Neurodevelopmental Impairments1 Year After Cerebral MalariaJohn T. Langfitt, PhD,a Michael P. McDermott, PhD,b Rachel Brim, MD,c,d Sebastian Mboma, MPH,d Michael J. Potchen, MD,d,e

Sam D. Kampondeni, MD,d,e Karl B. Seydel, MD,c,d Margaret Semrud-Clikeman, PhD,f Terrie E. Taylor, DOc,d

abstractBACKGROUND AND OBJECTIVES: Cerebral malaria (CM) causes significant mortality and morbidity insub-Saharan African children. Reliable morbidity estimates are scarce because ofmethodological variability across studies. We describe the incidence, course, and severity ofneurodevelopmental impairments in survivors of CM and the associated patientcharacteristics to inform epidemiologic estimates of malaria morbidity rates and preventionand treatment efforts.

METHODS:We conducted an exposure-control study of 85 survivors of CM and 100 age-matchedpatients in a control group who were enrolled at hospital discharge and assessed after 1, 6,and 12 months using caregiver interviews and standardized developmental, cognitive, andbehavioral measures.

RESULTS: Developmental or cognitive impairment (,10th percentile of the control distribution)and/or new onset of caregiver-reported behavior problems occurred in 53% of case patientscompared with 20% of the patients in the control group (odds ratio 4.5; 95% CI: 2.4 to 8.6;P , .001). In case patients, developmental or cognitive impairment at the 12-monthassessment was associated with HIV-positive status and short stature at presentation, moreprolonged fever and coma during admission, and severe atrophy or multifocal abnormalitiesbeing found on MRI at the 1-month assessment.

CONCLUSIONS: One-half of survivors of CM were neurodevelopmentally impaired at the 1-yearassessment. With these results, we support prevention trials of acute, neuroprotectiveinterventions and the allocation of resources to evaluation, education, and rehabilitationefforts to reduce the significant long-term burden of CM-associated impairments on familiesand their communities.

WHAT’S KNOWN ON THIS SUBJECT: In some studies,survivors of cerebral malaria have performed poorlyon measures of development, cognition, and behavior.It is unclear how many of these children are franklyimpaired, who are most likely to be impaired, andwhether impairment persists.

WHAT THIS STUDY ADDS: One-half of survivors ofcerebral malaria have some kind of impairment 1 yearafter the illness. Developmental or cognitiveimpairment is associated with having HIV, shorterstature, prolonged fever and coma during the initialillness, and abnormalities found on MRI.

To cite: Langfitt JT, McDermott MP, Brim R, et al.Neurodevelopmental Impairments 1 Year After CerebralMalaria. Pediatrics. 2019;143(2):e20181026

Departments of aNeurology, Psychiatry, bBiostatistics and Computational Biology, and eNeuroradiology, School ofMedicine and Dentistry, University of Rochester, Rochester, New York; cDepartment of Osteopathic MedicalSpecialties, College of Osteopathic Medicine, Michigan State University, East Lansing, Michigan; dBlantyre MalariaProject, College of Medicine, University of Malawi, Blantyre, Malawi; and fDepartment of Pediatrics, MedicalSchool, University of Minnesota, Minneapolis, Minnesota

Dr Langfitt assembled the data, designed and conducted the data analyses, drafted the initial

manuscript, and critically reviewed the manuscript for important intellectual content; Dr McDermott

designed and conducted the data analyses, drafted the initial manuscript, and critically reviewed the

manuscript for important intellectual content; Dr Brim contributed to the study design and

development of hypotheses for the current analyses, designed and managed data collection, and

critically reviewed the manuscript for important intellectual content; Mr Mboma supervised the

Cognitive Outcomes and Psychiatric Symptoms of Retinopathy-Positive Cerebral Malaria study team,

managed all aspects of data collection, and critically reviewed the manuscript for important

intellectual content; (Continued)

PEDIATRICS Volume 143, number 2, February 2019:e20181026 ARTICLE

Dow

nloaded from http://publications.aap.org/pediatrics/article-pdf/143/2/e20181026/1076071/peds_20181026.pdf by guest on 09 D

ecember 2021

Malaria is one of the most commoncauses of illness, death, andneurodisability worldwide.1,2

Parasitic infestation of the brain(cerebral malaria [CM]) affects∼575000 children in sub-SaharanAfrica annually, with an estimatedcase fatality rate of 15%.3,4

Approximately one-half of childrenwho are hospitalized with malariahave neurologic involvement.5

Potential pathogenic mechanismsinclude hypoxia, hypoglycemia,inflammation, increased intracranialpressure, and excitotoxicity.2 AcuteMRI reveals a range of brainpathology,6 but the relationshipsbetween MRI lesion locations,burdens, and outcomes in survivorsof CM have not been described.

It is difficult to obtain reliableestimates of the incidence, course,and severity of developmental,cognitive, and psychiatric morbiditiesrelated to CM because ofmethodologic variability amongprevious studies.7–12 Differentassessments are used at differentages. Few studies report rates ofimpairment on primary or summarymeasures of outcomes relative tolocal normative data. Characteristicsassociated with poor outcomes havebeen identified, but no studies haveexamined the unique association ofeach characteristic withoutcomes.4,9,13–17 Disruptivebehaviors (eg, aggression anddisobedience) have been noted in upto one-third of cases,4,18 but this hasnot been confirmed in prospectivestudies in which standardizedassessments are used.8,17

The Cognitive Outcomes andPsychiatric Symptoms ofRetinopathy-Positive Cerebral Malariastudy is a prospective, exposure-control study of CM outcomes inwhich standardized developmental,cognitive, and behavioral assessmentsover 5 years of follow-up are used.19

Here, we report the incidence, course,and severity of developmental,cognitive, and behavioral

impairments associated with CMacross the first 12 months of follow-up to inform epidemiologic estimatesof the burden of malaria. We alsoreport the associations of theseimpairments with demographic,clinical, and neuroimagingcharacteristics so that children at riskfor poor outcomes can be targeted forearly intervention.

METHODS

Ethics Approval

The study was approved by theresearch ethics bodies of theUniversity of Malawi College ofMedicine, Michigan State University,the University of Minnesota, and theUniversity of Rochester.

Participants

Recruitment and assessmentmethods, baseline characteristics, and1-month outcomes in the CognitiveOutcomes and Psychiatric Symptomsof Retinopathy-Positive CerebralMalaria study are described by Brimet al.19 Briefly, children who wereeligible were $6 months of age, hadno previous episodes of CM, and wereenrolled at the time of discharge fromthe major public hospital in southernMalawi during the peak malariaseasons of 2012–2014. Case patientshad been admitted for retinopathy-positive CM, which is defined by (1)the presence of Plasmodiumfalciparum parasitemia as determinedby using light microscopy ofperipheral blood film; (2)unarousable coma (Blantyre comascore [BCS] #2) with no otheridentifiable causes of coma afterexcluding hypoglycemia, meningitis,and postictal state; and (3) malariaretinopathy determined in anophthalmic examination by anexperienced ophthalmologist.20,21

Age-matched patients in the controlgroup had been admitted during thesame period for nonneurologicconditions (eg, dehydration,pneumonia, and other acute

infections). Case patients were morelikely to come from families withfewer resources (ie, rural, farmingfamilies with less parental educationand more rudimentary housing) andto have had a previous episode ofmalaria, but otherwise, they did notdiffer from patients in the controlgroup. Caregivers consented to returnsemiannually for interviews andassessments appropriate to thechild’s age at each visit. Children wereassessed semiannually up to age7 years and annually thereafter.

Assessments

At discharge from the hospital, allchildren with CM underwent anabbreviated neurologic examinationby a nonneurologist clinician toscreen for gross neurologicabnormalities. In children,5 years ofage, we assessed development atfollow-up visits using the 4 subscales(gross motor, fine motor, languagedevelopment, and socialdevelopment) of the MalawiDevelopmental Assessment Tool(MDAT),22 which is a developmentalscreening instrument culturallyadapted for Malawi. In children$5 years of age, we assessedcognitive function using thesequencing, learning, and delayedrecall subscales of the preschool andschool-age versions of the KaufmanAssessment Battery for Children,Second Edition (KABC-II).23

Caregivers of children $2 years ofage annually completed the preschooland school-aged versions of theBehavior Rating Inventory ofExecutive Function (BRIEF).24,25 Highscores on the BRIEF’s Inhibitory Self-Control Index (ISCI) reflect disruptivebehavior. The KABC-II and BRIEFwere translated but not culturallyadapted. During their interviews,caregivers were also asked todescribe in an open-ended mannerany new concerns about theirchildren’s behavior since the last visit.Using an approach similar to that ofBergemann et al,26 we createda single, age-normed measure of

2 LANGFITT et al

Dow

nloaded from http://publications.aap.org/pediatrics/article-pdf/143/2/e20181026/1076071/peds_20181026.pdf by guest on 09 D

ecember 2021

global development and cognition(DC), in which data from the age-appropriate assessment at a givenvisit were used (MDAT if the childwas ,5 years of age and KABC-II ifthe child was $5 years of age; seeSupplemental Information fordetails). The same approach was usedto create an ISCI age-normed z score.

Noncontrasted MRI was obtainedafter 1 month on a 0.35 T SignaOvation Excite MRI scanner (GeneralElectric Company, Boston, MA). Forcase patients only, 2neuroradiologists who were blind tooutcome data systematically rated thetype and severity of brain injury inspecific brain regions.27 Alldiscrepancies were resolved bydiscussion. Atrophy was rated on a 3-point scale (none, mild, or severe)that was used in a previous study.28

Focal abnormalities were classified asnone, unifocal, or multifocal on thebasis of focal injuries in 1 vs .1 of 6general brain regions (cerebralcortex, cerebral white matter,thalamus, basal ganglia, brainstemand pons, and cerebellum).

Statistical Analysis

Exposure-Control Comparisons

Bias in loss to follow-up wasevaluated by comparingdemographic, anthropometric, andhealth characteristics, and 1-monthoutcomes between children who wereexcluded from the 12-month analysesversus those who were included, byusing t tests and x2 tests. We testedthe primary hypothesis of groupdifferences in global DC across the 1-,6-, and 12-month visits using mixed-model repeated-measures analysis,with case-control status and timebeing the categorical variables. Anunstructured covariance matrix wasused to model the dependence of thewithin-subject observations. Changein global DC was assessed by usinga linear mixed-effects model thatincluded group membership and time(which was treated as a continuousvariable) and their interaction.

Similar analyses were conducted forthe ISCI scores and modified toaccount for observations only at the1- and 12-month assessments. Forthe global DC measure and ISCI, weused the mixed-model repeated-measures model to test (separately)whether a number of potentialconfounders (age, sex, height for age,HIV status, family resources, previousepisode of malaria, and MDAT versusKABC-II assessment) could be used toaccount for any observed groupdifferences in global DC. Fisher’sexact tests were used to comparegroup differences in the incidence ofpsychometrically defined DC orbehavior impairment at the 12-monthassessment (global DC or ISCI scoresin the decile of the patients in thecontrol group who were mostimpaired) and in the incidence of newbehavior concerns reported directlyby caregivers.

Characteristics Associated WithOutcomes in Case Patients

In case patients, we used separatelogistic regressions to identifydemographic and clinicalcharacteristics that were associatedwith an impaired-level DC score,impaired-level ISCI score, or newcaregiver-reported behavioralconcerns at the 12-monthassessment. Characteristics werechosen on the basis of previousstudies and included age, sex, heightand weight for age, HIV status, familyresources, maximum in-hospitaltemperature, fever clearance time,parasite load, parasite clearance time,white blood cell count on admission,lactate and oxygen saturation, hoursto a BCS of 3, and hours to full comaresolution (BCS = 5). We used the all-possible-regressions approach toempirically select the best subsets ofthese characteristics.29 We report theresults from the model that yieldedthe highest value of the x2 scorestatistic for the largest number ofcharacteristics such that adding $1characteristic did not significantlyenhance the model. We used analysis

of variance (supplemented withTukey’s test for pairwise groupcomparisons), Mantel-Haenszel x2

tests, and Fisher’s exact tests to testassociations between MRI findingsand outcomes at 12 months.

RESULTS

Participants

Of the 221 children who wereenrolled and returned for the 1-month assessment, 29 (13%) did notreturn for the 12-month assessment(Fig 1). Seven of the 192 children whoreturned lacked global DC and ISCInorm-based scores because of theirage (see Supplemental Information)and were excluded from furtheranalysis. The 36 children who wereexcluded from the 12-month analysisdid not differ from the 185 who wereincluded (85 case patients and 100patients in the control group) on mostbaseline characteristics nor on globalDC and ISCI scores, which wereobtained at the 1-month assessment(see Supplemental Table 2). Thosewho were excluded weighed more,and their caregivers were more likelyto have expressed new concernsabout behavior at the 1-monthassessment (19% excluded versus

FIGURE 1Patient flow through the study. Primary analy-ses were conducted for the 185 participantswho completed the 12-month follow-up visit.

PEDIATRICS Volume 143, number 2, February 2019 3

Dow

nloaded from http://publications.aap.org/pediatrics/article-pdf/143/2/e20181026/1076071/peds_20181026.pdf by guest on 09 D

ecember 2021

5% included; P = .006). Case patientswho were included were more likelythan patients in the control group tobe HIV-positive and have fewer familyresources, a previous episode ofmalaria, clinically abnormal MRIresults, and new behavioral concernsat the 1-month assessment. The vastmajority (95%) of case patients werejudged to be grossly neurologicallynormal at the time of discharge.

Associations Between CM and DCOutcomes

Global DC scores were lower for casepatients than patients in the controlgroup across all assessment periods(0.73 SD units; 95% confidenceinterval [CI]: 0.47 to 0.98; P , .001;Fig 2). There was no change in globalDC across assessments of casepatients (mean slope = 0.00; 95% CI:20.02 to 0.02) nor in assessments ofpatients in the control group (mean

slope = 0.00; 95% CI: 20.02 to 0.02).Mean slopes did not differ betweenthe groups (P = .88). The differencebetween case patients and patients inthe control group at the 12-monthassessment was 0.75 SD units (95%CI: 0.45 to 1.04; P , .001).Developmental or cognitiveimpairment at the 12-monthassessment (global DC scores ,10thpercentile of the control sample) waspresent in 29% of case patients and7% of patients in the control group(odds ratio [OR] 5.5; 95% CI: 2.3 to13.6; P , .001).

The difference in global DC betweencase patients and patients in thecontrol group persisted afteradjusting for HIV status (P , .001;Fig 2). Global DC differed betweenchildren who were HIV-positive andthose who were HIV-negative acrossall assessment periods (20.51 SD

units; 95% CI:20.07 to 0.97; P = .02).Developmental or cognitiveimpairment at the 12-monthassessment was observed in 8 of 21children (38%) who were HIV-positive versus 24 of 160 children(15%) who were HIV-negative (OR3.5; 95% CI: 1.3 to 9.3; P = .01). Case-control differences in global DCpersisted when accounting for age,sex, height for age, a previous episodeof malaria, family resources, and theassessment on which global DC wasbased (KABC-II versus MDAT) forthat observation (all P , .001).

Associations Between CM andBehavioral Problems

ISCI scores were available for 125children age $2 years (58 casepatients and 67 patients in thecontrol group). ISCI scores werehigher in case patients (greaterproblems with inhibition and self-control) than in patients in thecontrol group across follow-up (0.48SD units; 95% CI: 0.12 to 0.84;P = .02). ISCI scores did not changebetween assessments across groups(mean change = 20.10; 95% CI:20.35 to 0.15; P . .10), and this wasconsistent across groups. At the 12-month assessment, ISCI scores werehigher in case patients than inpatients in the control group (0.49 SDunits; 95% CI: 0.09 to 0.90; P , .01).Case-control differences on the ISCIpersisted when accounting for age,sex, height for age, a previous episodeof malaria, and family resources (allP , .05). Behavioral impairment atthe 12-month assessment (ISCI scores.90th percentile of the controlsample) was present in 32% of casepatients and 11% of patients in thecontrol group (OR 3.8; 95% CI: 1.4 to9.8; P = .005).

In response to the open-endedinterview question, caregiversreported new concerns aboutbehavior at some point during follow-up for 35% of case patients and 16%of patients in the control group (OR2.9; 95% CI: 1.4 to 5.7; P = .001).

FIGURE 2Mean DC z scores in case patients and patients in the control group who were either HIV-positive orHIV-negative. Case patients differed from patients in the control group (P , .001), and patients whowere HIV-positive differed from those who were HIV-negative (P = .02), with no change over time orinteractions (all P$ .25). The filled symbols indicate HIV-positive status; the unfilled symbols indicateHIV-negative status. CM+, case patients; CM2, patients in the control group; HIV-, HIV-negative; HIV+,HIV-positive.

4 LANGFITT et al

Dow

nloaded from http://publications.aap.org/pediatrics/article-pdf/143/2/e20181026/1076071/peds_20181026.pdf by guest on 09 D

ecember 2021

Among case patients, new concernswere reported for 9 patients at the1-month assessment, 11 patients atthe 6-month assessment, and 10patients at the 12-monthassessment. Among patients in thecontrol group, new concerns werereported for no patients at the 1-month assessment, 7 patients at the6-month assessment, and 9 patientsat the 12-month assessment. Newbehaviors were disruptive(aggression, arguing, hyperactivity,and disobedience) in 28% of casepatients and 8% of patients in thecontrol group (OR 4.2; 95% CI: 1.9to 10.7; P , .001). One-half of thecase patients (53%) haddevelopmental or cognitiveimpairment, or new concerns aboutbehavior, or both compared with20% of patients in the control group(OR 4.5; 95% CI: 2.4 to 8.6;P , .001).

Characteristics of Case PatientsAssociated With 12-Month Outcomes

A 3-characteristic model revealedthe strongest joint association withdevelopmental or cognitiveimpairment, which was comparedwith the best model with 2characteristics. Models with 4 to 6characteristics marginally improvedthe score statistic, but none of theadded characteristics weresignificantly associated withoutcomes. Included in the finalmodel were height for age incentimeters (OR 0.56; 95% CI: 0.35to 0.85; P = .009), fever clearancetime in hours (OR 0.97; 95% CI: 0.95to 0.99; P = .02), and coma durationin hours (OR 1.03; 95% CI: 1.01 to1.05; P = .0008). Developmental orcognitive impairment was morecommon among case patients whowere short for their age at admissionand took longer to clear their feverand emerge from a coma. Nocharacteristics were associated withhaving an ISCI score .90thpercentile or new onset of behaviorconcerns by the 12th month.

MRI Findings and Their AssociationWith Case Patient Characteristics

One-month MRIs were obtained for71 of 85 case patients. Abnormalitieswere present in 53 of 71 patients(75%). Atrophy was absent in 39% ofpatients, mild in 37% of patients, andsevere in 24% of patients. Focalabnormalities were absent in 45% ofpatients, unifocal in 37% of patients,and multifocal in 18% of patients.Some atrophy and $1 focalabnormality were present in 41% ofpatients, atrophy only was present in20% of patients, focal abnormalitiesonly were present in 14% of patients,and no abnormalities were present in25% of patients. Focal abnormalitieswere observed in the periventricularwhite matter (n = 29), cerebral cortex(n = 10), subcortical white matter(n = 9), corpus callosum (n = 7),globus pallidus (4), caudate (n = 7),putamen (n = 6), pons (n = 3),brainstem (n = 5), and cerebellum(n = 6).

Severe atrophy and focalabnormalities were associated withyounger age (atrophy P = .02; focalabnormalities P , .001; see the toppanel of Table 1). Case patients withsevere atrophy were in a coma longerthan those with mild or no atrophy(P , .001). Case patients with severeatrophy had a higher white blood cellcount than those with mild atrophy(P = .046).

MRI Association With Case PatientOutcomes at 12 Months

Global DC scores were lower in casepatients with severe atrophy ormultifocal abnormalities than in casepatients with mild or no atrophy(P = .0009) or unifocal or no focalabnormalities (P = .02; see thebottom 2 rows of Table 1).Developmental or cognitiveimpairment was observed in 6 of 28patients (21%) with no atrophy, 8 of26 patients (31%) with mild atrophy,and 10 of 17 patients (59%) withsevere atrophy (P = .01).Developmental or cognitiveTA

BLE1Associations

BetweenMRI

Findings,B

aselineCharacteristics,andOutcom

esat

the12-Month

Assessment

GroupDifferences

Atrophy

FocalAbnorm

alities

Severe

Versus

Mild,M

ean

Difference(95%

CI)

Mild

Versus

None,M

ean

Difference(95%

CI)

Severe

Versus

None,M

ean

Difference(95%

CI)

Multifocal

Versus

Unifocal,

MeanDifference(95%

CI)

UnifocalVersus

None,M

ean

Difference(95%

CI)

Multifocal

Versus

None,M

ean

Difference(95%

CI)

Age,y

21.45

(0.10to

23.00)

20.32

(21.67

to0.02)

21.77

a(2

3.30

to20.25)

21.04

(22.63

to0.55)

21.44

a(2

2.70

to20.20)

22.48

a(2

4.02

to20.93)

Comaduration,h

45.8a(19.1to

72.6)

6.8(16.6to

30.2)

52.6a(26.2to

79.0)

19.0(2

14.2to

52.2)

6.2(2

19.6to

32.1)

25.2(2

6.9to

57.4)

White

bloodcellcount,

cells

permm3

4.80

a(0.10to

9.50)

20.97

(25.20

to3.30)

3.80

(20.77

to8.40)

1.36

(23.90

to6.60)

0.89

(23.40

to5.12)

2.25

(22.90

to7.40)

DC,z

scoreunits

21.13

a(2

2.00

to20.25)

20.26

(21.02

to0.51)

21.39

a(2

2.25

to20.52)

21.15

a(2

2.15

to20.15)

0.14

(20.64

to0.91)

21.02

a(2

2.00

to20.05)

BRIEFISCI,z

scoreunits

0.44

(20.60

to1.50)

20.10

(21.00

to0.90)

0.37

(20.75

to1.50)

0.53

(20.52

to1.59)

0.47

(20.44

to1.38)

1.00

(20.07

to2.10)

aPairwisecomparisons

weresignificant

atthe0.05

levelwhenusingTukey’s

test.

PEDIATRICS Volume 143, number 2, February 2019 5

Dow

nloaded from http://publications.aap.org/pediatrics/article-pdf/143/2/e20181026/1076071/peds_20181026.pdf by guest on 09 D

ecember 2021

impairment was not associated withfocal abnormalities (x2 = 3.7; P = .17).MRI findings were not associatedwith ISCI scores nor were theyassociated with new onset ofbehavioral problems (all P . .10).Heat maps (Figs 3 and 4) reveal MRI

findings and global DC and ISCIscores for each case patient. Casepatients with severe atrophy and/ormultifocal lesions are among the mostdevelopmentally and cognitivelyimpaired. ISCI scores were .90thpercentile of the control sample in 5

of the 6 case patients with basalganglia abnormalities.

DISCUSSION

We found developmental, cognitive,or behavioral impairments inapproximately one-half (53%) of casepatients with CM. Impairments wereassociated with MRI abnormalities,which occurred in 75% of casepatients. This rate is ∼3 times the rateobserved in a sample of olderchildren who were developmentallynormal from the same population, forwhom the same MRI acquisition andrating approach was used.30

Impairments persisted up to 1 yearafter the index episode. Epidemiologicestimates of the morbidity associatedwith malaria, such as those used inthe Global Burden of Diseasestudies,31 cannot be used to accountsufficiently for these impairments.Previous estimates of the neurologicmorbidity of malaria3 have beenbased solely on the estimated 2% to4% of children who have clinicallyevident neurologic sequelae after anepisode of illness (eg, paresis, ataxia,hearing deficits, visual field deficits,aphasia, and convulsions). Consistentwith this, 5% of our survivors of CMwho were retinopathy-positive werenot grossly neurologically normalwhen discharged from the hospital.The Global Burden of Disease methodhas no specific disability weight forcase patients with malaria andcognitive impairment who appeargrossly neurologically normal, yet themethod assigns disability weights of0.024 to 0.436 to patients withcognitive impairment associated withJapanese encephalitis, ascariasis,trichuriasis, and hookworminfections. Results of this study maybe useful in informing more accurateestimates of the neurologic burden ofmalaria.

Developmental or cognitiveimpairment was associated with HIVcoinfection and shorter stature atpresentation, more prolonged fever

FIGURE 3Heat map revealing the association of MRI findings in 71 case patients ranked by degree ofdevelopmental and cognitive impairment (global DC z score). BST, brain stem; CAU, caudate nucleus;CC, corpus callosum; GBP, globus pallidus; PVW, periventricular white matter; SCW, subcortical whitematter.

6 LANGFITT et al

Dow

nloaded from http://publications.aap.org/pediatrics/article-pdf/143/2/e20181026/1076071/peds_20181026.pdf by guest on 09 D

ecember 2021

and coma during hospitalization, andatrophy and/or focal abnormalitiesbeing found on postacute MRI.Anthropometric measures and comaduration are among thecharacteristics that have been mostconsistently associated with

developmental or cognitive outcomesin previous CM cohorts.4,9,15–17 Thisis consistent with the deleteriouseffects of HIV-positive status, poornutritional status, and prolonged,increased intracranial pressure onneuronal survival and subsequent

developmental and cognitiveoutcomes.32–34 Associations betweenfever characteristics and outcomeshave been observed less consistently.Survivors of CM who are HIV-positive,have poor nutritional status, and takelonger to emerge from a fever andcoma are therefore most appropriatefor clinical surveillance and earlyintervention services. Theaccumulated evidence is used tosupport clinical trials of acuteinterventions for resolving feversooner and enhancingneuroprotection during the acutephase of the illness.

Approximately one-third of caregiversof children with CM reportedproblems with inhibition and self-control as determined ina standardized assessment or witha single, open-ended question. Wemay have slightly underestimated theincidence because those withbehavior problems at the 1-monthassessment were less likely to returnfor the 12-month follow-upassessment. The etiology of theseproblems remains unclear becausethey were not significantly associatedwith any specific characteristics. It isintriguing that 5 of the 6 case patientswith basal ganglia abnormalities hadISCI scores .90th percentile of thecontrol sample because basal ganglialesions in children have beenassociated with behavioraldisinhibition.35,36 However,confirmation of this observationrequires a larger sample.

Limitations of this study include ourenrollment of participants froma single country, which may limitgeneralizability to other malaria-endemic areas. We enrolled hospital-based patients for the control group,who differed from case patients inprevious histories of malaria andtheir families’ educational andhousing resources. However, groupdifferences persisted when thesedifferences were accounted for in theanalysis. This is the only study inwhich the relationship between HIV

FIGURE 4Heat map revealing the association of MRI findings in 54 case patients ranked by degree ofimpairment in inhibition and self-control (BRIEF ISCI z scores). BST, brain stem; CAU, caudate nucleus;CC, corpus callosum; GBP, globus pallidus; PVW, periventricular white matter; SCW, subcortical whitematter.

PEDIATRICS Volume 143, number 2, February 2019 7

Dow

nloaded from http://publications.aap.org/pediatrics/article-pdf/143/2/e20181026/1076071/peds_20181026.pdf by guest on 09 D

ecember 2021

status, chronic brain MRI findings,and outcomes in survivors of CM havebeen examined, but the sample sizewas too small to identify associationsbetween use of antiretroviral therapyand outcomes. Also, the MRI findingswere too heterogeneous to identifyassociations between damage tospecific brain areas and specificoutcomes beyond an expected effectof severe atrophy and multifocallesions on development or cognition.

CONCLUSIONS

Developmental, cognitive, andbehavioral impairments after CMrepresent a large and persistentburden on sub-Saharan Africanfamilies and communities for whomtreatment, educational, andrehabilitation resources are scarce.

Analyses of 5-year outcomes (whenthe majority of patients have reachedschool age) and additionalqualitative studies of the cohort willbe used to help clarify the long-termfunctional impact of theseimpairments. These findings shouldbe used to encourage policy makersin malaria-endemic regions toprioritize directing resources totrials of acute interventions forpreventing morbidity and to parentaleducation, teacher training in specialeducation, and rehabilitationprograms for survivors of CM whoare affected.

ACKNOWLEDGMENTS

We thank Clara Antonio for the initialtraining on the MDAT, KABC-II, andBRIEF and Chimwemwe Kalengo and

Edith Kafoteka for performing theassessments. David Bearden andGretchen Birbeck provided helpfulcomments on initial drafts.

ABBREVIATIONS

BCS: Blantyre coma scoreBRIEF: Behavior Rating Inventory

of Executive FunctionCI: confidence intervalCM: cerebral malariaDC: development and cognitionISCI: Inhibitory Self-Control IndexKABC-II: Kaufman Assessment

Battery for Children,Second Edition

MDAT: Malawi DevelopmentalAssessment Tool

OR: odds ratio

Drs Potchen and Kampondeni designed and supervised the collection of MRI data, interpreted the MRI scans, conceptualized their analysis, and critically reviewed

the manuscript for important intellectual content; Dr Seydel conceptualized and designed the study, oversaw the collection and quality of clinical data during

hospital admission, and critically reviewed the manuscript for important intellectual content; Dr Semrud-Clikeman conceptualized and designed the study,

contributed to the selection of the assessment instruments and training of the clinical evaluators, and critically reviewed the manuscript for important intellectual

content; Dr Taylor conceptualized and designed the study, provided administrative and budgetary oversight, and critically reviewed the manuscript for important

intellectual content; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

DOI: https://doi.org/10.1542/peds.2018-1026

Accepted for publication Oct 31, 2018

Address correspondence to John T. Langfitt, PhD, Strong Memorial Hospital, Box 673, 601 Elmwood Ave, Rochester, NY 14642. E-mail: john_langfitt@urmc.

rochester.edu

PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275).

Copyright © 2019 by the American Academy of Pediatrics

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

FUNDING: Funded by the Department of Psychiatry at the Michigan State University College of Osteopathic Medicine in East Lansing, Michigan. Dr Langfitt’s initial

work on this project was supported through a sabbatical granted by the University of Rochester.

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

REFERENCES

1. World Health Organization.World Malaria Report. Geneva,Switzerland: World HealthOrganization; 2017. Available at:https://www.who.int/malaria/publications/world-malaria-report-2017/en/. Accessed August 6, 2016

2. Idro R, Marsh K, John CC, Newton CR.Cerebral malaria: mechanisms of

brain injury and strategies forimproved neurocognitive outcome.Pediatr Res. 2010;68(4):267–274

3. Murphy SC, Breman JG. Gaps in thechildhood malaria burden in Africa:cerebral malaria, neurologicalsequelae, anemia, respiratory distress,hypoglycemia, and complications of

pregnancy. Am J Trop Med Hyg. 2001;64(suppl 1–2):57–67

4. Birbeck GL, Molyneux ME, Kaplan PW,et al. Blantyre Malaria Project EpilepsyStudy (BMPES) of neurologicaloutcomes in retinopathy-positivepaediatric cerebral malaria survivors:a prospective cohort study. LancetNeurol. 2010;9(12):1173–1181

8 LANGFITT et al

Dow

nloaded from http://publications.aap.org/pediatrics/article-pdf/143/2/e20181026/1076071/peds_20181026.pdf by guest on 09 D

ecember 2021

5. Idro R, Ndiritu M, Ogutu B, et al. Burden,features, and outcome of neurologicalinvolvement in acute falciparummalaria in Kenyan children. JAMA. 2007;297(20):2232–2240

6. Potchen MJ, Kampondeni SD, Seydel KB,et al. Acute brain MRI findings in 120Malawian children with cerebral malaria:new insights into an ancient disease. AJNRAm J Neuroradiol. 2012;33(9):1740–1746

7. Holding PA, Taylor HG, Kazungu SD, et al.Assessing cognitive outcomes in a ruralAfrican population: development ofa neuropsychological battery in KilifiDistrict, Kenya. J Int Neuropsychol Soc.2004;10(2):246–260

8. Carter JA, Ross AJ, Neville BG, et al.Developmental impairments followingsevere falciparum malaria in children.Trop Med Int Health. 2005;10(1):3–10

9. Boivin MJ, Bangirana P, Byarugaba J,et al. Cognitive impairment aftercerebral malaria in children:a prospective study. Pediatrics. 2007;119(2). Available at: www.pediatrics.org/cgi/content/full/119/2/e360

10. John CC, Bangirana P, Byarugaba J, et al.Cerebral malaria in children is associatedwith long-term cognitive impairment.Pediatrics. 2008;122(1). Available at: www.pediatrics.org/cgi/content/full/122/1/e92

11. Bangirana P, Musisi S, Boivin MJ, et al.Malaria with neurological involvementin Ugandan children: effect on cognitiveability, academic achievement andbehaviour. Malar J. 2011;10:334

12. Bangirana P, Opoka RO, Boivin MJ, et al.Severe malarial anemia is associatedwith long-term neurocognitive impairment.Clin Infect Dis. 2014;59(3):336–344

13. Holding PA, Stevenson J, Peshu N, MarshK. Cognitive sequelae of severe malariawith impaired consciousness. Trans RSoc Trop Med Hyg. 1999;93(5):529–534

14. Boivin MJ. Effects of early cerebralmalaria on cognitive ability inSenegalese children. J Dev BehavPediatr. 2002;23(5):353–364

15. Idro R, Carter JA, Fegan G, Neville BG,Newton CR. Risk factors for persistingneurological and cognitive impairmentsfollowing cerebral malaria. Arch DisChild. 2006;91(2):142–148

16. Kihara M, Carter JA, Holding PA,et al. Impaired everyday memory

associated with encephalopathy ofsevere malaria: the role of seizuresand hippocampal damage. Malar J.2009;8:273

17. Boivin MJ, Gladstone MJ, Vokhiwa M,et al. Developmental outcomes inMalawian children with retinopathy-positive cerebral malaria. Trop Med IntHealth. 2011;16(3):263–271

18. Idro R, Kakooza-Mwesige A, BalyejjussaS, et al. Severe neurological sequelaeand behaviour problems after cerebralmalaria in Ugandan children. BMC ResNotes. 2010;3:104

19. Brim R, Mboma S, Semrud-Clikeman M,et al. Cognitive outcomes andpsychiatric symptoms of retinopathy-positive cerebral malaria: cohortdescription and baseline results. AmJ Trop Med Hyg. 2017;97(1):225–231

20. World Health Organization. Severemalaria. Trop Med Int Health. 2014;19(suppl 1):7–131

21. Taylor TE, Fu WJ, Carr RA, et al.Differentiating the pathologies of cerebralmalaria by postmortem parasite counts[published correction appears in NatMed. 2004;10(4):435]. Nat Med. 2004;10(2):143–145

22. Gladstone M, Lancaster GA, Umar E,et al. The Malawi DevelopmentalAssessment Tool (MDAT): the creation,validation, and reliability of a tool toassess child development in ruralAfrican settings. PLoS Med. 2010;7(5):e1000273

23. Kaufman AS. KABC-II: KaufmanAssessment Battery for Children.London, England: Pearson Clinical; 2004

24. Gioia GA, Isquith PK, Guy SC, Kenworthy L.Behavior Rating Inventory of ExecutiveFunction: BRIEF. Odessa, FL: PsychologicalAssessment Resources; 2000

25. Familiar I, Ruisenor-Escudero H,Giordani B, et al. Use of the BehaviorRating Inventory of Executive Functionand Child Behavior Checklist inUgandan children with HIV or a historyof severe malaria. J Dev Behav Pediatr.2015;36(4):277–284

26. Bergemann TL, Bangirana P, Boivin MJ,Connett JE, Giordani BJ, John CC.Statistical approaches to assess theeffects of disease on neurocognitivefunction over time. J Biom Biostat.2012;(suppl 7):7310

27. Potchen MJ, Kampondeni SD, Ibrahim K,et al. NeuroInterp: a method forfacilitating neuroimaging research oncerebral malaria. Neurology. 2013;81(6):585–588

28. Seydel KB, Kampondeni SD, Valim C,et al. Brain swelling and death inchildren with cerebral malaria.N Engl J Med. 2015;372(12):1126–1137

29. Miller AJ. Subset Selection inRegression. 2nd ed. Boca Raton, FL:Chapman and Hall/CRC; 2002

30. Potchen MJ, Kampondeni SD, MallewaM, Taylor TE, Birbeck GL. Brainimaging in normal kids: a community-based MRI study in Malawianchildren. Trop Med Int Health. 2013;18(4):398–402

31. Mathers CD, Ezzati M, Lopez AD.Measuring the burden of neglectedtropical diseases: the global burden ofdisease framework. PLoS Negl Trop Dis.2007;1(2):e114

32. Abubakar A, Van Baar A, Van de VijverFJ, Holding P, Newton CR. PaediatricHIV and neurodevelopment in sub-Saharan Africa: a systematic review.Trop Med Int Health. 2008;13(7):880–887

33. Griffiths JK, Kikafunda JK. Childhoodthreats to adult cognition in sub-Saharan Africa: malaria, anemia,stunting, enteric enteropathy, and themicrobiome of malnutrition. In: MusisiSJ, Jacobson S, eds. BrainDegeneration and Dementia inSub-Saharan Africa. New York, NY:Springer; 2015:75–87

34. Uzzell BP, Obrist WD, Dolinskas CA,Langfitt TW. Relationship of acute CBFand ICP findings to neuropsychologicaloutcome in severe head injury.J Neurosurg. 1986;65(5):630–635

35. Krause M, Mahant N, Kotschet K,et al. Dysexecutive behaviourfollowing deep brain lesions–adifferent type of disconnectionsyndrome? Cortex. 2012;48(1):97–119

36. Westmacott R, McDonald KP, deVeber G,et al. Neurocognitive outcomes inchildren with unilateral basal gangliaarterial ischemic stroke and secondaryhemidystonia. Child Neuropsychol. 2018;24(7):923–937

PEDIATRICS Volume 143, number 2, February 2019 9

Dow

nloaded from http://publications.aap.org/pediatrics/article-pdf/143/2/e20181026/1076071/peds_20181026.pdf by guest on 09 D

ecember 2021