Neurodevelopmental performance among pre-schoolers treated for severe anaemia at Lira Regional Referral Hospital, Uganda

Background Severe anaemia is a common clinical problem among young children in sub-Saharan Africa. However, the effect of severe anaemia on neurodevelopment of these children is not well described. Therefore, we assessed the neurodevelopmental performance of preschool children diagnosed with severe anaemia in Northern Uganda. Methods We conducted a prospective cohort study among children < 5 years of age 14 days post discharge after an episode of severe anaemia (Hb < 5.0 g/dl; n = 171; mean Hb = 3.9g/dl) at Lira Regional Referral Hospital, Uganda. Neurodevelopmental outcomes (cognitive, language and motor) were assessed using Bayley Scales of Infant and Toddler Development, 3rd edition (Bayley-III). Age-adjusted z-scores for each domain were calculated using scores from healthy community control children (n = 88) recruited from the same environment for each age category. Multiple linear regression was used to compare z-scores in the cognitive, language and motor scales between the two groups after adjusting for weight-for-age z-score, socioeconomic status, mother’s education, and father’s employment on all the scales. Results The prevalence of neurodevelopmental impairment was 2.3% (95% CI: 0.8–6.1) for cognition, 1.7% (95%: 0.6–5.3) for language and 3.5% (95% CI: 1.6–7.6) for motor scales and 4.6% (95% CI: 2.3–9.1) for deficits in ≥1 area of neurodevelopment. Significant differences were observed between the two groups with the SA group performing worse on cognition [adjusted mean score, (Standard error, SE), P-value] [-0.20, (0.01) vs. 0.00, (0.01), P = 0.02]; language [-0.25, (0.01) vs. 0.00, (0.01), P< 0.001]; and motor [-0.17, (0.01) vs. 0.00, (0.01), P = 0.05] scales. Conclusion In children < 5 years of age, severe anaemia was associated with neurocognitive (cognition, language and motor) deficits in the immediate period post treatment. Further research is needed to identify risk factors and determine the long-term effects of poor neurodevelopment in young children with severe anaemia.


Introduction
Severe anaemia (SA) defined as haemoglobin (Hb) < 5g/dl is a global public health challenge commonly associated with childhood morbidity and mortality among children < 5 years of age in sub-Saharan Africa [1,2]. Anaemia prevalence is estimated at 47.4% globally among preschool children with a burden of 62.3% among African children [3,4]. In Uganda, the prevalence of anaemia among children < 5 years of age is estimated at 37.2% in the Northern part of Uganda [5]. The common causes of SA are multi-factorial including severe malaria, poor nutrition, micronutrient deficiencies (i.e. iron deficiency), sickle cell anaemia, acute illnesses and infections, but malaria is a major driver [6][7][8]. Despite several interventions to improve the iron status of children and reduce the anaemia burden, SA remains a challenge with variability in clinical management and community perceptions that affect seeking care and treatment [9,10]. These interventions include blood transfusion [11,12], iron supplementation/ giving haematinics [13,14], de-worming [15], and malarial presumptive treatment [16,17]. Severe anaemia in children can affect cerebral oxygen supply causing prolonged, repetitive and acute hypoxic-ischaemic events leading to cerebral damage, abnormalities and lesions in the basal ganglia, thalami, white and grey matter [18,19]. These are significant risk factors for poor early childhood neurodevelopment. However, data on the association between SA and neurodevelopmental impairment is still limited. With improved survival rates and lowered mortality among children <5 years of age in our setting [10], we may potentially have more numbers of children surviving with neurodevelopmental impairment persisting long after treatment and resolution of anaemia.
Deficits in neurocognition in children may significantly impact future learning, class performance and academic attainment [20]. However, in routine clinical practice in the African setting, assessment of neurodevelopment in children surviving illnesses like SA is not integrated in routine clinical care [21,22]. An earlier study that examined neurodevelopmental outcomes amongst Ugandan children < 5 years surviving severe malaria anaemia (SMA), a form of complicated malaria showed that SMA was associated with long-term cognitive impairment [23]. However, it remains unclear if the reductions in scores may be similarly observed for children surviving SA without malaria in the short-term post recovery. Given the association between anaemia and impaired cognition [24,25], severe anaemia may also affect neurodevelopment significantly affecting full developmental growth. It is therefore essential to examine the association between exposure to SA and neurodevelopmental outcomes.
There is scanty literature to date on neurodevelopmental outcomes among children < 5 years of age with SA without malaria. Previous studies have not particularly focused on the much younger children in their first 1000 days of life, a time of critical brain growth, neural circuitry and development [26,27]. Therefore, we aimed to estimate the prevalence of neurodevelopmental (cognition, language and motor) deficits among children below 5 years surviving SA at Lira Regional Referral Hospital in Northern Uganda where the prevalence of SA is 33.5% [28]. We hypothesized that children surviving SA will have poorer neurodevelopmental outcomes on all domains compared to healthy controls.

Materials and methods
This prospective cohort study was conducted at Lira regional referral hospital, Northern Uganda between August 2016 and June 2017. The hospital is a free-for-care hospital serving numerous districts in an area of all year high malaria transmission around the Lake Kyoga region [29]. Children are first seen in an outpatient section where those requiring admission are identified and taken to the paediatric ward as described elsewhere [10,28].

Study participants
Children aged 6-42 months with SA (n = 180) with mean Hb = 3.9g/dl were consecutively enrolled from an ongoing implementation research study where SA was defined as Hb < 5g/dl [10,28]. Children above 42 months and those with known sickle cell disease (a recognised specific cause of severe anaemia; equally associated with neurologic disability) [30] were excluded. All children with SA received a blood transfusion treatment at admission following the prevailing Ugandan Ministry of Health guidelines at the time of the study [31]. Classification, clinical guidelines and treatment management of these children is described elsewhere [10,28]. Parents of children with SA were informed about the need for healthy control children of the same age from their nuclear or extended family (siblings) within the household compound area or neighbourhood (playmate) to generate a comparison/ control group. They were requested to bring the eligible children along with their caregiver (if different from child with SA) to the hospital when they returned for their assessment.
The healthy community controls (CCs) (n = 90) were of similar age (6-42 months) as children with SA and currently healthy. The CCs were recruited mainly as a normative group to generate z-scores and to serve as comparison with children with SA. The CCs were examined at the time of enrolment to ensure they were healthy and did not have: (1) clinical pallor on clinical examination; or (2) a history of hospitalization for severe anaemia 6 months or other illnesses 4 weeks prior to enrolment; or (3) major neurological, developmental or clinical abnormalities on physical examination. The participants were grouped to create appropriate and balanced 4 age bands (6-12 mo.; 12.1-24 mo.; 24.1-36 mo.; 36.1-42 mo.). The controls were recruited as a normative group to generate z-scores with at least 20 controls for each age band. The control group provided a comparison for the SA group and was intended to reduce norm related bias and provide a local normative reference score of the children using Bayley Scales of Infant and Toddler Development, Third Edition (Bayley-III) [32].

Sample size
Sample size was calculated done assuming a prevalence of cognitive deficits in Ugandan children at 26% [33]. A sample size of 180 children with severe anaemia was calculated to have 80% power and delta (decrease in prevalence) = 0.2, at 0.05 level of significance to demonstrate a two-fold increase in the risk of neurodevelopmental dysfunction in severe anaemia children. This sample size did allow for 4% non-responsiveness. Additionally, we recruited 90 controls with at least 20 controls for each of the 4 age bands as described above.

Demographic and clinical assessment
Demographic and socioeconomic status (SES) data were collected using a questionnaire checklist of material possessions previously used among the Ugandan paediatric population [34]. We assessed items involving nutritional status, child's education level, caregiver education level, housing quality in which lower SES scores have been associated with worse cognitive functioning in healthy Ugandan children < 5 years old [34]. SES questionnaire has a checklist of indicators related to material possessions, house structure, living density, food resources and access to electricity, transport facility and clean water. Each indicator was given a score, and these scores are added to give a total SES score. A composite SES score using household's assets was determined and children categorized into wealth quintiles using principal components analysis [35,36]. Height was obtained using a stadiometer with a sliding horizontal headpiece and weight using a paediatric calibrated weighing scale. Physical growth z-score anthropometric indicators (weight-for-age (WAZ), height for age (HAZ) and weight for height (WHZ) were used to evaluate nutritional status using WHO Anthro survey analyser software for children under 5 years of age; published norms and standardized z-scores. The WAZ, HAZ, and WHZ scores of less than −2 were categorized as underweight, stunting and wasting respectively [37,38].

Neurodevelopmental assessment
The Bayley Scales of Infant and Toddler Development, 3 rd Edition (Bayley-III) [39] were used for neurodevelopmental assessment at 14 days post discharge after clinical recovery for the children with SA and at enrolment for the community controls or when appropriate for the caregiver to bring the child to the hospital for assessment. All assessments were conducted in a quiet and child friendly room at the clinic and lasted approximately 50-70 minutes. The assessments were done in Langi, a local dialect as most children and caregivers were more fluent in Langi than English by competent psychology graduates trained in the child assessment and the use of Bayley-III. The two assessors were blinded to the child's study group. The first author (Health psychologist) supervised the administration of tests and reviewed the record forms at the end of each week for completeness to ensure data quality. The assessors were periodically given review trainings by a psychology graduate independent of the study and with expertise in Bayley-III. For consistency, the testers were video recorded conducting the assessments at the beginning-, mid-and end of the study and these were reviewed, discussed and feedback shared by the two testers and first author.
The Bayley-III is a widely used play-based standardized assessment measure of neurodevelopment in young children 1-42 months on various domains [39]. It evaluates a child in three areas: Cognitive (91 items), Language (97 items; subdivided into receptive-49 items and expressive communication-48 items subtests), and Motor 138 items; subdivided into Fine-66 items and Gross motor-72 items subtests). The Bayley-III tool used for neurodevelopmental assessment has not been validated for our setting however has been used in other studies and adapted for appropriate use among children in rural Uganda [40][41][42].

Statistical methods
Data were entered into Filemaker 11.0 v3 (FileMaker Inc. US) database, and exported into IBM SPSS 23 for statistical analysis. We compared the children with severe anaemia to healthy controls on the three neurodevelopmental summary scores (cognition, language and motor).
We converted raw scores for each scale into an age-specific standardized z-score, based on the scores of CCs since age has an effect on neurodevelopmental assessments where younger children will have lower raw scores than older children [33]. The z-scores were computed as (actual score-mean score for a child's sex and age)/SD, where the mean score for a child's sex and age and SD were computed by fitting a linear regression model to data for all CC children (n = 88) [23]. Z-scores have a mean of 0 and SD of 1 in the CC reference population. For all the scales, deficit was defined as a z-score of less than -2. Neurodevelopmental deficit was defined as deficit in more than 1 area (cognition, language and motor) as there was no way of generating an overall neurodevelopmental score from the primary data of scores in individual areas. This has been used in previous cognitive studies among Ugandan children [33,43]. Neurodevelopmental impairment prevalence for each domain was computed by dividing the number of children who had cognitive impairment in the SA group by the total number of children in the SA group.
Multiple linear regression was used to compare age-and sex-adjusted z-scores between the two groups while adjusting for weight-for-age z-score, socioeconomic status, mother's education, and father's employment (that differed between the two groups) for all the scales. We used the Holm-Bonferroni method (Holm procedure) to adjust for comparisons due to multiple testing [44].

Ethical considerations
Makerere

Characterization of the study children
One hundred eighty children with SA were recruited for the study. Of these, 06 did not return for assessment at day 14 post-discharge, leaving 174 children. During the assessment visit, 03 children had incomplete assessments, leaving 171 children who were analysed for neurodevelopmental performance. Additionally, 90 community control children were invited and consented to participate in the study. All the 90 children reported for their assessment however, two (02) children did not complete their assessments and 88 CCs were analysed for this study (Fig 1).
Of and 47% were girls. The pre-schoolers assessed in this study resided in the same geographical region. The socio-demographic characteristics of the study participants are represented in Table 1.
With adjustment for multiple testing using Holm's procedure, significant differences were observed between the two groups with the SA group compared to CCs performing worse on cognition  (Table 3).

Discussion
The study assessed the prevalence of neurodevelopmental (cognition, language and motor) deficits among preschool children aged 6 to 42 months shortly after an episode of SA at Lira Regional Referral Hospital in Northern Uganda. The study found that children surviving SA had lower age unadjusted z-scores in the cognitive and language domains compared to CCs. After adjusting for SES, WAZ, maternal education and paternal employment and adjusting for multiple testing using Holm's procedure [44], children with SA had significantly lower scores in all the three domains (cognitive, language and motor).
In our study, SA was associated with small acute decreases in neurodevelopmental outcomes after multiple comparisons adjustment. The poor neurodevelopmental scores among children with SA compared to CCs may be subtle as we assessed the participants during recovery from severe anaemia. The study results reflect the potential effect of severe anaemia among pre-schoolers in resource-limited settings where the prevalence of SA among children is notably high [4]. Therefore, the current study provides evidence to date that SA is associated with immediate neurodevelopmental deficits in pre-schoolers in Northern Uganda. Hare [45] also observed that SA was associated with cerebral dysfunction and injury as the brain is vulnerable to anaemia-induced injury. Impaired cognitive function and memory among SA sufferers has been linked to a reduction of cerebral blood which affects neuronal function [25].
SA among children <5 years happens at a time of rapid child growth and development and increased iron requirements critical in proper brain growth with the potential to affect neurodevelopmental outcomes [46][47][48]. Children need iron to make haemoglobin, which carries oxygen to all body cells and iron deficiency has been associated with SA in pre-schoolers [49]. Therefore, the lack of iron during times of SA leads to decreased iron bioavailability and haemoglobin concentration to the central nervous system (CNS) thereby decreasing oxygen solubility of whole blood; inhibiting oxygen transportation [49]. This could lead to significant deficits in domains sensitive to CNS iron levels leading to neurodevelopmental impairment [50]. Additionally, low haemoglobin concentrations adversely affect cognitive and motor development which are important risk factors for the health and development among children [51].
Our study found that children with SA were at risk of poor overall language performance specifically in expressive communication implying that children with SA may have limited ability to vocalise and perform tasks such as object naming, babbling, and gesturing strongly associated with children's overall learning, academic skills and later development [52]. Language has been shown to facilitate learning and profiling a child's skills in other domains and prediction of a child's later socioemotional, pre-linguistic and comprehension skills [52]. These early language deficits in pre-schoolers with SA may affect their cognitive and motor abilities.
Our study findings agree with Santos, Rates [53] showing that childhood anaemia has the propensity to lead to language development alterations thereby affecting social and learning abilities among pre-schoolers. Furthermore, expressive communication deficits could account for deficits in cognitive development where early child language development ability has strong associations with cognitive functioning, verbal and nonverbal skills, later development and academic achievement at school age [25,54]. This may also affect the way in which the pre-schoolers attain language, use self-expressions when with peers of the same age and affects self-esteem and confidence.
Severe anaemia remains one of the significant public health problems among preschool children worldwide particularly in malaria prone regions and the alleviation of this condition has the potential to improve both the health of children and their neurodevelopment [23,55,56]. Identifying abilities that were affected provides information on how SA affects neurodevelopment and the domains that need to be evaluated post SA especially in children under 5 years. Therefore, understanding neurodevelopmental outcomes among children with SA may provide more insight on the nature and types of policies needed to address these potential child health and developmental challenges. The present study suggests that effects of SA are seen early in pre-schoolers below 5yrs of age, as lower adjusted z-scores were seen 14 days post-discharge. From our findings, SA among children below 5 years is associated with poor neurodevelopmental performance even in the context of decreasing malaria prevalence. Therefore, designing interventions that include neurodevelopmental assessment, especially in areas with high SA prevalence, could be of added value. The study results may provide direction for intervention planning and support strategies. Correction of SA among children under 5 years will enhance better neurodevelopmental outcomes supporting school performance and contributing to better health, quality of life and economic outcomes later in life [51]. Cognitive, behavioural, social and educational problems in later life could be alleviated when early identification of neurodevelopmental impairment and those at greater risk and appropriate interventions implemented among this population [51].
Our study examined neurodevelopmental performance among pre-schoolers surviving severe anaemia in Uganda using Bayley III as the primary neurodevelopmental outcome measure, which is a fairly recent neurodevelopmental performance assessment tool used worldwide and in our settings [40,41]. However, it still needs to be standardised and culturally adapted for low resource settings like Uganda.
Certain limitations should be considered in the interpretation of the findings of the present study. It was impossible to collect and compare neurodevelopmental scores of the SA children studied before the illness. Therefore we compared children with SA to age-matched CCs to cater for other factors i.e. socioeconomic status that may interfere with the neurodevelopmental outcomes [34]. Bayley-III has not been standardized and validated in Uganda, however we converted the raw scores into z-scores for each age group based on scores of CCs to make the assessment appropriate for this study population. We lacked Hb assessment for the CC group as at the time of the study it was only considered for the severely febrile and admitted children Note # Overall language z-score is a combined score of receptive communication and expressive communication scores.
� Overall motor z-score is a combined score of fine motor and gross motor scores.
Age-and sex-adjusted z-scores were computed using CC children as the reference population.
Mean z-scores were adjusted for nutrition status (weight-for-age), maternal education, paternal employment and socioeconomic status.
All p-values have been adjusted for multiple testing using Holm's procedure [Holm. S. (1979)].
https://doi.org/10.1371/journal.pone.0240694.t003 due to limited resources at the hospital. We also lack data related to our study sample on the underlying cause of severe anaemia common among African pre-schoolers i.e. helminth, iron deficiency, malnutrition, severe malaria, [49,51] which could have interfered with our findings as they are potential for neurodevelopmental decline in children. Further study is required to determine whether the cognitive differences seen are due to severe anaemia or to factors such as iron deficiency or malaria that may lead to severe anaemia and also to cognitive impairment. We suggest conducting a longitudinal study following up the children to their school age years.

Conclusion
This study demonstrates that severe anaemia is associated with neurocognitive (cognition, language and motor) deficits among infants and pre-schoolers. The poor performance can be observed from as early as 6 months of age. The study findings suggest the importance of early screening and detection of neurodevelopmental deficits in pre-schoolers post illness and admission which can facilitate timely interventions for these children and others with other related conditions i.e. iron deficiency, severe malaria, and malnutrition. We propose further more rigorous investigations to examine the effect of severe anaemia in relation to long-term neurodevelopmental outcomes, as the children get older.
Supporting information S1 File. Final Z scores by regression for neurodevelopmental outcomes data. File containing primary data used in the analysis of cognitive, language and motor outcomes.