Severe malarial anemia is associated with long-term neurocognitive impairment Paul Bangirana ...

27 Pages·2014·0.46 MB·English

MAJOR ARTICLE
Severe Malarial Anemia is Associated With Long-
term Neurocognitive Impairment
PaulBangirana,1RobertO.Opoka,2MichaelJ.Boivin,3RichardIdro,2JamesS.Hodges,4RegildaA.Romero,6
ElsaShapiro,5andChandyC.John4,5
Departmentsof1Psychiatry,and2PaediatricsandChildHealth,MakerereUniversityCollegeofHealthSciences,Kampala,Uganda;3Departmentof
D
Psychiatry,MichiganStateUniversitySchoolofHumanMedicine,EastLansing;Departmentsof4Biostatistics,and5Epidemiology,UniversityofMinnesota o
w
SchoolofPublicHealth,and6DepartmentofPediatrics,UniversityofMinnesotaMedicalSchool,Minneapolis;and7ClinicalPsychologyAssociatesof n
lo
NorthCentralFlorida,Gainesville a
d
e
Background. Cerebral malaria (CM) is associated with long-term neurocognitive impairment in children ≥5 d fro
years of age. No prospective studies to date have assessed neurocognitive impairment in children with CM <5 m
years of age, or in children with severe malarial anemia (SMA), a form of severe malaria estimated to affect as http
maMnyetahsod5sm. illCiohnildcrheinld<re5nyeaanrnsuoafllayg.epresentingtoMulagoHospital,Kampala,Uganda,withCM(n=80)orSMA s://ac
a
(n=86)wereassessedforoverallcognitiveability,attention,andassociativememory1weekafterdischargeand6 de
and12monthslater.Thezscoresforeachdomainwerecomputedbasedonscoresof61healthycommunitychildren mic
(CC), whowerealso tested at enrollment and 6 and 12 months later. Groups were compared using mixed linear .ou
models,adjustedforage,weightforage,andchild’seducation. p.c
o
Results. At12months,childrenwithCMhadloweradjustedscoresthanCCincognitiveability(P<.001),at- m
tention (P=.02), and associative memory,(P=.002). Children with SMA had lower scoresthan CC in cognitive /cid
ability (P=.01) but not attention orassociative memory.Cognitive abilityscores in children with CM and SMA /artic
didnotdiffersignificantly. le
-a
Conclusions. In children <5 years of age, SMA is associated with long-term impairment in cognitive ability, b
s
whereas CM is associated with additional impairment in the areas of attention and associative memory. SMA tra
c
maybeamajorcontributortolong-termneurocognitiveimpairmentinchildreninsub-SaharanAfrica. t/5
9
/3
Keywords. severemalarialanemia;cerebralmalaria;cognitive;impairment. /3
3
6
/2
8
9
5
4
Severe malarial anemia (SMA) and cerebral malaria visual spatial ability and executive function [4]. 0
1
(CM) are common complications of Plasmodium fal- Seizures, hypoglycemia, prolonged coma, and elevated by
ciparum malaria. Prospective studies in Ugandan chil- cerebrospinal fluid tumor necrosis factor levels at the gu
e
dren 5–12 years of age with CM found significant time of the CM episode all seem to contribute to the st o
n
cognitive impairment 2 years after the episode [1,2]. long-termimpairment [5].Emergingevidencesuggests 1
0
Retrospective studies suggest that this impairment thatthespectrumofmalariainfection,fromasymptom- A
p
may last 8 years or longer [3,4].The primary areas atictoseveredisease,affectscognitivefunctions[6–10]. ril 2
affected are attention, memory, speech and language, Giventheeffectoncognitionofnonsevereformsofma- 01
9
lariaandtheassociationofanemiawithimpairedcog-
nition [7, 9, 11], SMA, which affects as many as 5
Received3March2014;accepted9April2014;electronicallypublished24April
million children annually [12], mayalso affect cogni-
2014.
Correspondence:ChandyC.John,MD,MS,DivisionofGlobalPediatrics,Univer- tion.SMAismorecommonthanCMandcouldsignifi-
sityofMinnesotaMedicalSchool,717DelawareStSE,Room366,Minneapolis,
cantly hinder children from achieving their full
MN55414([email protected]).
cognitivepotential.
ClinicalInfectiousDiseases 2014;59(3):336–44
©TheAuthor2014.PublishedbyOxfordUniversityPressonbehalfoftheInfectious SMA and CM both occur most frequently in sub-
DiseasesSocietyofAmerica.Allrightsreserved.ForPermissions,pleasee-mail:
SaharanAfricanchildren<5yearsofage,butnostudy
[email protected].
DOI:10.1093/cid/ciu293 to date has prospectively examined neurocognitive
(cid:129) (cid:129)
336 CID 2014:59 (1August) Bangiranaetal
outcomesinchildrenwithCM<5yearsofageorinchildrenof medical or neurologic abnormalities at screening physical
anyagewith SMA. Critical events in brain growth, including examination.
synaptogenesis and myelination, occur between birth and age Written informed consent was obtained from parents or
5years[13,14].Dependingonthelocationandseverity,injuries guardians of study participants. Ethical approval was granted
inthiscriticalperiodcouldresultineithermoresevereneuro- bytheinstitutionalreviewboardsforhumanstudiesatMaker-
cognitive impairment than injuries sustained at older ages or ere University School of Medicine, University of Minnesota,
better recovery due to neural plasticity in the younger age andMichiganStateUniversity.
group [15]. In the present study, we prospectively evaluated
the risk of neurocognitive impairment in children <5 years ClinicalandDemographicAssessment
old with SMA or CM in the year after their discharge from Allchildrenunderwentamedicalhistoryandphysicalexamina-
thehospital. tion. Children with CM were assessed for malaria retinopathy
[16]by means of indirect ophthalmoscopy. Nutrition was as-
METHODS sessedbyheight-andweight-for-agezscores(EpiInfoversion
D
3.5.3; Centers for Disease Control and Prevention). Emotional ow
n
StudyParticipants stimulationinthehomewasmeasuredbyage-appropriatever- lo
a
The study was performed at Mulago Hospital, Kampala, sionsoftheHomeObservationfortheMeasurementoftheEn- de
d
Uganda. Children with CM or SMA or community children vironment [17].Socioeconomic status was measured using a fro
m
(CC) were enrolled if they were between 18 months and 5 scoring system described elsewhere, in which lower scores h
years of age. CM was defined as (1) coma (Blantyre Coma have been associated with cognitive functioning in healthy ttp
s
Scale [BCS] score ≤2), (2) P. falciparum on blood smear, Ugandanchildren≥5yearsold[18]. ://a
c
and(3)nootherknowncauseofcoma(eg,meningitis,apro- ad
e
longed postictal state, or hypoglycemia-associated coma CognitiveAssessment m
ic
reversed by glucose infusion). SMA was defined as the Childrenweretestedeitheraweekafterdischarge(CMorSMA .o
u
presence of P. falciparum on blood smear in children with a group)oratenrollment (CC)andthen6and12monthsafter p.c
o
hemoglobin level ≤5g/dL. Children with CM or SMAwere enrollment.TheMullenScalesofEarlyLearning[19]wereused m
/c
managedaccordingtothe UgandanMinistryofHealthtreat- tomeasurecognitiveability.Scoresfromfinemotor,visualre- id
/a
mentguidelinescurrentatthetimeofthestudy.Theseinclud- ception,receptivelanguage,andexpressivelanguagescaleswere rtic
edintravenousquininetreatmentfollowedbyoralquininefor summedtogivetheearlylearningcompositescore,ameasure le-a
b
severe malaria during hospital admission and artemisinin of overall cognitive ability. Associative memory was assessed s
tra
combination therapy for outpatient follow-up therapy. All usingtheColorObjectAssociationTest[20],inwhichchildren c
children with a hemoglobin levels <5g/dL (all with SMA, 31 are required to associate toys with specific color-coded boxes t/59
/3
withCM)received abloodtransfusion. and scored on the total number of toys placed in the correct /3
3
The CC were recruited from the nuclear family, extended boxes.AttentionwasassessedusingtheEarlyChildhoodVigi- 6/2
8
family, or household compound area of children with CM or lanceTest[21],inwhichachildwasrequiredtofocushisorher 9
5
4
SMA.EligibleCCwereaged18monthsto5yearsandcurrently gazeoncartoonsscreenedonacomputerforabout7minutes; 0
1
healthy.ParentsofchildrenwithCMorSMAweregiveninfor- the measure of attention is the percentage of time the child by
g
mation about the study, asked whether any eligible children spends gazing at the screen. Neuropsychology testers were u
e
s
were present in their extended family, and requested to bring blinded to the study group (CM, SMA, or CC) of the child t o
n
theeligiblechildrentothecenterforevaluation.Parentsofchil- beingtested.PilottestinginhealthyUgandanchildrenshowed 1
0
dreninthehouseholdcompoundofachildwithCMorSMA abroadrangeoftestscoresandhighreliabilityforthesetestsin A
p
werealsonotifiedaboutthestudyduringahomevisit.Children the present study context. The few other neurocognitive tests ril 2
wereenrollediftheymetinclusioncriteriaandhadnoexclusion previously used in young African children had limitations 01
9
criteria. Exclusion criteriaforall children included (1) known that made them less suitable for use in the present study (eg,
chronicillnessrequiringmedicalcare;(2)knowndevelopmen- the Bayley Scales of Development is forchildren <42 months
tal delay; or (3) historyof coma, head trauma, hospitalization old [22]and the Malawi Developmental Assessment Test [23]
formalnutrition,orcerebralpalsy.Additionalexclusioncriteria does not assess attention and memory, which are affected in
forchildrenwithSMAincluded(1)impairedconsciousnessat CM [1]). The Mullen scales and Color Object Association
physical examination, (2) other clinical evidence of central Testhavebeensuccessfullyusedtoevaluateinterventionstoim-
nervous system disease, or (3) >1 seizure before admission. provecognitioninUgandanchildrenexposedtohumanimmu-
AdditionalexclusioncriteriaforCCincluded(1)illnessrequir- nodeficiency virus (HIV) [24]and to evaluate the effect of
ing medical care within the previous 4 weeks or (2) major maternalhealthandcaregivingonchildren’scognition[25].
(cid:129) (cid:129)
SevereMalarialAnemiaandCognition CID 2014:59 (1August) 337
StatisticalAnalyses using the SAS MIXED procedure (version 9.3; SAS Institute)
Demographic characteristics were compared using t tests for with default settings. Adjusted averages are SAS least-squares
continuous measures, Pearson χ2 test forcategorical variables, means. Other analyses were performed with JMP software
or Fisherexact test for binary variables. Age-adjusted z scores (JMPPro10.0;SASInstitute).
for cognitive outcomes were created using the scores of the
CC[1,2].Foreachoutcome,thezscorewascomputedas(ac-
tual score−mean score for child’s age)/standard deviation RESULTS
(SD), where the mean score for age and SD were computed
byfittingamixedlinearmodeltodatafromallavailablevisits ClinicalandDemographicCharacteristicsofStudyParticipants
forCC(allowingcorrelatederrorsforachild’smultiplevisits);z From November2008 to January 2012, atotal of 102 children
scores have a mean of 0 and SD 1 in the reference population withCM,102withSMA,and64CCwererecruited.Asshown
(CC)overalltimepoints.Groupswerethencomparedaccord- in Figure 1, 13 children met exclusion criteria, 14 died during
ing to z scores using all 3 testing times and analyzed using a theinitialadmission,4diedduringfollow-up,5werewithdrawn
D
mixed linear model, which generalizes repeated-measures fromthestudy,and5werelosttofollow-up.Eightychildrenwith ow
analysis(againallowingcorrelatederrorsforachild’smultiple CM,86withSMA,and61CCunderwentneurocognitivetesting nlo
a
visits).BecauseCCdifferedfromchildrenwithCMorSMAin atbaselineandat6and12months.Of681possiblechildvisits, de
d
age,weight-for-ageandheight-for-agezscores,andchildedu- dataweremissingfor33forattention,14forassociativememory, fro
cational level, neurocognitive z scores between groups were and5foroverallcognitiveability.Forty-threechildren(54%)with m
h
compared with adjustment for these characteristics. Sample CMhadretinopathy,and37(46%)didnot. ttp
s
sizes of 60 CC and 80 children in each of the CM and SMA Demographic,socioeconomic,clinical,andlaboratoryfindings ://a
groups were estimated a priori to give >80% power (α=.05) inthestudychildrenaresummarizedinTable1.Childrenwith ca
d
to detect a 0.5 difference in z scores between each disease SMAwereyoungerthanCCandhadlowerweight-for-agezscores em
ic
groupandCC. thanCC.ComparedwiththeCC,fewerchildrenwithCMorSMA .o
u
For the CM and SMA groups separately, z scores for each had attended preschool. The 3 groups were otherwise similar p.c
o
cognitive domain were tested forassociation with clinical fea- (Table1).HemoglobinSS(1inCMand4inSMAgroup),HIV m
/c
tures using Pearson correlation for continuous variables and infection(2inCMand1inSMAgroup),andstoolhelminthin- id
/a
2-sample t tests for binary variables. To account for multiple fection (Table 1) were uncommon. Mean hemoglobin levels rtic
testing, differences were considered statisticallysignificant for (±SD) were lower in children with SMA (3.78±0.85 g/dL) le-a
these tests at P<.01 . Mixed linear models were estimated thaninthosewithCM(5.99±2.06g/dL;P<.001). bs
tra
c
t/5
9
/3
/3
3
6
/2
8
9
5
4
0
1
b
y
g
u
e
s
t o
n
1
0
A
p
ril 2
0
1
9
Figure1. Studydesignandfollow-up.
(cid:129) (cid:129)
338 CID 2014:59 (1August) Bangiranaetal
Table1. DemographicandClinicalCharacteristicsofStudyChildren
Characteristic CM(n=80) SMA(n=86) CC(n=61) PValuea
Age,mean±SD,y 2.8±0.6 2.5±0.6 2.9±0.7 <.001b
Femalesex,No.(%) 29(36) 36(42) 27(44) .60
Weightforagezscore,mean±SD −1.2±1.2 −1.7±1.7 −0.7±1.3 <.001c
Heightforagezscore,mean±SD −0.5±1.3 −1.2±1.4 −1.0±1.4 .011d
Socioeconomicstatusscore,mean±SD 8.4±2.5 9.3±3.4 9.5±3.2 .06
Homeenvironmentzscore,mean±SD −0.1±1.0 −0.1±0.8 0.0±1.0 .73
Maternaleducationallevel,No.(%)
Primary6orlower 37(46) 31(36) 24(39) .86
Primary7 16(20) 17(20) 12(20)
Secondaryorhigher 25(31) 34(40) 22(36)
Notknown 2(2) 4(5) 3(5) D
o
Paternaleducationallevel,No.(%) w
n
Primary6orlower 15(19) 25(29) 9(15) .37 loa
d
Primary7 15(19) 7(8) 12(20) e
d
Secondaryorhigher 33(41) 37(43) 26(43) fro
m
Notknown 17(21) 17(20) 14(23) h
Childwithpreschooleducation,No.(%) 6(8) 2(2) 14(23) .001e ttp
s
HemoglobinSS,No./totaltested,(%) 1/77(1) 4/83(5) 0/61(0) .23 ://a
HIVpositive,No./totaltested(%) 2/79(3) 1/86(1) 0/61(0) .63 ca
d
Stoolhelminth,No.(%)b 6/65(9) 3/60(5) 3/21(14) .36 em
ic
Abbreviations:CC,communitychildren;CM,cerebralmalaria;HIV,humanimmunodeficiencyvirus,SD,standarddeviation;SMA,severemalarialanemia. .o
u
aPvaluesbasedon1-wayanalysisofvarianceforcontinuousvariablesandPearsonχ2testforcategoricalvariablesotherthanhemoglobinSS,HIV,andstool p.c
helminthinfection,whichwerecomparedusingFisherexacttest. o
m
bSMAgroupdiffersfromCCandCMgroups. /c
cSMAgroupdiffersfromCCgroup. id/a
dSMAgroupdiffersfromCMgroup. rtic
eCMandSMAgroupsdifferfromCCgroup.Tukeytestwasusedtocomparecontinuousmeasures;forchild’seducation,Pearsonχ2testwasusedtocomparepairs le-a
ofgroups,withBonferroniadjustment. b
s
tra
c
t/5
9
/3
Long-termNeurocognitiveOutcomesinChildren cognitive ability and attention did not differ significantly /3
3
WithCMandSMA betweenchildrenwithCMandSMA(P>.05forboth),butchil- 6/2
Gross neurologic deficits were present in 43 children (51%) drenwithCMhadsignificantlylowerassociativememoryscores 895
4
with CM at discharge but had largely resolved at 6-month thanthosewithSMA(P=.001).Comparingadjustedscoresbe- 0
1
(4 children, 5%) and 12-month (4 children, 5%) follow-up. tweengroupsusingall3testingtimepoints,childrenwithCM by
g
However,at12-monthfollow-up,childrenwithCMhadsignif- hadlowerscoresthanCCinoverallcognitiveability,attention, u
e
icantlylowerage-adjustedzscoresthanCCforcognitiveability, andassociativememory,whereaschildrenwithSMAhadlower st o
n
attention, and associative memory (Figure 2). Children with scores than CC in overall cognitive ability and attention 1
0
SMAhadloweradjustedzscoresthanCCforoverallcognitive (Table 2). Analyses were also conducted excluding children A
p
abilityandshowedatrendtowardlowerscoresforattentionbut with any preschool education, children with hemoglobin SS ril 2
didnotdifferforassociativememory(Figure2).Themagnitude or children with HIV infection. Differences between groups 01
9
of difference in cognitive ability or attention scores between remained significant after these groups were excluded
childrenwith CMorSMA compared withCC did not change (SupplementaryTables1–3).
significantlyoverthestudyperiod(Figure2;P>.70fortestsof
theinteractionofstudygroupandvisitforall3outcomes).The ClinicalFeaturesAssociatedWithNeurocognitiveOutcomes
differences in age-adjusted neurocognitive z scores between Clinical features assessed as risk factors for adverse neuro-
children with CM or SMA and CC remained after additional cognitive outcomes for children with CM or SMA included
adjustment for weight-for-age and height-for-age z scores, hypoglycemia(glucose<2.2mmol/L),daysoffever,useofan-
andchildpreschooleducation,theonlyfactorsthatdifferedsig- timalarialmedicationsbeforeadmission,priorseizurehistory,
nificantly between study groups (Table 2). Adjusted scores for prior hospitalization, malnutrition, admission temperature,
(cid:129) (cid:129)
SevereMalarialAnemiaandCognition CID 2014:59 (1August) 339
D
o
w
n
lo
a
d
e
d
fro
m
h
ttp
s
://a
c
a
d
e
m
ic
.o
u
p
.c
o
m
/c
id
/a
rtic
le
-a
b
s
tra
c
t/5
9
/3
/3
3
6
/2
8
9
5
4
0
Figure2. Age-adjustedneurocognitivetestscoresinchildrenwithcerebralmalaria(CM)orseveremalarialanemia(SMA)andincommunitychildren(CC) 1 b
atbaselineandat6and12monthsforcognitiveability(A),attention(B),andassociativememory(C).Scoresaregivenasmeanswithstandarderrorsofthe y g
u
mean. e
s
t o
n
1
0
A
p
ril 2
systolic and diastolic blood pressure, hypoxia (pulse oxygen InchildrenwithSMA,noclinicalfeatureswereassociatedwith 01
9
saturation<92%), deeprespirations, lacticacidosis(lacticacid worsened neurocognitive outcomes. In children with CM, de-
level ≥5mmol/L), hyperparasitemia (>250000parasites/µL), creasedcognitiveabilitywas seeninchildrenwithlongercoma
hemoglobin level, white blood cell count, platelet count, and duration,lowerBCSscore,moreseizures,orneurologicdeficits
presenceofbacteremia.ForchildrenwithCM,additionalclin- at 6- or 12-month follow-up. Decreased attention was seen in
ical features included presence and numberof seizures before childrenwithlongercomadurationorneurologic deficits at 6-
admission, BCS score, coma duration, abnormal posturing, monthfollow-up(Tables3and4).Thepresenceofretinopathy
presence and numberof seizures afteradmission, and abnor- didnotmodifytheeffectofCMforanyneurocognitiveoutcome
mal neurologic examination findings at discharge or 6- or (allP>.05),andchildrenwithorwithoutretinopathyhadsimilar
12-monthfollow-up. zscoresforallcognitiveoutcomes(allP>.23).
(cid:129) (cid:129)
340 CID 2014:59 (1August) Bangiranaetal
Table2. NeurocognitiveOutcomezScoresinChildrenWithCMorSMAandinCCat12-MonthFollow-upandforAllStudyVisitsa
CM−CC SMA−CC
ScoreandOutcome CM SMA CC Differenceb PValue Differenceb PValue
zScoreat12mo,mean±SE
Cognitiveability −0.96±0.13 −0.63±0.13 −0.11±0.16 −0.85±0.21 <.001 −0.52±0.21 .01
Attention −0.45±0.12 −0.32±0.12 0.01±0.14 −0.46±0.19 .02 −0.33±0.19 .08
Associativememory −0.52±0.09 −0.09±0.08 −0.10±0.10 −0.41±0.14 .002 0.02±0.13 .89
zScoreforallvisits,mean±SEc
Cognitiveability −0.94±0.12 −0.57±0.12 −0.10±0.14 −0.85±0.19 <.001 −0.48±0.19 .01
Attention −0.47±0.09 −0.25±0.09 0.02±0.11 −0.49±0.14 <.001 −0.27±0.14 .05
Associativememory −0.39±0.06 −0.08±0.06 −0.02±0.07 −0.38±0.09 <.001 −0.07±0.09 .49
Abbreviations:CC,communitychildren;CM,cerebralmalaria;SE,standarderror;SMA,severemalarialanemia. D
aAdjustedforage,nutrition,andchildeducationallevel.Age-adjustedneurocognitivezscoreswerecalculatedusingCCasthecomparatorgroup.Nutritionwas ow
adjustedforusingforweight-for-ageandheightforagezscores. nlo
bEstimatedmeandifference±SE. ad
e
cMixedlinearmodelestimatesofmeanzscoresatenrollmentand6-and12-monthvisits.Themixedlinearmodelanalysisisageneralizationofrepeated-measures d
analysis. fro
m
h
ttp
s
://a
c
a
DISCUSSION contributortolong-termneurocognitiveimpairmentinAfrican de
m
children. ic
Toourknowledge,thepresentstudydemonstratesforthefirst Previousprospectivestudieshavedocumentedthatalthough .oup
timethatSMA,aformofseveremalariathataffectsbetween1.5 gross neurologic deficits in children with CM largely resolve .co
m
and 5 million children in sub-Saharan Africa annually [12], is overtime[1,26],CMisassociatedwithlong-termneurocogni- /c
id
associatedwithlong-termimpairmentinoverallcognitiveabil- tiveimpairment [1,2],neurologicdisability[27],and develop- /a
ity.Thezscoresforcognitiveabilityat12-monthfollow-up,ad- mental deficits [28] in children ≥3 years of age. The present rticle
justedforpotentialconfoundingfactors,werealmostafullSD study confirms these findings in children with CM as young -ab
(−0.85)lowerinchildrenwithCMcomparedwithCC,andhalf as 18 months and adds to these findings a new finding of stra
aSDlower(−0.52)inchildrenwithSMA,analogoustoadiffe- long-term cognitive impairment in children with SMA. In the ct/5
9
renceof13and8IQpointsasaresultofCMandSMA,respec- present study, children with CM had significantly reduced /3
/3
tively.GiventhelargenumberofchildrenwhohaveSMAeach memory scores, which were not seen in children with SMA, 36
yearandthesizeoftheeffectonoverallcognitiveabilityinchil- andoftheclinicalfactorsmeasured,thoseassociatedwithneu- /28
9
dren with SMA in the present study, SMA may be a major rocognitive deficits occurred only in children with CM (eg, 54
0
1
b
y
g
u
e
s
Table3. CategoricalVariablesSignificantlyAssociatedWithCognitiveAbilityorAttentioninChildrenWithCerebralMalaria t o
n
1
0
CognitiveAbility Attention Ap
Variable Group Mean(SE) PValuea Group Mean(SE) PValuea ril 20
1
BCSscoreb 1(n=17) −2.04(0.37) <.001 . . . . . . . .. 9
2(n=61) −0.52(0.20)
Neurologicdeficitat6mo Yes(n=4) −6.16(0.55) <.001 Yes(n=4) −1.86(0.47) .002
No(n=75) −0.59(0.1) No(n=72) −0.33(0.11)
Neurologicdeficitat12mo Yes(n=3) −6.78(0.67) <.001 . . . . . . . ..
No(n=76) −0.64(0.13) . . . . . .
Abbreviations:BCS,BlantyreComaScale;SE,standarderror.
aGroupswerecomparedusingttestsorPearsoncorrelationtozero.
bIn2childrentheBCSscorewasrecordedas≤2,buttheexactscorewasnotrecorded.TheBCSscorewas1or2inallchildrenwitharecordedscore(nonehada
scoreof0).
(cid:129) (cid:129)
SevereMalarialAnemiaandCognition CID 2014:59 (1August) 341
Table 4. Continuous Variables Significantly Associated With reducing long-term neurodevelopmental sequelae. Further
CognitiveAbilityorAttentioninChildrenWithCerebralMalaria studieswillberequiredtoassesswhetherimpairmentworsens
oversubsequentyears.Comadurationandnumberofseizures
CognitiveAbility Attention were strongly associated with cognitive outcomes in children
Pearson P Pearson P with CM, suggesting that more effective seizure prevention
Variable Correlation(r) Valuea Correlation(r) Valuea couldbeaneffectiveinterventiontodecreasecognitiveimpair-
Comadurationafter −0.59 <.001 −0.45 <.001 ment. Prior studies attempting to reduce seizures in children
admission(hours) with CM with phenobarbital led to increased mortality [37],
No.ofseizuresduring −0.68 <.001 . . . . ..
but new antiepileptic medications such as levetiracetam,
admission
whichdo not lead torespiratorydepression [38], mayprovide
aPvalueforcomparisonofPearsoncorrelationcoefficientwith0. a better option to prevent seizures and possibly reduce long-
termneurocognitiveimpairmentinCM.
Manyfactorscanaffectcognitioninchildren,andinstudies
D
comaandseizures).Thesefindingssuggestthatthepathwaysby ofseveremalariaitisimpossibletoconductin-depthneurocog- ow
n
which SMA and CM lead to neurocognitive impairment may nitive testing before the disease episode. We attempted to ad- lo
a
differanddemonstratetheimportanceoftestingmultipleneu- dress this limitation by excluding children with known prior de
d
rocognitive domains when assessing neurodevelopmental se- neurologicdisability;bychoosingCCfromtheextendedhouse- fro
m
quelaeofsevereillness. hold or neighborhood of children with CM or SMA, to mini- h
Inpreviousstudies,cerebrospinalfluidtumornecrosisfactor mize differences between groups in education, socioeconomic ttp
s
levelswereassociatedwithneurocognitiveimpairmentinatten- status,andmalariaexposure;andbyconductingthemostcom- ://a
c
tionandworkingmemoryinchildrenwithCM[29],suggesting prehensive assessment to date of other factorsthat mayaffect ad
that central nervous system inflammation may play a role in cognitive outcomes, including parental and child educational em
ic
neurocognitive impairment in children with CM. A prior levels,childnutrition,childhomeenvironment,familysocioe- .o
u
p
study demonstrated no difference in neurologic outcomes in conomicstatus,andpresenceofHIVorhemoglobinSS.Differ- .c
o
children with CM with vs without retinopathy [30]. We ences between groups in neurocognitive outcome scores were m
/c
found,similarly,nodifferences inneurocognitiveoutcomes in notalteredbyadjustingpotentialconfoundingfactorsthatdif- id
/a
children with CM with or without retinopathy, although our feredbetweenchildrenwithCMorSMAandCCorafterexclu- rtic
studylackedthepowertodetectsmalldifferencesinneurocog- sionofchildrenwithfactorsthatmightstronglyaffectcognition le-a
b
nitive outcomes. Cognitive impairment in children with SMA (eg,HIV,hemoglobinSS,andpreschooleducation). s
could be due to iron deficiencyor metabolic acidosis, both of In summary, the present study demonstratesthat SMA and trac
t/5
whicharecommoninSMA[31,32]andareassociatedwithcog- CMareassociatedwithneurocognitiveimpairmentinchildren 9
/3
nitive impairment in other disease processes [5,33].Cognitive <5 years of age. SMA affects overall cognitive ability, whereas /3
3
impairmentinchildrenwithSMAcouldalsobeduetothesys- CMaffectsmultipleareasofneurocognitivefunction,including 6/2
temicinflammationseeninSMA[34]ortoadirecteffectoflow cognitive ability, attention, and associative memory. Further 89
5
hemoglobin on cognition [11].The present study did not find study is required to determine the mechanisms by which CM 40
1
anassociationbetweencognitiveoutcomesandeitherhemoglo- and SMA lead to neurocognitive impairment, because this by
g
binlevelorlacticacidosisinchildrenwithSMAorCM,butbase knowledge could lead to adjunctive neuroprotective therapy u
e
deficit,whichwedidnotmeasure,maybeabettermeasureof for children with severe malaria and decrease the long-term st o
metabolicacidosisinseveremalaria[35],andasinglehemoglo- neurocognitivesequelae.Thesefindingsprovidestrongfurther n 1
binvaluemaynotadequatelyreflecthemoglobinlevelsoverthe rationale for malaria elimination programs, because they indi- 0 A
p
length of follow-up. Iron deficiency was not assessed in this catethateliminatingmalariamaysignificantlydecreasethebur- ril 2
study, but memory deficits, which are common in children den of neurocognitive impairment in sub-Saharan African 01
9
with iron deficiency [36], were not present in children with children.
SMAinthepresentstudy,suggestingthatadditionalpathways
maybeinvolvedinneurocognitiveimpairmentinSMA.
SupplementaryData
Differences between study groups in neurocognitive scores
did not change significantlyover the 1-year period of evalua- SupplementarymaterialsareavailableatClinicalInfectiousDiseasesonline
tion, indicating that cognitive damage associated with CM or (http://cid.oxfordjournals.org).Supplementarymaterialsconsistofdata
SMA occurs early and does not change greatly in the first providedbytheauthorthatarepublishedtobenefitthereader.Theposted
yearaftertheepisode.Thesefindingssuggestthatinterventions materialsarenotcopyedited.Thecontentsofallsupplementarydataarethe
soleresponsibilityoftheauthors.Questionsormessagesregardingerrors
given early in the treatment course may be most effective in shouldbeaddressedtotheauthor.
(cid:129) (cid:129)
342 CID 2014:59 (1August) Bangiranaetal
Notes DevelopmentalCognitiveNeuroscience,2nded.MassachusettsInsti-
tuteofTechnology,2008:213–36.
Acknowledgments. Wethankthechildrenandtheirparentswhopar- 15. AndersonV,Spencer-SmithM,WoodA.Dochildrenreallyrecoverbet-
ticipatedinthisstudy,thestudyteamfortheirdedicatedeffortintreating ter?neurobehaviouralplasticityafterearlybraininsult.Brain2011;
thechildrenandcollectingthedata,andBriannaYundforscoringtheEarly 134:2197–221.
ChildhoodVigilanceTestvideos. 16. BirbeckGL,BeareN,LewallenS,etal.Identificationofmalariaretinop-
Disclaimer. Thecontentissolelytheresponsibilityoftheauthorsand athyimprovesthespecificityoftheclinicaldiagnosisofcerebralmalaria:
doesnotnecessarilyrepresenttheofficialviewsoftheNationalInstitutesof findingsfromaprospectivecohortstudy.AmJTropMedHyg2010;
Health. 82:231–4.
Financialsupport. ThisworkwassupportedbytheNationalInstitute 17. CaldwellBM,BradleyRH.Homeinventoryadministrationmanual.3rd
ofNeurologicalDisordersandStrokeandtheFogartyInternationalCenter ed.LittleRock,AR:UniversityofArkansas,2001.
(grantsR01NS055349andD43NS078280). 18. BangiranaP,JohnCC,IdroR,etal.Socioeconomicpredictorsofcog-
Potentialconflictsofinterest. Allauthors:Noreportedconflicts. nitioninUgandanchildren:implicationsforcommunityinterventions.
AllauthorshavesubmittedtheICMJEFormforDisclosureofPotential PLoSOne2009;4:e7898.
ConflictsofInterest.Conflictsthattheeditorsconsiderrelevanttothecon- 19. MullenEM.MullenScalesofEarlyLearning:AGSedition.American
tentofthemanuscripthavebeendisclosed. GuidanceServices,Inc,CirclePines,MN.1995.
20. JordanCM,JohnsonAL,HughesSJ,ShapiroEG.TheColorObjectAs- D
o
sociationTest(COAT):thedevelopmentofanewmeasureofdeclara- w
References tivememoryfor18-to36-month-oldtoddlers.ChildNeuropsychol nlo
2008;14:21–41. ad
e
1. BoivinMJ,BangiranaP,ByarugabaJ,etal.Cognitiveimpairmentafter 21. GoldmanDZ,ShapiroEG,NelsonCA.Measurementofvigilancein2- d
cerebralmalariainchildren:aprospectivestudy.Pediatrics2007;119: year-oldchildren.DevNeuropsychol2004;25:227–50. fro
e360–6. 22. CarloWA,GoudarSS,PashaO,etal.Randomizedtrialofearlydevel- m
h
2. JohnCC,BangiranaP,ByarugabaJ,etal.Cerebralmalariainchildrenis opmentalinterventiononoutcomesinchildrenafterbirthasphyxiain ttp
associatedwithlong-termcognitiveimpairment.Pediatrics2008;122: developingcountries.JPediatr2013;162:705–12e3. s
e92–9. 23. GladstoneM,LancasterGA,UmarE,etal.TheMalawiDevelopmental ://a
c
3. CarterJA,RossAJ,NevilleBG,etal.Developmentalimpairmentsfol- AssessmentTool(MDAT):thecreation,validation,andreliabilityofa ad
lowingseverefalciparummalariainchildren.TropMedIntHealth tooltoassesschilddevelopmentinruralAfricansettings.PLoSMed em
2005;10:3–10. 2010;7:e1000273. ic
4. KiharaM,CarterJA,NewtonCR.TheeffectofPlasmodiumfalciparum 24. Boivin MJ, Bangirana P, Nakasujja N, et al. A year-long care- .ou
p
on cognition: a systematic review. Trop Med Int Health 2006; giver training program to improve neurocognition in preschool .c
11:386–97. Ugandan HIV-exposed children. J Dev Behav Pediatr 2013; om
5. IdroR,MarshK,JohnCC,NewtonCR.Cerebralmalaria:mechanisms 34:269–78. /c
id
ofbraininjuryandstrategiesforimprovedneurocognitiveoutcome.Pe- 25. Bodeau-LivinecF,CotM,KouraG,BoivinM.Assessingtheeffects /a
diatrRes2010;68:267–74. of maternalanemia onchilddevelopmentinBenin.In:BoivinMJ, rtic
6. AlSerouriAW,Grantham-McGregorSM,GreenwoodB,CostelloA. Giordani B, eds. Neuropsychology of Children in Africa: Pers- le-a
Impactofasymptomaticmalariaparasitaemiaoncognitivefunction pectives on Risk and Resilience. New York, NY: Springer, 2013: b
s
andschoolachievementofschoolchildrenintheYemenRepublic.Par- 203–14. tra
asitology2000;121:337–45. 26. vanHensbroekMB,PalmerA,JaffarS,SchneiderG,KwiatkowskiD. ct/5
7. FernandoD,deSilvaD,WickremasingheR.Short-termimpactofan Residualneurologicsequelaeafterchildhoodcerebralmalaria.JPediatr 9
acuteattackofmalariaonthecognitiveperformanceofschoolchildren 1997;131(1Pt1):125–9. /3/3
livinginamalaria-endemicareaofSriLanka.TransRSocTropMed 27. BirbeckGL,MolyneuxME,KaplanPW,etal.BlantyreMalariaProject 36
Hyg2003;97:633–9. EpilepsyStudy(BMPES)ofneurologicaloutcomesinretinopathy- /2
8
8. HoldingPA,StevensonJ,PeshuN,MarshK.Cognitivesequelaeofse- positivepaediatriccerebralmalariasurvivors:aprospectivecohort 95
veremalariawithimpairedconsciousness.TransRSocTropMedHyg study.LancetNeurol2010;9:1173–81. 40
1999;93:529–34. 28. BoivinMJ,GladstoneMJ,VokhiwaM,etal.Developmentaloutcomes 1 b
9. Nankabirwa J, Wandera B, Kiwanuka N, Staedke SG, Kamya MR, inMalawianchildrenwithretinopathypositivecerebralmalaria.Trop y g
Brooker SJ. Asymptomatic Plasmodium infection and cognition MedIntHealth2011;16:263–71. ue
ainmUongagnpdraim.AamryJscThroooplcMhielddreHnygin2a01h3ig;h88m:1a1l0a2ri–a8t.ransmissionsetting 29. cJoyhtonkCinCe,lePvaenlsoasknadltcsoisg-nMitoivretairmipAa,irOmpeonktainROce,reetbraal.lCmearleabrriao.sApimnaJlTflruoipd st on
10. ThuilliezJ,SissokoMS,ToureOB,KamateP,BerthélemyJ-C,Doumbo MedHyg2008;78:198–205. 10
OK.MalariaandprimaryeducationinMali:alongitudinalstudyinthe 30. PostelsDG,TaylorTE,MolyneuxM,etal.Neurologicoutcomesinret- Ap
11. vOilllnaegyeDofKD,PonolélgitutéEb,oKugaoriug.eSroPcKS,ceitMale.dY2o0u1n0g;Z71a:n3z2i4b–a3ri4.childrenwith i7n9o:1p2a6t8h–y7-2n.egative cerebral malaria survivors. Neurology 2012; ril 20
1
iron deficiency, iron deficiency anemia, stunting, or malaria have 31. Cserti-GazdewichCM,DhabangiA,MusokeC,etal.Inter-relationships 9
lowermotoractivityscoresandspendlesstimeinlocomotion.JNutr ofcardinalfeaturesandoutcomesofsymptomaticpediatricPlasmodi-
2007;137:2756–62. umfalciparummalariain1,933childreninKampala,Uganda.AmJ
12. MurphySC,BremanJG.GapsinthechildhoodmalariaburdeninAf- TropMedHyg2013;88:747–56.
rica:cerebralmalaria,neurologicalsequelae,anemia,respiratorydis- 32. DeMastQ,VanDongen-LasesEC,SwinkelsDW,etal.Mildincreases
tress,hypoglycemia,andcomplicationsofpregnancy.AmJTropMed inserumhepcidinandinterleukin-6concentrationsimpairironincor-
Hyg2001;64(1-2Suppl):57–67. porationinhaemoglobinduringanexperimentalhumanmalariainfec-
13. KolbB,WhishawIQ.Fundamentalsofhumanneuropsychology.6thed. tion.BrJHaematol2009;145:657–64.
NewYork,NY:WorthPublishers,2009. 33. Grantham-McGregorS,AniC.Areviewofstudiesontheeffectofiron
14. KostovicI,JudasM,PetanjekZ.Structuraldevelopmentofthehuman deficiencyoncognitivedevelopmentinchildren.JNutr2001;131(2S-2):
prefrontal cortex. In: Nelson CA, Luciana M, eds, Handbook of 649S–66S;discussion66S-68S.
(cid:129) (cid:129)
SevereMalarialAnemiaandCognition CID 2014:59 (1August) 343
34. PerkinsDJ,WereT,DavenportGC,KempaiahP,HittnerJB,Ong’echa in newborn infants of diabetic mothers. Pediatr Res 2004; 55:
JM.Severemalarialanemia:innateimmunityandpathogenesis.IntJ 1034–41.
BiolSci2011;7:1427–42. 37. CrawleyJ,WaruiruC,MithwaniS,etal.Effectofphenobarbitalon
35. von Seidlein L, Olaosebikan R, Hendriksen ICE, et al. Predicting seizure frequency and mortality in childhood cerebral malaria:
theclinicaloutcomeofseverefalciparummalariainAfricanchildren: a randomised, controlled intervention study. Lancet 2000; 355:
findings from a large randomized trial. Clin Infect Dis 2012; 54: 701–6.
1080–90. 38. CormierJ,ChuCJ.Safetyandefficacyoflevetiracetamforthetreatment
36. Siddappa AM, Georgieff MK, Wewerka S, Worwa C, Nelson CA, ofpartialonsetseizuresinchildrenfromonemonthofage.Neuropsy-
DeregnierRA.Irondeficiencyaltersauditoryrecognitionmemory chiatrDisTreat2013;9:295.
D
o
w
n
lo
a
d
e
d
fro
m
h
ttp
s
://a
c
a
d
e
m
ic
.o
u
p
.c
o
m
/c
id
/a
rtic
le
-a
b
s
tra
c
t/5
9
/3
/3
3
6
/2
8
9
5
4
0
1
b
y
g
u
e
s
t o
n
1
0
A
p
ril 2
0
1
9
(cid:129) (cid:129)
344 CID 2014:59 (1August) Bangiranaetal