Menu
Menu
Menu
Menu
Background: The frequency of recovery from undernutrition after an episode of severe malaria, and the relationship between undernutrition during severe malaria and clinical and cognitive outcomes are not well characterized. Methods: We evaluated undernutrition and cognition in children in Kampala, Uganda 18 months to 5 years of age with cerebral malaria (CM), severe malarial anemia (SMA) or community children (CC). The Mullen Scales of Early Learning was used to measure cognition. Undernutrition, defined as 2 SDs below median for weight-for-age (underweight), height-for-age (stunting) or weight-for-height (wasting), was compared with mortality, hospital readmission and cognition over 24-month follow-up. Results: At enrollment, wasting was more common in CM (16.7%) or SMA (15.9%) than CC (4.7%) (both p < 0.0001), and being underweight was more common in SMA (27.0%) than CC (12.8%; p = 0.001), while prevalence of stunting was similar in all three groups. By 6-month follow-up, prevalence of wasting or being underweight did not differ significantly between children with severe malaria and CC. Undernutrition at enrollment was not associated with mortality or hospital readmission, but children who were underweight or stunted at baseline had lower cognitive z-scores than those who were not {underweight, mean difference [95% confidence interval (CI)]-0.98 (-1.66,-0.31),-0.72 (-1.16,-0.27) and-0.61 (-1.08,-0.13); and stunted,-0.70 (-1.25,-0.15),-0.73 (-1.16,-0.31) and-0.61 (-0.96,-0.27), for CM, SMA and CC, respectively}. Conclusion: In children with severe malaria, wasting and being underweight return to population levels after treatment. However, being stunted or underweight at enrollment was associated with worse long-term cognition in both CC and children with severe malaria.
This study was conducted in Kampala, Uganda. Kampala is the capital and largest city in Uganda with an estimated population of 1 507 080 according to the 2014 census [33]. Children with CM and SMA were enrolled from Mulago Hospital which is the national referral hospital in Uganda. In 2019, Uganda accounted for about 5% of all malaria cases globally [34]. Uganda also has a high prevalence of undernutrition among children under 5 years (29% stunting, 11% underweight and 4% wasting in 2016) [35]. Between 2008 and 2013, Ugandan children aged 18 months to 5 years with CM, SMA and asymptomatic CC with no acute illness were enrolled in Kampala, Uganda. Children with CM and SMA meeting the study criteria were recruited and enrolled from the Acute Care Unit, which is the emergency pediatric unit of Mulago Hospital. On admission, national treatment guidelines were used to provide emergency care to children with CM or SMA. After stabilization, the caregivers of children meeting the eligibility criteria were approached for participation in the study. Children were enrolled after written informed consent was provided. The CC were recruited from the nuclear family, extended family or household compound area of children with CM or SMA. Parents of children with CM or SMA were given information about the study, asked whether any eligible children were present in their extended family, and requested to bring the eligible children to the center for evaluation. Parents of children in the household compound of a child with CM or SMA were notified about the study during a home visit. CM was defined as: (i) Plasmodium falciparum on blood smear; (ii) coma [Blantyre Coma Scale (BCS) score ≤2, Glasgow Coma Scale (GCS) ≤ 8] after ruling out other causes of coma (hypoglycemia, postictal state and meningitis). SMA was defined as the presence of P.falciparum on blood smear and a hemoglobin level ≤5 g/dl. Ugandan Ministry of Health national guidelines define SMA as a hemoglobin ≤5 g/dl with any P.falciparum parasitemia. No parasitemia threshold is required to meet this definition. We made our study definition of SMA to be consistent with the Ugandan Ministry of Health guidelines. Children with CM and severe anemia were classified as CM. Exclusion criteria for all children included: (i) previous history of head trauma, coma or prior hospitalization for undernutrition; (ii) cerebral palsy. Exclusion criteria for children with SMA included: (i) impaired consciousness to any degree on physical exam (e.g. GCS < 15 or BCS < 5); (ii) seizure activity prior to admission; (iii) any other clinical evidence of central nervous system (CNS) disease. Among children with severe malaria, those with concomitant infections were included but we excluded those with a known chronic illness. Exclusion criteria for CC included (i) any active illness or illness within the past 4 weeks requiring medical care; (ii) chronic illness requiring medical care; (iii) major medical abnormalities on screening history or physical exam; (iv) known developmental delay; and (v) prior history of coma. Children who had known HIV infection at the time of recruitment were not included in the study, but those who were diagnosed on enrollment (all children were tested) and were not symptomatic were not excluded from the study. Symptomatic HIV is known to affect cognition in children, so we excluded children with a known diagnosis of HIV. The number of children with asymptomatic, newly diagnosed HIV infection was small, and the effect of this infection on cognition was not known, so we did not exclude them from the study. Socioeconomic status (SES) was assessed using a scoring instrument that included material possessions, house structure, living density, food resources and access to electricity and clean water [36]. Emotional stimulation in the home was measured by age-appropriate versions of the Home Observation for the Measurement of the Environment (HOME). The Infant–Toddler version of the HOME Inventory was used for the children in the younger age group (<3 years), and the Early Childhood version used for the older age group (3–6 years) [37]. These age-appropriate versions of the HOME assessment were administered by trained staff during a home visit. Data on SES, home environment, dietary habits, parent and child education were collected at the home visit after discharge therefore it was missing for children who died or were lost to follow-up. Children with CM or SMA were managed according to the Ugandan Ministry of Health treatment guidelines. These included intravenous quinine or artesunate until a patient is alert and then oral quinine and artemether–lumefantrine therapy for outpatient [5]. In addition to the antimalarials, all children with SMA were transfused with blood at a dose of 20 mg/kg of whole blood or 10 mg/kg of packed cells. After correction of the severe anemia, the children were put on oral hematinic—ferrous sulfate at a dose of 2 mg/kg for 3 months [38]. A child’s weight was measured with a digital weight scale. Children who could stand were weighed on the scale. Those unable to stand were weighed in their caregiver’s arms, and the caregiver’s weight was subtracted from the child’s weight. Recumbent length was taken for all children with CM or SMA at enrollment. For CC at enrollment and follow-up and for children with CM or SMA at follow-up, recumbent length of children <2 years was determined using a stadiometer, and a wall mounted-tape measure was used to assess the standing height of children ≥2. Heights and weights were converted into z-scores (height-for-age, weight-for-age and weight-for-height) based on 2006 World Health Organization (WHO) growth standards for infants and children <5 years and 2007 WHO growth references for children ≥5 years by follow-up visits [39, 40]. We defined nutritional status using standard anthropometric z-score cutoffs: more than 2 SD units below the reference median for weight-for-age (underweight), height-for-age (stunted) or weight-for-height (wasted). Children underwent cognitive evaluation either a week after discharge (CM or SMA) or at enrollment (CC) and then at 6, 12 and 24 month follow-up. Neuropsychology testers were blinded to study group. The Mullen Scales of Early Learning [41], which have been used and validated in Ugandan children in previous studies [4, 5], were used to measure cognitive ability. The early learning composite score was calculated from the sum of fine motor, visual reception, receptive language and expressive language scores [5]. To account for differences in child age, we converted each raw score into a z-score using scores of the CC children. The z-scores were computed as (actual score—mean score for a child’s age)/SD, where the mean score for a child’s age and SD were computed by fitting a quadratic mixed effects model, including a random intercept for the child and where correlations within a child were based on time between visits, to data for all visits for all CC children. Children who crossed the 5-year age limit for the Mullen Scales of Early Learning were tested by the Kauffman Assessment Battery for Children-II (KABC-II) when ≥ 5 years. Outcomes from the Mullen Scales and the KABC-II are not directly comparable, as they measure different components of cognition, so only results from the Mullen Scales are reported. At later follow-up time points, more children crossed the 5 year age threshold and therefore did not have Mullen Scales data available (Supplementary Fig. S1). Written informed consent was obtained from parents/guardians of study participants. Ethical approval was granted by the Institutional Review Boards at Makerere University School of Medicine and University of Minnesota. Additional regulatory approval was obtained from The Uganda National Council of Science and Technology. Comparisons between categorical variables were assessed using Pearson χ2 and one-way analysis of variance for continuous variables. Pairwise comparisons of continuous variables were assessed using Student’s t-test. Logistic regression was used to assess the odds ratio (OR) and 95% CI of mortality by nutritional status (stunting, underweight and wasting) at enrollment. Negative binomial regression was used to assess incidence rate ratio (IRR; 95% CI) of hospital readmission by nutritional status at enrollment. Multiple linear mixed effects (LMEs) models were used to evaluate the correlation between demographic or social factors and weight-for-age, height-for-age and weight-for-height z-scores over the full 24-month study period. Within subject observations were correlated using a subject specific intercept and time-points were treated as categorical variables. A banded diagonal covariance matrix was used to model the within-subject variance–covariance errors, and the mixed model was fit by restricted maximum likelihood. Kenward–Roger approximations were used to estimate the denominator degrees of freedom. Age at enrollment, sex, SES, home environment, weekly meat consumption, preschool education, maternal and paternal education were evaluated as independent predictor variables. To assess whether a child’s nutritional status at enrollment was associated with differences in cognition in children with CM, SMA or in CC, we used multiple LME models with child cognitive z-score as the dependent variable. Cognitive testing at all time points was incorporated. Cognitive models included a subject specific intercept and time-points were treated as a categorical variable. We adjusted for age at enrollment, sex, SES, home environment, preschool education, maternal and paternal education. To adjust for multiple comparisons, the Benjamini–Hochberg false discovery rate was applied at a threshold of 0.05. All analyses used Stata version 15 (StataCorp, College Station, TX).
