|dc.description.abstract||Background: The former Transkei region of the Eastern Cape (EC) is a deep rural area characterised by a high prevalence of poverty and underdevelopment. Subsistence farming is a major source of food and maize consumption is part of a culturally distinct dietary pattern and ethnic tradition. It has furthermore been well-documented that the home-grown maize in these rural areas are extremely high in mycotoxins. Mycotoxins are low-molecular-weight metabolites that are produced by fungi that grow on the maize. Aflatoxins (AF), deoxynivalenol (DON), zearalenone (ZEA) and fumonisin (FB) are some of the major mycotoxins that influence human health. The mycotoxins are associated amongst other thing with hepatitis, liver cancer, stunting and immune suppression, gastro-intestinal disorders, anorexia, nausea, emesis, headache, chills, giddiness and convulsions, precocious pubertal changes in children, early menarche and possibly infertility and an increased risk of oesophageal and liver cancer, neural tube defects and stunting.
Very little is known about the association of maternal exposure and anthropometric measures of infants at birth. Although there is some evidence that AF affects human foetal growth, none or very little, human research has been done on the association of other mycotoxins such as DON, ZEA and FB. Therefore, the aim of this study is to determine the association between maternal multi-mycotoxin exposure and anthropometric outcomes at birth of mothers and their infants in rural areas of the EC, South Africa.
Methods: This sub-study was part of a larger prospective study (PhilaSana). The PhilaSana study was a longitudinal study focussing on the factors affecting infant and young child feeding and their growth patterns during the first 1 000 days. The study used systematic and snowball sampling to recruit pregnant women at various villages within the pre-selected area. Women were included if they were pregnant and lived in the area. Ethical approval was obtained from the Health Research Ethics Committee (HREC) at North-West University. Informed consent was obtained in the participants’ first language, isiXhosa. Data were collected from mothers and children and included socio-demographic information, maternal and child general health (self-reported), maternal and infant dietary intake, anthropometric measures and maize samples. To determine mycotoxin concentration on maize, 1 kg home-grown maize samples were collected per household. The maize was analysed by LC-MS/MS for DON, ZEA and FB concentrations. Mycotoxin (DON, ZEA and FB) exposures were measured, expressed as probable daily intake (PDI) in μgkg-1 body weight (bw) day-1 (the deterministic approach). Based on the mycotoxin concentrations found on the maize (determined from another sub-study), and the mean raw maize intake, mycotoxin exposure was calculated. Furthermore, infant birth anthropometric measurements including weight, length, head circumference (HC) and gestational age (GA) were recorded based on the information in the Road to Health Booklet (RtHB). Lastly, the association between maternal mycotoxin exposure and infant birth anthropometric measurements and GA were determined. Pregnant women (n=92), and infants at birth were included. Only women consuming home-grown maize were included. Amount of cooked maize consumed in a day was determined by the portion size (in grams) multiplied by the number of portions consumed in a day. Monthly intake was determined by consumption frequency per week or per month multiplied by intake at a time. Monthly intake was divided by 28 days to give a mean daily intake of cooked maize meal. Mean daily intake of cooked maize was then converted to raw maize according to recipes obtained during the development of the questionnaire and portion size photographs. Mycotoxin exposures of pregnant women were determined according to maize contamination levels. The mean total FB (FB1 + FB2 +FB3), DON and ZEA levels were obtained from the analysis of home-grown maize collected from the same households. Mean contamination levels were found to be 24.5 μgkg-1 for DON, 31.0 μgkg-1 for ZEA and 1 035.0 μgkg-1 for FB.
Data were captured and cleaned in Excel. Statistical analyses were conducted with SPSS version 25. Data were tested for normality using the Shapiro-Wilk Test and was not normally distributed. Median, and IQR were reported for the mycotoxin exposures (DON, ZEA and FB), as well as for infant weight, length, HC and GA. Maternal age, weight and total raw maize intake was also reported as median and inter quartile range (IQR). Birth weight, length, HC and GA were divided into tertiles (tertile 1 = lower third, tertile 2 = the middle third and tertile 3 = the upper third). The Krusskal-Wallis test were used to compare the tertiles against mycotoxin exposure levels. Generalized linear regressions were performed to determine the impact of maternal age, weight and maize intake (as confounders). Significant levels were set at p ≤ 0.05. Data from alcohol consumption, tobacco use, HIV and TB were not included due to large numbers of missing values.
Results and discussion: Maternal exposure levels for DON and ZEA were lower than that of FB. Although the median of the DON exposure was under the probable maximum tolerance daily intake (PMTDI), the IQR indicated that there were participants with levels above the expected safety levels. The median and IQR for ZEA was below PMTDI levels, as opposed to the median for FB, that was much higher than the estimated PMTDI. The IQR indicated exposure levels as high as 52 μgkg-1 bw day-1 compared to the PMTDI of < 2 μgkg-1 bw day-1. Results indicated that DON exposure might be associated with infant birth weight, with higher DON exposure in the upper tertile, while FB exposures were associated with lower birth weight of the infants. For both DON and FB higher exposures were associated with a smaller HC. There were indications for all three mycotoxins that higher exposures resulted in infants with a lower GA, although no significant associations were found. However, generalized linear regressions indicated that although DON and FB were both associated with infant weight, maternal weight might have influenced the results. The same was found for the association of DON and FB with HC, which might have been influenced by maternal age. After adjustment for confounders no association between ZEA and any of the anthropometric measures were found. Thus, although data indicate that DON exposure was associated with greater infant birth weight and FB with smaller infant birth weight and that both DON and FB exposures were associated with smaller HCs, associations could have been influenced by confounding factors such as maternal weight and age. More research is required to better understand the associations between DON, ZEA and FB and infant anthropometric measures at birth. Furthermore, more research is needed to determine additional factors that might influence maternal health and the anthropometric measures of infants at birth.||en_US