Agricultural and nutritional education interventions for reducing aflatoxin exposure to improve infant and child growth in low- and middle-income countries.

backgroundAflatoxins are carcinogenic mycotoxins that contaminate many food crops. Maize and groundnuts are prone to aflatoxin contamination, and are the major sources of human exposure to aflatoxins, due to their high intake as staple foods, particularly in low- and middle-income countries (LMICs). Observational studies suggest an association between dietary exposure to aflatoxins during pregnancy and early childhood and linear growth in infants and young children.

objectivesTo assess the effects on pre- and postnatal growth outcomes when agricultural and nutritional education interventions during the post-harvest period that aim to reduce aflatoxin exposure are compared to usual support or no intervention. We assessed this in infants, children, and pregnant and lactating women at the household or community level in LMICs. SEARCH

methodsIn July and August 2019, we searched: CENTRAL, MEDLINE, Embase, CINAHL, Web of Science Core Collection, Africa-Wide, LILACS, CAB Abstracts, Agricola, and two trials registers. We also checked the bibliographies of the included studies and contacted relevant mycotoxin organisations and researchers for additional studies. SELECTION CRITERIA: We included randomised controlled trials (RCTs) and cluster-RCTs of agricultural education and nutritional education interventions of any duration, at the household or community level, aimed at reducing aflatoxin intake by infants, children, and pregnant and lactating women, in LMICs during the post-harvest period, compared to no intervention or usual support. We excluded studies that followed participants for less than four weeks. We assessed prespecified prenatal (at birth) and postnatal growth outcomes (during infancy, childhood, and adolescence), with linear growth (as the primary outcome), infectious disease morbidity, and unintended consequences. DATA COLLECTION AND ANALYSIS: Two authors independently assessed study eligibility using prespecified criteria, extracted data, and assessed risk of bias of included RCTs. We evaluated the certainty of the evidence using GRADE, and presented the main results in a 'Summary of findings' table. MAIN

resultsWe included three recent cluster-RCTs reporting the effects of agricultural education plus post-harvest technologies, compared to usual agricultural support or no intervention. The participants were pregnant women and their children, lactating women and their infants (< 6 months), women of childbearing age, and young children (< 59 months), from rural, subsistence maize-farming communities in Kenya, Zimbabwe, and Tanzania. Two trials randomised villages to the intervention and control groups, including a total of at least 979 mother-child pairs from 60 villages. The third trial randomised 420 households, including 189 mother-child pairs and 231 women of childbearing age. Duration of the intervention and follow-up ranged between five and nine months. Due to risk of attrition bias, the overall risk of bias was unclear in one trial, and high in the other two trials. None of the included studies addressed the effects of nutritional education on pre- and postnatal growth. One trial reported outcomes not prespecified in our review, and we were unable to obtain unpublished growth data from the second trial, even after contacting the authors. The third trial, in lactating women and their infants in Tanzania, reported on the infants' weight-for-age z-score (WAZ) after six months. This trial found that providing agricultural education aimed at changing farmers' post-harvest practices to reduce aflatoxin exposure, by using demonstrations (e.g. handsorting, de-hulling of maize, drying sheets, and insecticides), may improve WAZ in infants from these farmers' households, on average, by 0.57 (95% confidence interval (CI) 0.16 to 0.98; 1 study; 249 participants; very low-certainty evidence), compared to infants from households where the farmers received routine agricultural extension services. Another way of reporting the effect on WAZ is to compare the proportion of underweight infants (WAZ > 2 SD below the reference median value) per group. This trial found that the intervention may reduce the proportion of underweight infants in the intervention households by 6.7% (95% CI -12.6 to -1.4; 249 participants; very low-certainty evidence) compared to control households. No studies reported on unintended effects of agricultural and nutritional education. AUTHORS'

conclusionsEvidence on the effects on child growth in LMICs of agricultural or nutritional education interventions that reduce aflatoxin exposure was very limited; no included study reported on linear growth. Very low-certainty evidence suggested that agricultural education aimed at changing farmers' post-harvest practices to reduce aflatoxin exposure by using demonstrations, may result in an increase in WAZ, when compared to usual or no education.