ABSTRACT.
Malaria in pregnancy (MIP) causes poor birth outcomes, but its impact on neurocognitive development has not been well characterized. Between 2012 and 2014, we enrolled 307 mother–infant pairs and monitored 286 infants for neurocognitive development using the Malawi Developmental Assessment Tool at 6, 12, and 24 months of age. MIP was diagnosed from peripheral blood and placental specimens. Cord blood cytokine levels were assessed for a subset of neonates. Predictors of neurodevelopment were examined using mixed-effect logistic regression for developmental delay. Among the participants, 78 mothers (25.4%) had MIP, and 45 infants (15.7%) experienced severe neurocognitive delay. MIP was not associated with differences in cord blood cytokine levels or neurocognitive development. Preterm birth, low birthweight, increasing maternal education level, and increasing interleukin 6 levels were associated significantly with delay. The results highlight the prevalence of severe delay and a need for broad access to early childhood support in this setting.
INTRODUCTION
Malaria in pregnancy (MIP), including Plasmodium falciparum infection in the peripheral blood or placenta, is common and a well-characterized cause of poor birth outcomes in malaria-endemic areas.1 However, long-term effects of MIP have not been well evaluated.
MIP is a known risk factor for low birthweight (LBW),2–4 which is associated with increased risk of infant death and has been linked to developmental delay.5–7 MIP may also lead to developmental delay independently of LBW.8 MIP can induce an inflammatory intrauterine environment and restrict fetal growth, both of which could impair fetal brain development. Elevated maternal levels of inflammatory cytokines may lead to abnormal neonatal brain development and cognitive deficits in early childhood.9,10 During pregnancy, elevated inflammatory cytokines are more common in women with MIP compared with those without it,11 although it remains unclear whether MIP-associated cytokine elevation is sufficient to disturb fetal brain development in humans. In a murine model, placental malaria was associated with placental inflammation and persistent deficits in memory and affective behavior in the offspring.12
Few studies have assessed the effect of MIP on neurocognitive development.8,13–15 More data are needed to determine which inflammatory markers may be related to neurodevelopment. Using data from a mother–infant cohort monitored prospectively from pregnancy to 24 months after birth, we tested the association between MIP and neonatal cytokine levels on neurodevelopment.
METHODS
Mother–infant pairs were recruited from participants enrolled in a randomized controlled trial in Blantyre, Malawi, between 2012 and 2014 (ClinicalTrials.gov Identifier: NCT01443130) described previously.16 Women enrolled in the trial were HIV negative, in their first or second pregnancy, and were monitored from 20 to 26 weeks’ gestation until delivery. The occurrence of malaria and LBW did not differ across the trial study arms. Singleton infants of trial participants were offered enrollment in the study described here and were monitored to 24 months.
Women had monthly prenatal examinations and were encouraged to visit the clinic whenever ill. Finger-prick dried blood spots (DBSs) on filter paper for detection of peripheral malaria were collected at every prenatal visit; DBSs for detection of placental malaria were collected from placental blood. Serum obtained from cord blood was cryopreserved at –80°C for cytokine quantification. Full-thickness placental biopsies were collected and examined for the presence of parasites or hemozoin pigment. Real-time quantitative polymerase chain reaction (qPCR) for P. falciparum 18S recombinant RNA was used to detect malaria from DBSs.17 Peripheral malaria was defined as the detection of malaria from a finger-prick DBS by qPCR during pregnancy. Placental malaria was defined as the presence of malaria hemozoin pigment or parasites in the placenta, or parasites detected by qPCR in the placental blood. Mothers who had either peripheral or placental malaria detected during pregnancy were defined as having “any malaria” for analysis. Gestational age was measured in the second trimester by ultrasound.16
Neonates with at least 1 year of follow-up were eligible for cytokine quantification. Cytokines in cord blood were quantified for all eligible neonates born to mothers with MIP (n = 23), and from 25 randomly selected deliveries of women with no MIP. Meso Scale Discovery (MSD) (Meso Scale Diagnositcs, Rockville, MD) plates were used to quantify the concentrations of interferon-γ, interleukin (IL) 13, IL-12, IL-10, IL-1β, IL-2, IL-4, IL-6, and tumor necrosis factor-α (human V-PLEX Proinflammatory Panel 1 Kit); C-reactive protein (CRP; human V-PLEX Vascular Injury Panel 2 Kit); and transforming growth factor (TGF)-β (MSD 96-well multi-array human TGFβ1 assay). Analytes were detected according to the manufacturer’s instruction. Serum was diluted 1:2 for the Proinflammatory Panel, 1:1000 for the Vascular Injury Panel, and 1:8 for the TGF-β assay. Plate readings were performed on an MSD QuickPlex SQ120 imager (Meso Scale Diagnositcs), and raw data were analyzed using MSD Discovery Workbench version 4 (Meso Scale Diagnositcs).
Children were evaluated at 6, 12, and 24 months using the Malawi Developmental Assessment Tool (MDAT), which assesses neurocognitive development across four domains: gross motor, fine motor, language, and social.18 The MDAT was developed and validated for use in Malawi and has been used in more than 15 African countries for both programmatic and research purposes.18–21 MDAT z-scores were calculated using gestational age-specific means and variances of MDAT domain scores from a reference population of Malawian children.18 MDAT results were analyzed three ways: 1) continuous distribution of z-scores, 2) any developmental delay (z-score < –1), and 3) severe developmental delay (z-score < –2).
The following covariates were examined for predictors of neurodevelopmental delay: trial arm, maternal education level, socioeconomic status from principal components analysis,22 preterm birth (birth < 37 weeks’ gestation), and LBW (< 2,500 g at birth).
We examined predictors of neurodevelopment by domain at any time point using χ2 tests for dichotomized MDAT results and mixed-effect logistic regression for delay over time, accounting for clustering resulting from repeated measures among individuals.
Differences in cytokine levels between those with and without MIP were assessed using Wilcoxon rank sum tests. Undetectable samples were assigned a value =1/2 lower limit of detection (Supplemental Table S1).
We used random forest machine learning to select the cytokines that best predicted domain-specific delay or delay in any domain. Predictive cytokines identified by random forests were selected for mixed-effect logistic regression to determine the association between important cytokines and delay, examining each predictive cytokine individually and adjusting for other predictive cytokines and preterm birth. Logistic regression was only done for the outcome “any delay’” as a result of few observations with severe delay. Cytokines were modeled as continuous if cytokine levels had a linear relationship with delay; otherwise, cytokine levels were categorized.
RESULTS
Participant characteristics of 307 mother–infant pairs are described in Table 1. Higher maternal education level and socioeconomic status were associated with decreased rates of MIP. MIP was not associated with either preterm birth or LBW.
Table 1.
Enrollment characteristics of 307 Malawian mother–infant pairs by maternal malaria status
| Characteristic | n (%) | Maternal malaria status, n (%) | |||
|---|---|---|---|---|---|
| None | Any malaria | Placental | Peripheral | ||
| n (% of total) | 307 (100) | 229 (74.6) | 78 (25.4) | 39 (12.7) | 39 (12.7) |
| Maternal age, years; mean (SD) | 21.1 (3.5) | 21.3 (3.6) | 20.5 (3.2) | 19.5 (2.6) ‡ | 21.5 (3.4) |
| Maternal education | – | – | † | † | – |
| Less than high school | 92 (30.0) | 62 (27.1) | 30 (38.5) | 16 (41.0) | 14 (35.9) |
| Some high school | 102 (33.2) | 71 (31.0) | 31 (39.7) | 17 (43.6) | 14 (35.9) |
| High school diploma or higher | 113 (36.8) | 96 (41.9) | 17 (21.8) | 6 (15.4) | 11 (28.2) |
| Marital status | |||||
| Married, living with spouse | 237 (77.2) | 175 (76.4) | 62 (79.5) | 31 (79.5) | 31 (79.5) |
| Single | 28 (9.1) | 20 (8.7) | 8 (10.3) | 4 (10.3) | 4 (10.3) |
| Other | 42 (13.7) | 34 (14.9) | 8 (10.3) | 4 (10.3) | 4 (10.3) |
| Socioeconomic status | – | – | † | † | † |
| Low | 77 (25.1) | 48 (21.0) | 29 (37.2) | 16 (41.0) | 13 (33.3) |
| Medium | 169 (55.0) | 130 (56.8) | 39 (50.0) | 16 (41.0) | 23 (59.0) |
| High | 61 (19.9) | 51 (22.3) | 10 (12.8) | 7 (17.9) | 3 (7.7) |
| Primigravid | |||||
| Yes | 168 (54.7) | 128 (55.9) | 40 (51.3) | 21 (53.9) | 19 (48.7) |
| No | 139 (45.3) | 101 (44.1) | 38 (48.7) | 18 (46.2) | 20 (51.3) |
| Infant gender | |||||
| Female | 150 (48.9) | 111 (48.5) | 39 (50.0) | 18 (46.2) | 21 (53.9) |
| Male | 157 (51.1) | 118 (51.5) | 39 (50.0) | 21 (53.9) | 18 (46.2) |
| Preterm birth | |||||
| Full term | 293 (95.4) | 216 (94.3) | 77 (98.7) | 39 (100) | 38 (97.4) |
| Preterm | 14 (4.6) | 13 (5.7) | 1 (1.3) | 0 (0.0) | 1 (2.6) |
| Low birthweight | |||||
| No | 280 (91.2) | 206 (90.0) | 74 (94.9) | 38 (97.4) | 36 (92.3) |
| Yes | 27 (8.8) | 23 (10.0) | 4 (5.1) | 1 (2.6) | 3 (7.7) |
P < 0.05.
P < 0.01.
P < 0.001.
Of 286 infants with MDAT scores, 45 (15.7%) experienced severe delay in any domain during the study follow-up. “Any delay” was prevalent (24.8%) in the cohort; however, MDAT z-scores increased significantly across all domains over 24 months (Supplemental Figure S1). In bivariate analysis, infant gender, LBW, preterm birth, study arm, and increasing maternal education level were associated with severe delay in at least one domain (Table 2). Study arm, preterm birth, and LBW were associated independently with a lower MDAT z-score in the fine motor domain and any domain. In longitudinal models, most predictor variables were not associated with severe developmental delay (Figure 1).
Table 2.
Number of infants with severe delay (age-adjusted z-score < –2) at any time point by Malawi Developmental Assessment Tool (MDAT) domain among 286 infants with MDAT data
| Variable | Severe gross motor delay, n (%) | Severe Fine Motor Delay | Severe Language Delay | Severe Social Delay | Any Severe Delay |
|---|---|---|---|---|---|
| Total (N = 286) | 23 (8.04) | 33 (11.5) | 16 (5.6) | 15 (5.2) | 45 (15.7) |
| Maternal malaria status | |||||
| No malaria (n = 218) | 21 (9.6) | 27 (12.4) | 13 (6.0) | 13 (6.0) | 37 (17.0) |
| Any malaria in pregnancy (n = 68) | 2 (2.9) | 6 (8.8) | 3 (4.4) | 2 (2.9) | 8 (11.8) |
| Placental malaria (n = 34) | 1 (2.9) | 2 (5.9) | 2 (5.9) | 1 (2.9) | 3 (8.8) |
| Peripheral malaria (n = 34) | 1 (2.9) | 4 (11.8) | 1 (2.9) | 1 (2.9) | 5 (14.7) |
| Trial Arm | – | * | – | – | * |
| CQ IPT (n = 92) | 5 (5.4) | 4 (4.4) | 3 (3.3) | 2 (2.2) | 8 (8.7) |
| CQ prophylaxis (n = 100) | 9 (9.0) | 18 (18.0) | 5 (5.0) | 4 (4.0) | 23 (23.0) |
| SP IPT (n = 94) | 9 (9.6) | 11 (11.7) | 8 (8.5) | 9 (9.6) | 14 (14.9) |
| Infant gender | – | – | – | * | – |
| Female (n = 139) | 12 (8.6) | 14 (10.1) | 7 (5.0) | 4 (2.9) | 22 (15.8) |
| Male (n = 147) | 11 (7.5) | 19 (12.9) | 9 (6.1) | 11 (7.5) | 23 (15.7) |
| Low birthweight | – | * | – | – | * |
| Yes (n = 26) | 3 (11.5) | 6 (23.8) | 3 (11.5) | 2 (7.7) | 8 (30.8) |
| No (n = 260) | 20 (7.7) | 27 (10.4) | 13 (5.0) | 13 (5.0) | 37 (14.2) |
| Preterm birth | – | – | – | – | * |
| Yes (n = 13) | 2 (15.4) | 3 (23.1) | 2 (15.4) | 1 (7.7) | 5 (38.5) |
| No (n = 273) | 21 (7.7) | 30 (11.0) | 14 (5.1) | 14 (5.1) | 40 (14.7) |
| Primigravid | |||||
| Yes (n = 155) | 12 (7.7) | 17 (11.0) | 8 (5.2) | 9 (5.9) | 24 (15.5) |
| No (n = 131) | 11 (8.4) | 16 (12.2) | 8 (6.1) | 6 (4.6) | 21 (16.0) |
| Marital status | |||||
| Married, living with spouse (n = 222) | 19 (8.6) | 25 (11.3) | 13 (5.9) | 12 (5.4) | 35 (15.8) |
| Single (n = 25) | 1 (4.0) | 5 (20.0) | 1 (4.0) | 1 (4.0) | 5 (20.0) |
| Other (n = 39) | 3 (7.7) | 3 (7.7) | 2 (5.1) | 2 (5.1) | 5 (12.8) |
| Socioeconomic status | |||||
| Low (n = 70) | 4 (5.7) | 9 (12.9) | 4 (5.7) | 4 (5.7) | 10 (14.3) |
| Medium (n = 160) | 15 (9.4) | 19 (11.9) | 9 (5.6) | 9 (5.6) | 27 (16.9) |
| High (n = 56) | 4 (7.1) | 5 (8.9) | 3 (5.4) | 2 (3.6) | 8 (14.3) |
| Maternal education | * | – | – | * | – |
| Less than high school (n = 82) | 2 (2.4) | 7 (8.5) | 2 (2.4) | 1 (1.2) | 8 (9.8) |
| Some high school (n = 95) | 8 (8.4) | 12 (12.6) | 6 (6.3) | 5 (5.3) | 15 (15.8) |
| High school diploma or higher (n = 109) | 13 (12.0) | 14 (2.8) | 8 (7.3) | 9 (8.3) | 22 (20.2) |
CQ = chloroquine; IPT = intermittent preventive treatment; SP = sulfadoxine-pyrimethamine.
P < 0.05.
Figure 1.
Odds ratios for possible predictors of severe developmental delay (z-score < –2) at any time point by birth exposure from unadjusted mixed-effect logistic regression models. The vertical dotted line indicates an odds ratio equal to 1.0. Any malaria, placental malaria, and peripheral malaria compared with no malaria in pregnancy. Single and other compared with married. Medium (Med) and high socioeconomic (SES) compared with low SES. Maternal education levels compared with individuals with some primary education only. Chloroquine (CQ) prophylaxis and sulfadoxine-pyrimethamine (SP) intermittent preventive treatment in pregnancy (IPTp) compared with CQ IPTp. LBW = low birthweight. This figure appears in color at www.ajtmh.org.
There was no association between MIP and MDAT z-scores in any domain at any time (Supplemental Table S2), nor was there any association between MIP and severe delay on bivariate analysis (Table 2), longitudinal analysis (Figure 1, Supplemental Table S3), or when adjusted for gender, LBW, preterm birth, study arm, or maternal education level (Supplemental Table S4). Results were similar when examining “any delay” (Supplemental Tables S5 and S6, and Supplemental Figure S2).
IL-13 levels were significantly less (P = 0.034) among infants born to mothers with MIP than those born to mothers without any MIP. Random forests models predicted developmental delay with an accuracy ranging from 78% to 90%. Important cytokines identified from random forest models were consistent between models when using either “any delay” or “severe delay” as the outcome. No cytokines were associated with “delay” in unadjusted models (Supplemental Figure S3). In adjusted models, a greater IL-6 level was associated with increased odds of gross motor developmental delay (odds ratio, 7.64; 95% CI,1.21–48.39), and a lower CRP level was associated with increased odds of gross motor developmental delay (odds ratio, 13.02–95% CI, 1.14–148.8). Other cytokines tested were not associated significantly with either developmental delay or MIP, and adjustment for preterm birth did not alter results.
DISCUSSION
MIP was not correlated with neurodevelopmental outcomes in this population of Malawian infants. Nor was MIP associated with previously identified mediators of fetal brain injury, including cord blood cytokines and preterm birth. However, preterm birth and LBW were associated with delayed cognitive development, supporting the validity of MDAT assessments that can, in this context, detect developmental delay.
We found that greater rates of IL-6 and lower levels of CRP in cord blood were associated with increased risk of neurodevelopmental delay. Previous studies examining the association between CRP at birth and neurocognitive outcomes have yielded conflicting results.8 We also found low levels of IL-13 among infants born to mothers with MIP, potentially indicating immune priming in utero.
Our study was limited by small numbers, particularly in the analysis of cord blood cytokines, and may have been underpowered to detect small effects, leading to large CIs for estimates. In addition, postnatal factors including adverse child experiences and family situation, contribute significantly to neurodevelopmental delays, potentially confounding our findings. The timing of MIP is likely important for predicting both inflammatory responses and neurodevelopment; however, we were unable to account for timing of exposure to MIP because of inadequate numbers. Last, all mothers included in this study received malaria chemoprophylaxis during pregnancy, presumably limiting MIP exposure and lessening the likelihood of finding an association.
Few previous community studies exist in African settings demonstrating the link between preterm birth and development, particularly with reliable ultrasound results. Although nearly all children achieved normal development by 24 months of age, severe delay was common at 6 and 12 months, and was associated with preterm birth. Given the importance of the postnatal environment for child growth and development, our findings highlight the need for broad access to early identification and support of infants at risk of delay in this setting.
Supplemental files
ACKNOWLEDGMENTS
We thank the women who volunteered to participate in this study, and the nurse–midwives of the Ndirande Health Centre maternity ward who supported this study. We are grateful to the clinical team of the Blantyre Malaria Project Ndirande clinic for their dedication and hard work on this study. The authors confirm that all ongoing and related trials for this drug/intervention are registered (#NCT01443130). This trial is registered at ClinicalTrials.gov (#NCT01443130, https://clinicaltrials.gov/ct2/show/NCT01443130).
Note: Supplemental figures and tables appear at www.ajtmh.org.
REFERENCES
- 1. Rogerson SJ Desai M Mayor A Sicuri E Taylor SM van Eijk AM , 2018. Burden, pathology, and costs of malaria in pregnancy: new developments for an old problem. Lancet Infect Dis 18: e107–e118. [DOI] [PubMed] [Google Scholar]
- 2. Guyatt HL Snow RW , 2004. Impact of malaria during pregnancy on low birth weight in sub-Saharan Africa. Clin Microbiol Rev 17: 760–769. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. De Beaudrap P et al. 2013. Impact of malaria during pregnancy on pregnancy outcomes in a Ugandan prospective cohort with intensive malaria screening and prompt treatment. Malar J 12: 139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Akum AE Kuoh AJ Minang JT Achimbom BM Ahmadou MJ Troye-Blomberg M , 2005. The effect of maternal, umbilical cord and placental malaria parasitaemia on the birthweight of newborns from south-western Cameroon. Acta Paediatr 94: 917–923. [DOI] [PubMed] [Google Scholar]
- 5. Blencowe H et al. 2013. Preterm birth-associated neurodevelopmental impairment estimates at regional and global levels for 2010. Pediatr Res 74 (Suppl 1): 17–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Villar J et al. 2015. Postnatal growth standards for preterm infants: the Preterm Postnatal Follow-up Study of the INTERGROWTH-21(st) Project. Lancet Glob Health 3: e681–e691. [DOI] [PubMed] [Google Scholar]
- 7. Upadhyay RP Naik G Choudhary TS Chowdhury R Taneja S Bhandari N Martines JC Bahl R Bhan MK , 2019. Cognitive and motor outcomes in children born low birth weight: a systematic review and meta-analysis of studies from South Asia. BMC Pediatr 19: 1–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Weckman AM et al. 2021. Neurocognitive outcomes in Malawian children exposed to malaria during pregnancy: an observational birth cohort study. PLoS Med 18: e1003701. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Rudolph MD et al. 2018. Maternal IL-6 during pregnancy can be estimated from newborn brain connectivity and predicts future working memory in offspring. Nat Neurosci 21: 765–772. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Ghassabian A Albert PS Hornig M Yeung E Cherkerzian S Goldstein RB Buka SL Goldstein JM Gilman SE , 2018. Gestational cytokine concentrations and neurocognitive development at 7 years. Transl Psychiatry 8: 1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Fried M et al. 2017. Systemic inflammatory response to malaria during pregnancy is associated with pregnancy loss and preterm delivery. Nephrol Dial Transplant 65: 1729–1735. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. McDonald CR et al. 2015. Experimental malaria in pregnancy induces neurocognitive injury in uninfected offspring via a C5a–C5a receptor dependent pathway. PLoS Pathog 11: e1005140. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Conroy AL et al. 2019. Case report: birth outcome and neurodevelopment in placental malaria discordant twins. Am J Trop Med Hyg 100: 552–555. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Garrison A Boivin MJ Fiévet N Zoumenou R Alao JM Massougbodji A Cot M Bodeau-Livinec F , 2022. The effects of malaria in pregnancy on neurocognitive development in children at 1 and 6 years of age in Benin: a prospective mother–child cohort. Clin Infect Dis 74: 766–775. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Lawford HL et al. 2021. Associations between malaria in pregnancy and neonatal neurological outcomes. Int J Infect Dis 112: 144–151. [DOI] [PubMed] [Google Scholar]
- 16. Divala TH et al. 2018. Chloroquine as weekly chemoprophylaxis or intermittent treatment to prevent malaria in pregnancy in Malawi: a randomised controlled trial. Lancet Infect Dis 18: 1097–1107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Boudova S Divala T Mungwira R Mawindo P Tomoka T Laufer MK , 2017. Placental but not peripheral Plasmodium falciparum infection during pregnancy is associated with increased risk of malaria in infancy. J Infect Dis 216: 732–735. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Gladstone M Lancaster GA Umar E Nyirenda M Kayira E van den Broek NR Smyth RL , 2010. The Malawi Developmental Assessment Tool (MDAT): the creation, validation, and reliability of a tool to assess child development in rural African settings. PLoS Med 7: e1000273. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Gomez LA et al. 2018. Improved neurodevelopment after initiation of antiretroviral therapy in human immunodeficiency virus-infected children. Pediatr Infect Dis J 37: 916–922. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Uwemedimo OT Howlader A Pierret G , 2017. Parenting practices and associations with development delays among young children in Dominican Republic. Ann Glob Health 83: 568–576. [DOI] [PubMed] [Google Scholar]
- 21. Gladstone MJ et al. 2019. Independent and combined effects of improved water, sanitation, and hygiene (WASH) and improved complementary feeding on early neurodevelopment among children born to HIV-negative mothers in rural Zimbabwe: substudy of a cluster-randomized trial. PLoS Med 16: e1002766. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Vyas S Kumaranayake L , 2006. Constructing socio-economic status indices: how to use principal components analysis. Health Policy Plan 21: 459–468. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.