RESEARCH ARTICLE Open Access
The associa ion o mala ia mo bidi y wi h
linea g ow h, hemoglobin, i on s a us, and
de elopmen in young Malawian child en:
a p ospec i e coho s udy
Jaden Bendabenda
1,2*
, Noel Pa son
1,5
, Lo a Hallamaa
2
, John Mbo wa
6,7
, Cha les Mangani
1
, John Phuka
1
,
Elizabe h L. P ado
4
, Yin Bun Cheung
3
, Ulla Asho n
2
, Ka h yn G. Dewey
4
, Pe Asho n
2
and Kenne h Male a
1
Abs ac
Backg ound: Al hough poo complemen a y eeding is associa ed wi h poo child g ow h, nu i ion in e en ions
only ha e modes impac on child g ow h, due o high bu den o in ec ions. We aimed o assess he associa ion o
mala ia wi h linea g ow h, hemoglobin, i on s a us, and de elopmen in child en aged 6–18 mon hs in a se ing o
high mala ia and unde nu i ion p e alence.
Me hods: P ospec i e coho s udy, conduc ed in Mangochi dis ic , Malawi. We en olled six-mon hs-old in an s
and collec ed weekly da a o ‘p esumed’mala ia, dia hea, and acu e espi a o y in ec ions (ARI) un il age 18 mon hs.
Change in leng h- o -age z-sco es (LAZ), s un ing, hemoglobin, i on s a us, and de elopmen we e assessed a age
18 mon hs. We used o dina y leas squa es eg ession o con inuous ou comes and modi ied Poisson eg ession o
ca ego ical ou comes.
Resul s: O he 2723 child en en olled, 2016 (74.0%) had comple e measu emen s. The mean (s anda d de ia ion)
incidences o ‘p esumed’mala ia, dia hea, and ARI, espec i ely we e: 1.4 (2.0), 4.6 (10.1), and 8.3 (5.0) episodes/child yea .
P e alence o s un ing inc eased om 27.4 o 41.5% om 6 o 18 mon hs. ‘P esumed’mala ia incidence was associa ed
wi h highe isk o s un ing ( isk a io [RR] = 1.04, 95% con idence in e al [CI] = 1.01 o 1.07, p= 0.023), anemia (RR = 1.02,
95%CI = 1.00 o 1.04, p= 0.014) and be e socio-emo ional sco es (B = −0.21, 95%CI = −0.39 o −0.03, p= 0.041), bu no
wi h change in LAZ, haemoglobin, i on s a us o o he de elopmen al ou comes. Dia hea incidence was associa ed wi h
change in LAZ (B = −0.02; 95% CI = −0.03 o −0.01; p= 0.009), s un ing (RR = 1.02; 95% CI = 1.01 o 1.03; p=0.005),and
slowe mo o de elopmen . ARI incidence was no associa ed wi h any ou come excep o poo e socio-emo ional
sco es.
Conclusion: In his popula ion o young child en li ing in a mala ia-endemic se ing, wi h ac i e su eillance and
ea men , ‘p esumed’mala ia is no associa ed wi h change in LAZ, hemoglobin, o i on s a us, bu could be
associa ed wi h s un ing and anemia. Dia hea was mo e consis en ly associa ed wi h g ow h han was mala ia o
ARI. The indings may be di e en in con ex s whe e ac i e mala ia su eillance and ea men is no p o ided.
T ial egis a ion: NCT00945698 (July 24, 2009) and NCT01239693 (No embe 11, 2010).
Keywo ds: Child en, G ow h al e ing, Mala ia, Mo bidi y, In ec ions, S un ing, iLiNS s udies, Longi udinal s udies
* Co espondence: [email p o ec ed]
1
College o Medicine, Depa men o Public Heal h, School o Public Heal h,
Uni e si y o Malawi, Maha ma Gandhi Road, P i a e Bag 360, Blan y e 3,
Malawi
2
Facul y o Medicine and Li e Sciences, Cen e o Child Heal h Resea ch,
Uni e si y o Tampe e, Tampe e, Finland
Full lis o au ho in o ma ion is a ailable a he end o he a icle
© The Au ho (s). 2018 Open Access This a icle is dis ibu ed unde he e ms o he C ea i e Commons A ibu ion 4.0
In e na ional License (h p://c ea i ecommons.o g/licenses/by/4.0/), which pe mi s un es ic ed use, dis ibu ion, and
ep oduc ion in any medium, p o ided you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o
he C ea i e Commons license, and indica e i changes we e made. The C ea i e Commons Public Domain Dedica ion wai e
(h p://c ea i ecommons.o g/publicdomain/ze o/1.0/) applies o he da a made a ailable in his a icle, unless o he wise s a ed.
Bendabenda e al. BMC Pedia ics (2018) 18:396
h ps://doi.o g/10.1186/s12887-018-1378-2
In oduc ion
Al hough poo complemen a y eeding is associa ed wi h
poo child g ow h, many in e en ions designed o imp o e
complemen a y oods only ha e modes impac on g ow h
[1], possibly due o a high bu den o in ec ions in child en
[2,3]. S udies in which mo bidi y ea men was in eg a ed
wi h a complemen a y eeding in e en ion demons a ed
imp o ed linea g ow h [4] and de elopmen al ou comes
in child en [5,6], sugges ing he impo ance o educing
he bu den o in ec ions along wi h imp o ed die o
p omo e child g ow h and de elopmen .
Longi udinal s udies ha e epo ed a signi ican in e se
associa ion o dia hea wi h g ow h [7–9]. Howe e , s udies
on he associa ion o mala ia wi h g ow h and de elopmen
ha e ei he epo ed inconsis en esul s o had c oss-sec-
ional designs, which makes i di icul o assess causali y o
di ec ionali y o associa ion [10–13]. This has p e en ed
he inclusion o mala ia as a de e minan o s un ing in he
Li es Sa ed Tool (LiST) model [14].
The In e na ional Lipid-based Nu ien Supplemen s
(iLiNS) P ojec DOSE and DYAD-M s udies we e an-
domized con olled ials conduc ed in Malawi o s udy
he impac o lipid-based nu ien supplemen s (LNS) on
g ow h o child en [15,16]. The aim o his analysis was
o assess he associa ion o mala ia wi h linea g ow h,
hemoglobin, i on s a us, and child de elopmen . Ou
hypo hesis was ha linea g ow h, hemoglobin, i on s a us,
and de elopmen al ou comes a age 18 mo would be
poo e in child en wi h highe incidence o mala ia
om age 6 o 18 mo. We also analyzed he associa ion
o dia hea and acu e espi a o y in ec ions (ARI) wi h
linea g ow h, hemoglobin, i on s a us, and de elopmen al
ou comes.
Me hods
S udy se ing
The iLiNS-DOSE and iLiNS-DYAD-M s udies we e
conduc ed in one public dis ic hospi al (Mangochi),
one mission hospi al (S Ma ins), and wo u al public
heal h cen e s (Lungwena and Namwe a) in Mangochi
Dis ic , Sou he n Malawi. The o al ca chmen popula ion
o 180,000 la gely subsis ed on a ming and ishing. In
Malawian child en aged < 5 yea s, he p e alence o
epo ed e e (a p oxy o mala ia), dia hea and ARI
was 29, 22 and 5%, espec i ely, wi h seasonal luc ua ions
[17]. Mala ia is endemic in Malawi and he s udy a ea has
high mala ia ansmission wi h high empe a u e and e-
quen ain all om Oc obe h ough Ap il [18].
S udy design and da a collec ion
In he iLiNS-DOSE s udy, 6-mo old child en we e
andomly alloca ed o one o i e in e en ion g oups
p o ided wi h di e en doses o o mula ions o LNS
o o a con ol g oup ha did no ecei e LNS du ing
he 12-mo s udy pe iod, be ween No embe 2009 and
May 2012. In he iLiNS-DYAD-M s udy, p egnan women
< 20 weeks’ges a ion we e andomly alloca ed o one o
h ee g oups o ecei e i on and olic acid (IFA), mul iple
mic onu ien s (MMN) o a small-quan i y (20 g) o LNS
daily. A e deli e y, women in he IFA g oup ecei ed pla-
cebo able s, while MMN and LNS supplemen a ion was
con inued up o 6 mo pos pa um. Child en o mo he s in
he LNS g oup also ecei ed LNS 10 g wice daily om
age 6 o 18 mo. This s udy was conduc ed om Feb ua y
2011 o Ap il 2015. De ails o s udy design, andomiza ion
and en olmen o he wo s udies we e explained in he
main ou come pape s [15,16].
In bo h s udies, esea ch assis an s isi ed he child en’s
homes e e y week om age 6 o 18 mo o in e iew he
gua dians abou he child’s heal h in he p e ious 7 days
using a s uc u ed ques ionnai e. The in o ma ion was
complemen ed by a pic u e calenda illed ou by he
gua dians daily o aid memo y o hei child en’smo -
bidi y s a us. The use o ma e nal in e iews as a means
o collec ing da a on child mo bidi y has been alida ed
in p e ious s udies [19,20]. The esea ch assis an s
e e ed all cases o ‘p esumed’mala ia (p esence o
e e ) o he nea by heal h acili y o ea men wi h
lume an ine/a eme he , he na ionally ecommended
an imala ial d ug. The child en we e ollowed h oughou
he yea , co e ing pe iods o bo h high and low mala ia
ansmission.
An h opome ic measu emen s we e aken a age 6 mo
and 18 mo. S udy an h opome is s measu ed he in an ’s
leng h wi h a high-quali y leng h boa d (Ha penden
In an ome e ; Hol ain Limi ed) and eco ded i o he
nea es 1 mm. They weighed unclo hed in an s wi h elec-
onic in an weighing scale (SECA 735; Seca GmbH &
Co), eco ding o he nea es 10 g. The an h opome is s
we e ained and hei measu emen eliabili y was e i-
ied a he s a o he s udy and a 6-mo in e als he e-
a e wi h me hods adap ed om he p ocedu es used in
he WHO Mul icen e G ow h Re e ence S udy [21]. The
an h opome is s calib a ed all equipmen wi h s anda d
weigh s and leng h ods daily.
We assessed i on s a us a age 6 mo and 18 mo by
measu ing he zinc p o opo phy in (ZPP) concen a ion
in unwashed enous blood sample using a hema o luo om-
e e (206D, AVIV Biomedical Inc., Lakewood, NJ, USA).
Abou 5–7 ml o blood was collec ed by enepunc u e
using a 23-gauge needle in o 7.5 ml e acua ed, ace
elemen - ee polye hylene ubes con aining li hium hepa in
(Sa s ed AG & Co, Nümb ech , Ge many). The samples
we e kep co e ed in aluminium and away om ligh , in a
e ige a o o on ice, and p ocessed wi hin 2 h o collec-
ion. We measu ed blood hemoglobin (Hb) concen a ion
a age 6 mo and 18 mo om a d op o blood aken om a
inge p ick and collec ed in a mic ocu e e. Hb analysis
Bendabenda e al. BMC Pedia ics (2018) 18:396 Page 2 o 12
was conduc ed on-si e using a Hemo-Cue ins umen
(Hemocue 201+, HemoCue® AB, Ängelholm, Sweden).
We assessed ine and g oss mo o de elopmen a age
18 mo using he Kili i De elopmen al In en o y (KDI)
de eloped in Kenya [22]. Language de elopmen was
assessed using a 100-wo d ocabula y checklis by ma e nal
in e iew based on he MacA hu -Ba es Communica i e
De elopmen In en o y [23] adap ed o he local languages,
and 18-mo socio-emo ional de elopmen was assessed using
he P o ile o Social and Emo ional De elopmen (PSED),
also de eloped in Kenya. The child’s mood du ing he KDI
assessmen was a ed as posi i e (smiling/laughing) o no
posi i e (c ying/inconsolable, changeable/mood swings, o
no isible emo ions). The child’s in e ac ion wi h he asses-
so du ing he KDI was a ed as posi i e ( iendly) o no
posi i e (a oidan and wi hd awn, clings o amily membe ,
hesi an /when app oached will accep eluc an ly, di i-
cul o engage in asks, o inapp op ia e app oaches o
he assesso ). The child’s ac i i y le el du ing he KDI
was a ed as posi i e (ac i e and main ains in e es ) o
no posi i e (una ousable, sleepy and can ha dly be awak-
ened, sleepy bu easily awakened, does no spon aneously
engage in ac i i y, and awake bu loses in e es ). The KDI,
ocabula y, and PSED sco es showed high in e - a e
ag eemen and mode a e o high es - e es eliabili y in
his s udy se ing [24,25].
De ini ion o he p edic o s and he ou comes
We used a p esump i e diagnosis o mala ia de i ed om
episodes o e e du ing he p e ious week, epo ed by
he gua dians. To ensu e he diagnoses we e mu ually
exclusi e, we c ea ed an algo i hm whe eby any e e wi h
a dia hea episode ( h ee o mo e loose s ools in 24 h) was
ca ego ized as dia hea; any e e in he p esence o any
espi a o y symp oms (cough, apid o di icul b ea hing
and nasal discha ge) was ca ego ized as ARI. ‘P esumed’
mala ia was de ined as any e e episode in he absence o
dia hea and espi a o y symp oms.
An episode o ‘p esumed’mala ia, ARI o dia hea
was de ined as he pe iod s a ing om he day he child
had he symp oms when p eceded by a leas 2 days o
no symp oms o no da a. The episode ended on he las
day he child had he symp oms which was hen
ollowed by a leas 2 symp om- ee days. Incidence o
‘p esumed’mala ia, ARI o dia hea o each child om
age 6 o 18 mo was calcula ed as o al episodes / o al
ollow up yea s a isk.
Longi udinal p e alences o common childhood symp-
oms ( e e , dia hea, and cough) om age 6 o 18 mo we e
de ined as he numbe o days wi h he symp om di ided by
he o al numbe o days o obse a ion o each child [26].
We calcula ed age- and sex-s anda dized an h opome ic
indices [leng h- o -age z sco e (LAZ), weigh - o -age z
sco e (WAZ), and weigh - o -leng h z sco e (WLZ)] based
on he WHO Child G ow h S anda ds [21] and conside ed
alues below –2.0 indica i e o unde weigh , s un ing and
was ing, espec i ely. Change in LAZ o each child was
calcula ed as he di e ence be ween LAZ a age 18 mo
and LAZ a age 6 mo.
I on de iciency a age 6 mo and 18 mo was de ined as
whole blood ZPP > 70 μmol/mole heme [27]. Anemia a
age 6 mo was de ined as blood Hb concen a ion < 105 g/L
[28] while anemia a age 18 mo was de ined as blood Hb
concen a ion < 110 g/L [29].
F om he child de elopmen da a a age 18 mo, ine
mo o sco es we e calcula ed as he sum o 34 KDI ine
mo o i ems, each sco ed 0 o 1, g oss mo o sco es
we e calcula ed as he sum o 35 KDI g oss mo o i ems,
each sco ed 0 o 1 [22] and ocabula y sco e was he
ma e nal- epo ed child exp essi e ocabula y ou o he
100-wo d checklis . Fo hese ou comes, mode a e o
se e e delay was de ined as he bo om 25% o he sample.
The socio-emo ional sco e was calcula ed as he sum o
19 PSED i ems. Mode a e o se e e delay was de ined as
he op 25% o ou sample (a highe sco e indica es less
ad anced socio-emo ional de elopmen ).
S a is ical analysis
We included in he analysis child en who had ou comes
measu ed a age 18 mo. Fo all con inuous ou comes,
we used o dina y leas squa es eg ession o assess he
associa ion be ween mala ia incidence and each ou come;
and o all bina y ou comes, we used modi ied Poisson
eg ession (wi h a obus a iance es ima o ) [30].
We i s assessed whe he he ela ionship be ween he
p edic o and each ou come di e ed be ween he wo
s udies. Howe e , he in e ac ion e m was no s a is ically
signi ican indica ing ha his ela ionship was no di e -
en be ween he wo s udies he e o e we pooled da a
om he wo coho s. We hen cons uc ed mul i a ia e
models o de e mine which a iables independen ly p e-
dic ed he ou comes. We included all heo e ically ele-
an a iables, ega dless o whe he hey we e s a is ically
signi ican o no a e he bi a ia e analysis. The ollowing
a iables collec ed a age 6 mo we e included in he
models: child sex, LAZ, WLZ, Hb, i on s a us, ma e nal
educa ion and household ood insecu i y access (HFIA)
sco e gene a ed by summing he alue o esponses o
nine ques ions ega ding ood insecu i y [31]. We also in-
cluded in he models, om age 6 o 18 mo, he incidence
o dia hea and ARI, and whe he he child ecei ed an
in e en ion (LNS) o no . Fo he isk o s un ing a age
18 mo, we included in he model s un ing a age 6 mo (in
place o LAZ). In addi ion, all de elopmen al ou comes
we e adjus ed o he child’s mood, ac i i y le el, age and
in e ac ion wi h he assesso .
We assessed collinea i y among he a iables (e.g. LAZ
s WLZ a age 6 mo). I he a iables we e highly collinea
Bendabenda e al. BMC Pedia ics (2018) 18:396 Page 3 o 12
(> 0.5), we d opped he one ha was less s ongly associ-
a ed wi h he ou comes. We accoun ed o in aclus e
co ela ion due o wins using gene alised es ima ing
equa ions [32].
We also pe o med explo a o y analyses by using e-
quency o mala ia episodes ( om age 6 o 18 mo) as a ca -
ego ical a iable (no episode, one episode, and > 1 episodes
g oups). In addi ion, we conduc ed s a i ied analyses by
s un ing a age 6 mo. Al hough we pe o med bi a ia e
analyses o each indi idual a iable, we will only epo
he esul s om mul i a ia e analysis.
We used S a a e sion 14 (S a aCo p, Texas, USA) o
all he analyses.
Resul s
Baseline cha ac e is ics and desc ip i e s a is ics
O he 2723 child en en olled in he wo s udy coho s,
2016 (74.0%) had leng h measu ed bo h a age 6 mo and
18 mo (1417 child en om he iLiNS DOSE s udy and
599 child en om he iLiNS DYAD-M s udy). These we e
included in he inal analysis (Fig. 1). The cha ac e is ics
o hese child en a age 6 mo a e summa ized in Table 1.
The 2016 child en included in he inal analysis con-
ibu ed 1647.9 child yea s o ollow up, i.e. he mean
(s anda d de ia ion) [SD] du a ion o ollow up was 298
(61) days / child. A o al o 24,024 mo bidi y episodes
we e epo ed du ing he home isi s. O hese, 9.7%
(2324/24024) we e episodes o ‘p esumed’mala ia. The
es o he mo bidi y episodes we e due o: acu e espi a o y
in ec ions (ARI), 55.6% (13,360/24024); dia hea, 33.6%
(8083/24024); and mino condi ions, 1.1% (257/24024).
O e all, he mean (SD) incidence o all illnesses com-
bined was 14.8 (6.8) episodes pe child yea . The mean
(SD) incidence o ‘p esumed’mala ia was 1.4 (2.0) episodes
pe child yea . The mean (SD) incidence o ARI was 8.3
(5.0) episodes pe child yea and he mean (SD) incidence
o dia hea was 4.6 (10.1) episodes pe child yea . The
longi udinal p e alences o common childhood symp oms
( e e , dia hea, and cough) om age 6 o 18 mo we e:
7.5%; 3.4%; and 11.7%, espec i ely.
Du ing he 12-mo ollow up pe iod, 39.0% (787/2016)
o he child en did no epo any episode o ‘p esumed’
mala ia, 30.3% (611/2016) epo ed one episode and 30.7%
(618/2016) epo ed > 1 mala ia episodes. The child en
who epo ed > 1 mala ia episodes we e esponsible o
869 p egnan women en olled in comple e
ollow up, andomized o ecei e ei he LNS,
MMN o IFA
791 li e bi hs, including 5 se s o wins
599 child en wi h leng h da a a bo h
ime poin s
656 child en comple ed ollow up
- Leng h assessmen
- Zinc p o opo phy in (i on s a us)
- Haemoglobin assessmen
- De elopmen al assessmen
42 dea hs
93 d op ou s
57 missing leng h da a
ei he a age 6 o 18 mo
iLiNS DYAD S udy iLiNS DOSE S udy
2136 age-eligible in an s in i ed o he ial
o ice o a de ailed eligibili y assessmen
118 missing leng h da a
ei he a age 6 o 18 mo
1417 child en wi h leng h da a a bo h
ime poin s
2016 child en pooled
om he wo s udies
73 d op ou s
20 misca iages o
s illbi hs
110 unde age
53 o e age
7ou o ca chmen a ea
20 e used
14 easons no known
1932 in an s en olled and andomized in o 6 g oups
- Leng h assessmen
-Weigh assessmen
- Zinc p o opo phy in (i on s a us)
- Haemoglobin assessmen
791 in an s came o clinic isi a age 6 mo
- Leng h assessmen
- Weigh assessmen
- Zinc p o opo phy in (i on s a us)
- Haemoglobin assessmen
78 dea hs
319 d op ou s
Mo bidi y da a collec ion
1535 child en comple ed ollow up
- Leng h assessmen
- Zinc p o opo phy in (i on s a us)
- Haemoglobin assessmen
- De elopmen al assessmen
Included in he inal analysis
Age 18 mo
Age 6 mo
Fig. 1 Flow cha o he child en en olled and included in he inal analysis. The igu e shows he numbe o child en en olled, child en los o
ollow up, and child en who we e e en ually included in he s udy om he iLiNS DOSE and iLiNS DYAD-M coho s
Bendabenda e al. BMC Pedia ics (2018) 18:396 Page 4 o 12
73.7% (1713/2324) o all ‘p esumed’mala ia episodes
epo ed in he wo s udies.
A age 18 mo, he mean (SD) leng h- o -age z-sco es
(LAZ), weigh - o -age z-sco es (WAZ) and weigh - o -
leng h (WLZ) sco es we e −1.8 (1.1), −1.0 (1.1) and −0.2
(1.1) espec i ely. The p opo ions o child en who we e
s un ed, unde weigh and was ed we e 41.5, 16.6 and
5.0%, espec i ely. The median (25 h, 75 h cen ile) zinc
p o opo phy in (ZPP) concen a ion was 74 (51, 114)
μmole/mole heme and he p opo ion wi h i on de iciency
was 54.1%. The mean (SD) hemogobin (Hb) concen a ion
was 108.5 (15.1) g/L and he p opo ion wi h anemia was
50.5%. The mean (SD) sco es o ine mo o , g oss mo o ,
language and P o ile o Social and Emo ional De elop-
men (PSED) we e 20.9 (2.2), 17.3 (2.6), 26.2 (5.0) and 16.2
(5.4) espec i ely.
Associa ion o mala ia wi h linea g ow h
The mean (SD) change in LAZ om age 6 o 18 mo was −
0.44 (0.77). In mul i a ia e analysis, he e was no associa ion
be ween he incidence o ‘p esumed’mala ia and change in
LAZ omage6 o18mo,adjus ed o LAZa age6mo
(B = −0.02, 95% CI = −0.04 o 0.01, p= 0.069) (Table 2).
The p opo ion o child en who we e s un ed inc eased
om 27.4% a age 6 mo o 41.5% a age 18 mo. In mul i-
a ia e analysis, he incidence o ‘p esumed’mala ia was
associa ed wi h highe isk o s un ing a age 18 mo,
adjus ed o s un ing a age 6 mo (RR = 1.04, 95% CI = 1.01
o 1.07, p= 0.023) (Table 2). When ca ego ized by equency
o mala ia episodes and adjus ed o s un ing a age 6 mo,
child en wi h > 1 mala ia episodes om age 6 o 18 mo had
highe isk o s un ing a age 18 mo compa ed o child en
wi h ze o mala ia episodes (RR = 1.39, 95% CI = 1.13 o 1.70,
p= 0.002).
Associa ion o mala ia wi h hemoglobin, anemia and i on
s a us
The incidence o ‘p esumed’mala ia om age 6 o 18 mo
was associa ed wi h highe isk o anemia a age 18 mo,
adjus ed o hemoglobin a age 6 mo (RR = −0.12; 95%
CI = −0.20 o −0.04; p= 0.002) bu no wi h hemoglobin
o i on de iciency a age 18 mo (Table 3).
Associa ion o mala ia wi h child de elopmen
The associa ion o incidence o ‘p esumed’mala ia om
age 6 o 18 mo wi h child de elopmen was signi ican o
PSED sco es (B = −0.21; 95% CI = 0.39 o −0.03; p= 0.041),
bu no o he o he domains o child de elopmen ,
adjus ed o he co a ia es lis ed in he oo no es o
Tables 4and 5.
Associa ion o dia hea and ARI wi h linea g ow h,
hemoglobin, i on s a us, and de elopmen al ou comes
In mul i a ia e analysis, incidence o dia hea om age 6
o 18 mo was associa ed wi h change in LAZ om age 6
o 18 mo (B = −0.02; 95% CI = −0.03 o −0.01; p=0.009),
highe isk o s un ing a age 18 mo (RR = 1.02; 95% CI =
1.01 o 1.03; p=0.005)(Table2), lowe g oss mo o sco es
Table 1 Pa icipan cha ac e is ics a age 6 mo
Va iable DOSE
a
DYAD-M
a
Pooled da a
a
Numbe o child en
b
1417 599 2016
P opo ion o boys 50.5% 47.6% 49.6%
Mean (SD) weigh , kg 7.0 (1.0) 7.3 (1.0) 7.1 (1.0)
Mean (SD) leng h, cm 63.4 (2.4) 64.2 (2.6) 63.6 (2.5)
Mean (SD) weigh - o -age z-sco e −0.7 (1.1) −0.5 (1.2) −0.7 (1.2)
Mean (SD) leng h- o -age z-sco e −1.4 (1.0) −1.2 (1.1) −1.4 (1.1)
Mean (SD) weigh - o -leng h z-sco e 0.3 (1.1) 0.4 (1.1) 0.3 (1.1)
Mean (SD) hemoglobin, g/L 103.6 (16.1) 103.7 (15.3) 103.6 (15.9)
P opo ion wi h LAZ < −2 sco es 29.1% 23.4% 27.4%
P opo ion wi h Hb < 105 g/L 50.6% 54.4% 51.7%
P opo ion wi h ZPP > 70 μmole/mole heme
c
68.7% 69.8% 69.0%
P opo ion wi h mala ia
d
16.7% 9.7% 14.5%
Mean (SD) ma e nal educa ion, comple ed yea s 4.7 (3.6) 3.8 (3.5) 4.4 (3.6)
Mean (SD) ma e nal age, yea s 26.1 (6.2) 25.1 (6.0) 25.8 (6.1)
Mean (SD) Household Food Insecu i y Access Sco e 6.5 (6.0) 5.0 (4.3) 6.0 (5.6)
a
Values a e n, mean (SD) o p opo ions
b
Child en who had leng h da a a age 6 mo and 18 mo
c
Measu ed om unwashed enous blood
d
Measu ed by mala ia an igen Rapid Diagnosis Tes (mRDT)
Bendabenda e al. BMC Pedia ics (2018) 18:396 Page 5 o 12

a age 18 mo (B = −0.02; 95% CI = −0.03 o −0.01; p<
0.001), and highe isk o g oss mo o delay (RR = 1.01;
95% CI = 1.00 o 1.02; p< 0.001) and ine mo o delay
(RR = 1.01; 95% CI = 1.00 o 1.02; p= 0.011) a age 18 mo
(Tables 4and 5).
The incidence o ARI om age 6 o 18 mo was no
signi ican ly associa ed wi h g ow h o o he ou comes
excep o PSED sco es (B = 0.08; 95% CI = 0.03 o 0.14;
p= 0.004), and PSED delay (RR = 1.02; 95% CI = 1.00 o
1.04; p= 0.025) (Tables 4and 5).
Table 2 Associa ion o in ec ious disease mo bidi y om age 6 o 18 mo wi h change in LAZ and s un ing a age 18 mo
Va iables Mean change in LAZ om age 6 o 18 mo S un ing a age 18 mo
(N= 2016) (N= 2016)
Reg ession coe icien
a
(95%
CI)
P- alue Risk a io
b
(95% CI) P- alue
Incidence
c
o ‘p esumed’mala ia −0.02 (−0.04 o 0.01) 0.069 1.04 (1.01 o 1.07) 0.023
> 1 mala ia episodes ( s no mala ia
episode)
−0.04 (−0.12 o 0.04) 0.363 1.23 (1.02 o 1.49) 0.034
Incidence
c
o dia hea −0.02 (−0.03 o −0.01) 0.009 1.02 (1.01 o 1.03) 0.005
Incidence
c
o ARI 0.01 (−0.01 o 0.01) 0.614 0.99 (0.98 o 1.00) 0.097
O he p edic o s
d
:
Female sex ( s. male) 0.09 (0.02 o 0.16) 0.012 0.92 (0.78 o 1.09) 0.352
LAZ a age 6 mo −0.27 (−0.30 o −0.24) < 0.001 No included in he
model
No included in he
model
WLZ a age 6 mo 0.14 (0.11 o 0.17) < 0.001 0.76 (0.69 o 0.82) < 0.001
Hemoglobin (g/L) a age 6 mo 0.01 (−0.01 o 0.01) 0.943 0.98 (0.97 o 0.99) 0.009
S un ing a age 6 mo No included in he model No included in he
model
3.02 (2.72 o 3.35) < 0.001
HFIA sco e −0.01 (−0.01 o 0.01) 0.155 1.02 (1.00 o 1.03) 0.025
ARI acu e espi a o y in ec ion, CI con idence in e al, HFIA household ood insecu i y access, LAZ leng h o age z-sco e, WLZ weigh o leng h z-sco e
a
Ob ained by o dina y leas squa es eg ession
b
Ob ained by modi ied poisson eg ession (wi h a obus a iance es ima o )
c
To al episodes/child yea s a isk
d
Only p edic o s ha showed s a is ical signi icance in any o he mul i a ia e models a e p esen ed. O he a iables en e ed in he eg ession, bu no signi ican
in any model, we e: i on s a us a age 6 mo; ma e nal educa ion; and whe he he child ecei ed an in e en ion (LNS) du ing he s udy pe iod
Table 3 Associa ion o in ec ious disease mo bidi y om age 6 o 18 mo wi h hemoglobin, anemia and i on de iciency a age 18 mo
Va iables Mean hemoglobin (g/L) a age 18 mo Anemia
a
a age 18 mo I on de iciency
b
a age 18 mo
(N= 1157) (N= 1157) (N= 1707)
Reg ession coe icien
c
P- alue Risk a io
d
P- alue Risk a io
d
P- alue
(95% CI) (95% CI) (95% CI)
Incidence
e
o ‘p esumed’mala ia −0.43 (−1.21 o 0.34) 0.273 1.02 (1.00 o 1.04) 0.014 1.01 (0.99 o 1.02) 0.223
> 1 mala ia episodes ( s no mala ia episode) −1.80 (−3.74 o 0.14) 0.068 1.14 (1.02 o 1.27) 0.022 1.05 (0.95 o 1.17) 0.304
Incidence
e
o dia hea 0.01 (−0.06 o 0.08) 0.767 1.00 (0.99 o 1.01) 0.370 1.00 (0.99 o 1.01) 0.776
Incidence
e
o ARI 0.09 (−0.11 o 0.31) 0.392 1.00 (0.99 o 1.01) 0.671 1.00 (0.99 o 1.01) 0.772
O he p edic o s
:
WLZ a age 6 mo 1.10 (0.21 o 1.99) 0.015 0.98 (0.94 o 1.02) 0.248 0.99 (0.96 o 1.03) 0.811
Hemoglobin (g/L) a age 6 mo 0.21 (0.15 o 0.28) < 0.001 0.98 (0.97 o 0.99) < 0.001 1.05 (0.96 o 1.14) 0.283
ZPP > 70 μmole/mole heme a age 6 mo −0.64 (−2.42 o 1.14) 0.483 1.03 (0.92 o 1.15) 0.632 1.95 (1.71 o 2.23) < 0.001
ARI acu e espi a o y in ec ion, CI con idence in e al, WLZ weigh o leng h z-sco e, ZPP zinc p o opo phy in
a
De ined as blood hemoglobin concen a ion < 110 g/L [29]
b
De ined as whole blood ZPP > 70 μmol/mole heme [27]
c
Ob ained by o dina y leas squa es eg ession
d
Ob ained by modi ied poisson eg ession (wi h a obus a iance es ima o )
e
To al episodes/child yea s a isk
Only p edic o s ha showed s a is ical signi icance in any o he mul i a ia e models a e p esen ed. O he a iables en e ed in he eg ession, bu no signi ican
in any model, we e: leng h o age z-sco e a age 6 mo; child sex; ma e nal educa ion; household ood insecu i y access sco e; and whe he he child ecei ed an
in e en ion (lipid-based nu ien supplemen s) du ing he s udy pe iod
Bendabenda e al. BMC Pedia ics (2018) 18:396 Page 6 o 12
Table 4 Associa ion o in ec ious disease mo bidi y om age 6 o 18 mo wi h de elopmen al sco es a age 18 mo
Va iables Mean de elopmen al sco es a age 18 mo
(N= 2016)
Fine mo o sco es G oss mo o sco es Language sco es PSED sco es
Reg ession coe icien
a
P- alue Reg ession coe icien
a
P- alue Reg ession coe icien
a
P- alue Reg ession coe icien
a
P- alue
(95% CI) (95% CI) (95% CI) (95% CI)
Incidence
b
o ‘p esumed’mala ia −0.06 (−0.13 o 0.01) 0.117 0.07 (−0.17 o 0.01) 0.070 −0.01 (−0.09 o 0.06) 0.692 −0.21 (−0.39 o −0.03) 0.041
> 1 mala ia episodes ( s no mala ia episode) −0.09 (−0.34 o 0.16) 0.488 −0.17 (−0.45 o 0.12) 0.250 −0.09 (−0.35 o 0.16) 0.460 −0.48 (−1.11 o 0.15) 0.132
Incidence
b
o dia hea −0.06 (−0.11 o −0.01) 0.050 −0.02 (−0.03 o −0.01) < 0.001 0.01 (−0.05 o 0.05) 0.984 −0.02 (−0.04 o 0.01) 0.126
Incidence
b
o ARI 0.01 (−0.02 o 0.04) 0.481 0.03 (−0.01 o 0.05) 0.067 0.01 (−0.02 o 0.04) 0.432 0.08 (0.03 o 0.14) 0.004
O he p edic o s
c
Female sex ( s. male) −0.27 (−0.54 o −0.01) 0.043 −0.67 (−0.94 o −0.39) < 0.001 0.13 (−0.13 o 0.39) 0.315 −1.28 (−1.92 o −0.64) < 0.001
LAZ a age 6 mo 0.20 (0.07 o 0.33) 0.002 0.32 (0.18 o 0.45) < 0.001 0.28 (0.15 o 0.41) < 0.001 0.16 (−0.48 o 0.16) 0.322
HFIA sco e 0.04 (0.02 o 0.06) 0.001 0.01 (−0.01 o 0.03) 0.392 −0.02 (−0.04 o −0.01) 0.044 0.01 (−0.04 o 0.06) 0.719
Child’s mood 0.95 (0.70 o 1.19) < 0.001 0.92 (0.64 o 1.19) < 0.001 −0.20 (−0.44 o 0.04) 0.108 −0.75 (−1.36 o −.14) 0.016
Ac i i y le el 0.48 (0.10 o 0.88) 0.014 0.46 (0.02 o 0.90) 0.041 0.39 (−0.01 o 0.78) 0.051 0.94 (−0.04 o 1.92) 0.061
ARI acu e espi a o y in ec ion, CI con idence in e al, HFIA household ood insecu i y access, LAZ leng h o age z-sco e, PSED P o ile o Social and Emo ional De elopmen
a
Ob ained by o dina y leas squa es eg ession
b
To al episodes/child yea s a isk
c
Only p edic o s ha showed s a is ical signi icance in any o he mul i a ia e models a e p esen ed. O he a iables en e ed in he eg ession, bu no signi ican in any model, we e: weigh o leng h z-sco es a age
6 mo; hemoglobin concen a ion; i on s a us; ma e nal educa ion; in e ac ion wi h he assesso du ing he Kili i De elopmen al In en o y (KDI) assessmen ; and whe he he child ecei ed an in e en ion (lipid-based
nu ien supplemen s) du ing he s udy pe iod
Bendabenda e al. BMC Pedia ics (2018) 18:396 Page 7 o 12
Table 5 Associa ion o in ec ious disease mo bidi y om 6 o 18 mo wi h de elopmen al delay a 18 mo
Independen Va iables De elopmen al delay
a
(N= 2016)
Fine mo o delay G oss mo o delay Language delay PSED delay
Risk a io
b
P- alue Risk a io
b
P- alue Risk a io
b
P- alue Risk a io
b
P- alue
(95% CI) (95% CI) (95% CI) (95% CI)
Incidence
c
o ‘p esumed’mala ia 1.03 (0.96 o 1.09) 0.435 1.04 (1.00 o 1.09) 0.059 0.99 (0.94 o 1.06) 0.926 0.95 (0.88 o 1.02) 0.145
> 1 mala ia episodes ( s no mala ia episode) 0.96 (0.78 o 1.19) 0.723 1.06 (0.89 o 1.27) 0.508 1.00 (0.82 o 1.23) 0.971 0.82 (0.66 o 1.03) 0.094
Incidence
c
o dia hea 1.01 (1.00 o 1.02) 0.011 1.01 (1.00 o 1.02) < 0.001 1.00 (0.99 o 1.01) 0.129 1.00 (0.99 o 1.01) 0.953
Incidence
c
o ARI 0.99 (0.97 o 1.01) 0.367 0.99 (0.98 o 1.01) 0.445 0.99 (0.97 o 1.01) 0.338 1.02 (1.00 o 1.04) 0.025
O he p edic o s
d
Female sex ( s. male) 1.13 (0.91 o 1.41) 0.253 1.50 (1.22 o 1.85) < 0.001 0.92 (0.74 o 1.14) 0.438 0.77 (0.61 o 0.96) 0.023
LAZ a age 6 mo 0.90 (0.81 o 1.00) 0.059 0.82 (0.73 o 0.91) < 0.001 0.87 (0.78 o 0.97) 0.014 0.97 (0.86 o 1.08) 0.533
WLZ a age 6 mo 0.84 (0.76 o 0.93) 0.001 0.97 (0.88 o 1.07) 0.517 0.97 (0.88 o 1.07) 0.542 0.92 (0.83 o 1.03) 0.145
HFIA sco e 0.98 (0.96 o 0.99) 0.034 0.99 (0.98 o 1.01) 0.702 1.01 (0.99 o 1.03) 0.091 1.02 (0.99 o 1.04) 0.075
Child’s mood 0.38 (0.29 o 0.49) < 0.001 0.67 (0.55 o 0.83) < 0.001 1.08 (0.89 o 1.33) 0.435 0.78 (0.63 o 0.97) 0.023
Ac i i y le el 0.92 (0.67 o 1.25) 0.587 0.67 (0.51 o 0.87) 0.003 0.91 (0.65 o 1.29) 0.610 1.57 (1.12 o 2.21) 0.009
ARI acu e espi a o y in ec ion, CI con idence in e al, HFIA household ood insecu i y access, LAZ leng h o age z-sco e, PSED P o ile o Social and Emo ional De elopmen
a
De ined as he bo om 25% o ou sample
b
Ob ained by modi ied poisson eg ession (wi h a obus a iance es ima o )
c
To al episodes/child yea s a isk
d
Only p edic o s ha showed s a is ical signi icance in any o he mul i a ia e models a e p esen ed. O he a iables en e ed in he eg ession, bu no signi ican in any model, we e: hemoglobin concen a ion; i on
s a us; ma e nal educa ion; in e ac ion wi h he assesso du ing he Kili i De elopmen al In en o y (KDI) assessmen ; and whe he he child ecei ed an in e en ion (lipid-based nu ien supplemen s) du ing he
s udy pe iod
Bendabenda e al. BMC Pedia ics (2018) 18:396 Page 8 o 12
Discussion
We es ed he hypo hesis ha he linea g ow h,
hemoglobin, i on s a us, and de elopmen al ou comes a
age 18 mo would be poo e in child en wi h highe inci-
dence o ‘p esumed’mala ia. In a sample o 2016 Malawian
child en aged 6–18 mo, we ound ha mala ia was no
associa ed wi h change in LAZ, ine mo o sco es, g oss
mo o sco es, language de elopmen , i on s a us o
hemoglobin concen a ion. Highe incidence o ‘p esumed’
mala ia was associa ed wi h highe isk o s un ing and
anemia (i.e: one addi ional episode o ‘p esumed’mala ia
pe yea was associa ed wi h 4 and 2% highe isk o
s un ing and anemia, espec i ely). Highe incidence o
‘p esumed’mala ia was also associa ed wi h lowe
socio-emo ional sco es (i.e: one addi ional episode o
‘p esumed’mala ia pe yea was associa ed wi h a educ-
ion in PSED sco es by 0.21), sugges ing ha child en wi h
highe mala ia incidence ended o ha e ewe socio-emo-
ional p oblems, possibly because mala ia causes le h-
a gy and inac i i y which may mani es as ewe
beha io al p oblems.
Ou s udy had se e al s eng hs: he weekly home
mo bidi y da a collec ion o 1 yea p o ided comp ehen-
si e da a co e ing pe iods o bo h high and low mala ia
ansmission; he longi udinal design made i possible o
co ela e he mala ia exposu e wi h he ou comes and
in e p e he di ec ionali y o associa ion; and pooling o
da a om wo s udies helped us d aw conclusions om a
la ge sample. Al hough we did no calcula e pos -hoc
powe o his analysis, wi h a la ge sample size o 2016
child en and na ow con idence in e als ob ained o
mos o he mo bidi y ou comes, we belie e he s udy was
powe ed o de ec clinically meaning ul associa ions.
Ou esul s should be in e p e ed wi h cau ion because
we excluded child en who did no ha e he ou comes
measu ed a age 18 mo (26% o he sample), esul ing in
possible su i al bias. Howe e , his a i ion a e is simi-
la o ha o o he s udies wi h a long ollow up pe iod
[9,10]. Mo eo e , we expec ha he child en los o ollow
upmayha ehadwo seou comes,whichp obablywould
ha e inc eased he s eng h o associa ion in ou indings.
Ano he possible cause o bias is he p esump i e diag-
nosis o mala ia, which could lead o misclassi ica ion.
The e is o e lap in he symp oms o mala ia, dia hea and
ARI [33] which may a ec he sensi i i y and speci ici y o
p esump i e mala ia diagnosis depending on he in ensi y
o mala ia ansmission. Fo example, p esump i e mala ia
diagnosis usually has highe sensi i i y and lowe speci i-
ci y in a eas o high mala ia ansmission compa ed o
a eas o low ansmission [34,35]. Wi h he global decline
in mala ia incidence and a ailabili y o mala ia Rapid
Diagnos ic Tes s (mRDTs), he WHO in 2010 ecom-
mended an imala ial ea men be p o ided when he e is
e idence o a posi i e mala ia es esul [36]. Howe e , a
he ime o conduc ing ou s udy, mRDTs had no been
olled ou na ionwide, hence p esump i e mala ia diagno-
sis was used no only in his s udy bu also in na ional
p e alence su eys [18,37], acco ding o he p ac ice o
In eg a ed Managemen o Childhood Illness (IMCI)
[38,39]. Fu he mo e, in explo a o y analysis using
hospi al diagnosed mala ia (con i med by mRDT, albei
wi h a lo o missing da a), he di ec ion o he associa-
ions was simila , sugges ing ha ou indings a e s ill
alid (da a no shown).
Ou esea ch assis an s e e ed he child en suspec ed
o mala ia o a clinic o ea men ; he ac i e su eillance
and ea ly ea men may ha e helped imp o e he s udy
ou comes, which may ha e esul ed in unde es ima ion o
he associa ions.
I is also possible ha he associa ion o mala ia wi h
s un ing, anemia, and PSED sco es was signi ican by
chance due o mul iple es ing [40]. Howe e , we belie e
he chance inding was less likely o he signi ican associa-
ions o dia hoea and ARI wi h he ou comes because
hese associa ions we e ela i ely s ong based on p- alues.
The a ailable e idence on he associa ion o mala ia
wi h g ow h and o he ou comes is inconclusi e. Some
s udies ha e epo ed signi ican associa ions o mala ia
wi h s un ing, hemoglobin concen a ion, i on s a us and
child de elopmen [11,12,41–44]. Child en li ing in se -
ings whe e in ec ious diseases a e equen and comple-
men a y ood is o poo quali y o en ail o achie e ca ch
up g ow h a e illness episodes [45], hence equen
mala ia could be associa ed wi h g ow h al e ing. Ea lie
e idence sugges ed ha ano exia, omi ing, and a ca abolic
s a e a e esponsible o he poo g ow h associa ed wi h
eb ile illnesses in child en [46]. Howe e , o he s udies
ha e epo ed no associa ion o mala ia wi h g ow h ou -
comes [47–50]. Simila o ou s udy, mos o hese s udies
p o ided ac i e mala ia diagnosis and ea men , which
may ha e a enua ed he s eng h o he associa ion.
In he s udies ci ed abo e, di e en exposu e and ou -
come measu es we e used, which may also explain he
inconsis ency in he indings. Fo example, we de ined
linea g ow h as change in LAZ and s un ing. Change in
LAZ indica es he g ow h a e be ween wo ime-poin s
and he e o e p o ides mo e in o ma ion abou linea
g ow h al e ing han s un ing s a us assessed a one ime-
poin [51,52]. The e o e, i is possible ha he e is no
associa ion be ween mala ia and linea g ow h (based on
change in LAZ), hemoglobin o i on s a us, o he associ-
a ion is e y weak. I is also possible ha he associa ion
be ween mala ia and hese ou comes is only seen in child en
in he le end o he cu e (i.e. LAZ < −2 o hemoglobin <
110 g/L), hence he signi ican associa ion o mala ia wi h
s un ing and anemia bu no change in LAZ o hemoglobin
concen a ion. In con as , dia hea incidence was associa ed
wi h an en i e le wa d shi in LAZ and g oss mo o sco es
Bendabenda e al. BMC Pedia ics (2018) 18:396 Page 9 o 12