaning. Studies of infantfeeding norms therefore have implications for under-standing the contextual parameters of feeding cues ininfants. Moreover, in the case of Young & Drewett

(2000), they provide insights into the impact of differentkinds of food (savoury vs. sweet) on eating behaviour.

Individual psychological factors

An important indication from several studies is thekey role mothers play in interpreting feeding cues.Crucially, this highlights the dyadic nature of feed-ing interactions. As noted, mothers’ interpretationof cues is not based solely on infant behaviour butalso infant characteristics and external cues such astime (Anderson et al. 2001; Hodges et al. 2008).Importantly, studies indicate that hunger cues aremore salient to mothers than satiation cues (Hodgeset al. 2013). The role of mothers in interpretingfeeding cues is also evident in associations betweenmaternal characteristics and how feeding cues areperceived. In particular, maternal characteristicssuch as obesity appear to be associated with lowerresponsiveness to infant fullness (Hodges et al.2013).

Evidence suggests that infant temperament mayinfluence feeding behaviour in terms of enjoymentof novel foods or intake of food (Darlington &Wright 2006; Wasser et al. 2011; Forestell & Mennella2012; McMeekin et al. 2013). Most studies in this areahave been concerned with associations between tem-perament and weight gain or temperament and ma-ternal feeding practices. Several explanations mayaccount for these associations, making it difficult todraw simple conclusions. Darlington and Wright’s(2006) finding that infants with high distress to limita-tions gained weight quickly may be explained in rela-tion to maternal responses to these babies. Infantswith high distress to limitations were reported tosleep less and to fuss more, and may have receivedadditional feeds to comfort them. This interpretationis supported by McMeekin et al.’s (2013) finding thatmothers of difficult infants were more likely to feedthem as a soothing strategy. Alternatively, mothersin Darlington et al.’s (2006) study may have feddemanding babies more as a result of misinterpretingfractiousness as hunger. A further possibility is thatthis group of infants may simply have been more hun-gry and demanding because of rapid growth (Darlington& Wright 2006).



Darlington and Wright’s (2006) finding that infantswith high fearfulness scores showed slower weight gainis harder to explain. The authors suggest such infantsmay have difficulty expressing their needs, althoughno evidence is provided for this. The precise mecha-nisms behind associations between temperament andinfant weight therefore remain unclear. The picture isfurther complicated by findings that maternal charac-teristics may shape responses to infants withdemanding temperaments (Darlington & Wright,2006; Wasser et al. 2011). While these findingsconfound attempts to identify causal relationshipsbetween infant feeding and infant temperament, theyagain highlight the bidirectional nature of feedinginteractions.

A further difficulty in interpreting the infant tem-perament and feeding behaviour literature arisesfrom differences in study characteristics (Bergmeieret al. 2014). Different temperament measures wereused by McMeekin et al. (2013) from those usedby Darlington & Wright (2006) and Wasser et al.(2011) (the STSI and the IBQ, respectively). Fur-thermore, infants in Darlington’s study were youn-ger than those in Wasser et al.’s (2011) andMcMeekin’s (2013) research (8–12 and 8–72weeks).In addition, the cross-sectional nature of muchresearch to date limits how far conclusions can bedrawn regarding associations between infant tempera-ment and weight gain.

The impact of physical characteristics

As discussed, infant age appears to affect how feed-ing cues are expressed, while Llewellyn et al.’s(2012) twin study provides evidence that appetitivebehaviours are also determined in part by genotype.Llewellyn et al.’s (2011) large scale study of infantappetite lends credibility to the idea that character-istics such as sex and birthweight influence appetiteand therefore the expression of hunger and satia-tion. There is additional evidence that characteristicssuch as sex, birthweight or prematurity influencefeeding behaviours and potentially feeding cues(Nisbett & Gurwitz 1970; Hwang, 1978; Wright1986, Stevenson et al. 1990). Such studies involveddirect observation with appropriate procedures

taken in relation to this (inter-rater reliability andobserver blindness). This is a relative strength.However, findings from some studies have beenbrought into question by more recent research. Sim-ilar levels of breastmilk consumption by males andfemales led Wright (1986) to conclude that reportsof different hunger cues in male and female infantsarose from maternal perceptions rather than infantbehavio

Communicating hunger and satiation in the first 2 years of life: a...

Documents

Transcript of Communicating hunger and satiation in the first 2 years of life: a...