Aflatoxin Exposure Among Mothers and Their Infants from the Western Highlands of Guatemala

Abstract

Objectives

We examined breast milk of mothers and urine of infants before and after introduction of supplementary foods for aflatoxin M₁ (AFM₁) and the association between AFM₁ with maternal and infant diet.

Methods

A prospective cohort study was conducted among mothers and infants ages 0–6 months and 7–12 months from June-October 2014. Sociodemographic, dietary, birth, and health data were collected. A breast milk sample was collected from each mother and a urine sample from each infant at baseline (time point 1) and monthly for 2 time points thereafter; samples collected at baseline and time point 3 were tested for AFM₁.

Results

Almost 5% of breast milk and 15.7% of urine samples tested AFM₁-positive. The median AFM₁ in breast milk was 0.020 ng/mL and in urine 0.077 ng/mg creatinine. At time point 3, infants of 5 of the 6 mothers in each group who were AFM₁-positive in breast milk were also AFM₁-positive in urine. Mothers’ consumption of cooked maize/maize dough ≥ 3 days per week (OR 2.96, 95% CI = 1.19–7.34) and mothers’ consumption of tamales made from maize ≥ 3 days per week (OR 0.28, 95% CI = 0.10–0.73) were significantly associated with AFM₁ in infant urine.

Conclusion

This is the first study in Guatemala documenting aflatoxin exposure in both breast milk of lactating mothers and infantś urine during the first year of life. This may have important implications in understanding the multicausality of the high rates of stunting among children < 5 years old in Guatemala.

Introduction

Forty-seven percent of children < 5 years old in Guatemala are stunted, and the country is ranked sixth in the world for chronic malnutrition (United States Agency for International Development, 2018). Possible factors that contribute to malnutrition and stunting are poor socioeconomic conditions, poor housing or sanitation, infections, non-potable water, and diarrhea (Sereebutra et al., 2006). Population groups in Guatemala with the highest incidence of malnutrition also have the highest poverty levels. The most affected children tend to live in rural areas and are of indigenous descent (Pan American Health Organization, 2018). Poverty affects food security, adequate care, and health, the three determinants of the nutritional status of children (Black et al., 2008).

Stunting is such a prevalent problem in Guatemala that it cannot be addressed only by the amount of food a child eats, but also by infant feeding, mother’s diet, and breastfeeding practices. Although the Guatemalan diet may include some meat, fish, and poultry, much of the diet is maize-based, especially among rural Guatemalans (United States Agency for International Development, 2018). According to the 2006 National Economic Survey in Guatemala, maize tortilla is the most important source of energy intake, with a national average of about 30% but may be above 40% in the Northern and Northwestern region (Instituto de Nutrición de Centroamérica y Panamá, 2013). A nutrition study involving rural maternal and child populations (predominantly of Mayan descent) in the Departments of Quiché and Huehuetenango showed that the contribution of maize to energy intake for children < 24 months old was around 35%, 58% for pregnant women, and 67% for lactating women (Food & Nutritional Technical Assistance, 2014). This finding highlights the low diet diversity and the relative high contribution of maize to the diet. This situation places these populations at high risk of nutrient deficiencies, but also at high risk of aflatoxin exposure that may contribute to negative health and nutritional outcomes (United States Agency for International Development, 2018).

The Guatemalan tropical environment and pre- and post-harvest factors contribute to fungal growth and proliferation of mycotoxins such as aflatoxins and fumonisins in maize (Mendoza et al., 2017). Aflatoxin is produced by the fungi Aspergillus flavus and Aspergillus parasiticus, and flourishes in crops such as maize, groundnuts, legumes, and other grains that form the staple diet in many developing countries (Cardwell, 2000). Of the four main aflatoxin chemotypes (B₁, B₂, G₁ and G₂), Aflatoxin B₁ (AFB₁) is the most carcinogenic form (Park et al., 2002). Aflatoxin M₁ (AFM₁) is a metabolite of AFB₁ that is found in milk and urine.

Aflatoxin exposure can adversely affect reproduction and birth outcomes in humans. A study of birth outcomes among pregnant women positive for AFB₁ in peripheral and cord blood showed that women in the highest AFB₁ quartile were twice as likely to have low birth weight infants compared to those in the lowest quartile (Shuaib et al., 2010). A previous study showed that AFB₁ levels in maternal blood were a strong predictor of height and weight gain of Gambian infants from birth to one year, with lower gain in infants with higher AFB₁ (Turner et al., 2007). Aflatoxin exposure in children has been associated with stunting and several pathways have been proposed, including early exposure in utero, during breastfeeding and while on supplemental food during weaning. Studies conducted in populations with high aflatoxin exposure showed that the levels of aflatoxin in children rose significantly up to three years old, were higher in infants who had stopped breastfeeding, and were correlated with maize consumption (Gong et al., 2003). Furthermore, a dose–response relationship between aflatoxin contamination levels and the degree of stunting was demonstrated (Gong et al., 2003, 2004).

Previous research reported that aflatoxin levels in maize samples from local markets in some departments of Guatemala were above import limits set by the United States and European Union (Torres et al., 2015). In addition, a study that calculated aflatoxin exposure level based on maize consumption reported a negative correlation between aflatoxin exposure and height for age z score in children 3 months to 5 years in rural communities in the Western Highlands of Guatemala (Voth-Gaeddert et al., 2018). Maize is one of the first complementary foods that the infant receives after 6 months. This raises concerns about the risk of aflatoxin exposure at an early age and the implications this may have for health and nutrition during growth, development, and later in life. This study examined aflatoxin exposure in mothers and their infants at two critical periods, before and after the introduction of foods other than breast milk to the infant, and the association of AFM₁ levels in infant urine with sociodemographic factors, maternal and infant diet, gestational age, and infant health and growth.

Methods

Study Design and Participants

A prospective cohort study was conducted among mothers and their infants attending clinics in three municipalities (Nebaj, Chajul, and Cotzal) of Quiché in the Western Highlands of Guatemala. These are rural/agricultural communities with high nutritional vulnerability and a prevalence of chronic malnutrition of about 73%. However, these populations have overall high rates of breastfeeding (> 90%) and exclusive breastfeeding rates of about 50%. The study sample consisted of two groups of mothers and their infants; one group consisted of 60 mothers and their infants 0–6 months old that were exclusively breastfed and the other of 84 mothers and infants ages 7–12 months with a breast milk and supplementary diet. This study did not include an intervention, but followed the traditional practices of introduction of complementary foods.

Eligibility Criteria

To be eligible to participate a mother had to live in one of the three municipalities of Quiché and be willing to provide her breast milk and urine samples from her infant. Infants in the 0–6 month-old age group had to be exclusively breastfed and infants in the 7–12 month-old age group had to have a combined diet of breast milk and supplementary foods. Potential participants were defined as mothers ≥ 16 years old who were not currently pregnant and had an infant 0–12 months old. Women and their infants who did not meet these inclusion criteria and infants with acute illnesses such as diarrhea, lower or upper respiratory infection, or who were on antibiotic treatment for any infection were excluded from the study.

Recruitment and Data Collection

Potential participants were told of the study by the clinic staff and community coordinators and asked if they would be willing to participate. Upon indication of willingness to participate, the clinic staff arranged for the mothers to visit the clinic to be introduced to the study staff. After the study staff explained the study and answered questions from the mothers, each mother provided signed informed consent. An interviewer-administered questionnaire was used to collect information from the mothers on sociodemographic factors (age, marital status, educational level, employment, religion, and living situation), number of pregnancies, number of children, and infant health. A breast milk sample (self-pumped using a manual disposable pump) was collected from each mother and a urine sample was collected from each infant using a pediatric urine collection bag. At the baseline visit, the mothers were scheduled to make two additional visits at monthly intervals to the clinic and at each visit, a breast milk sample and an infant urine sample were collected, and the length and weight of the infants measured. WHO Growth Charts were used to characterize nutritional status (World Health Organization, 2009). Diet was assessed by asking the mother about the foods consumed by the infant the day before the interview and by a food frequency questionnaire focused on the mother’s intake of foods made from maize such as tortillas and tamales during the week before the baseline interview. We also tried to distinguish between cooked maize (e.g. boiled on the cob or some other way) and maize dough.

Breast milk and urine samples collected at baseline and the third time point were tested for AFM₁ levels. We did not always have samples from the third time point because of loss to follow-up; therefore, 28 samples from the second time point were included with samples from the third time point.

Quantification of Aflatoxin M₁ Metabolite in Urine and Breast Milk

AFM₁ in infant urine was measured using the method of Warth et al., (2012) and outlined in the dissertation by Elmore (2016). AFM₁ in breast milk was quantified using the method of the Association of Official Analytical Chemists (Method 2000.08; AOAC, 2000) and outlined in Maki et al., (2016). In all, 573 samples (120 breast milk and 120 urine for group 1 and 166 breast milk and 167 urine for group 2) were tested.

Statistical Analysis

One hundred and forty-four mothers and their infants for whom AFM₁ measurements were conducted were included in the analysis. Sixty infants were 0–6 months old (group 1) and 84 were 7–12 months old (group 2). The participants were categorized into AFM₁-positive and AFM₁-negative groups. Bivariate analyses were conducted to determine statistical associations between sociodemographic, diet, and infant variables and AFM₁ positivity. Chi-square tests were used for categorical variables and t-tests for continuous variables. The odds of AFM₁ positivity were estimated for each predictor variable significantly associated in bivariate analyses (p-value < 0.1) using unadjusted logistic regression analyses. Finally, variables which were significantly associated in bivariate analyses and which had group sizes larger than five were included in a multivariable model to determine the extent to which each predicted the odds of AFM₁ positivity alone.

Results

AFM₁ Levels in Breast Milk of Mothers and Urine of the Infants

Overall, 14/286 (4.9%) breast milk and 45/287 (15.7%) urine samples were AFM₁-positive. The median AFM₁ in breast milk was 0.020 (range = 0.004–0.333) ng/mL, and the median AFM₁ in urine was 0.077 (range = 0.013–0.978) ng/mg creatinine. Table 1 shows the AFM₁ levels in breast milk and urine at baseline and time point 3 for groups 1 and 2. The percentages of AFM₁-positive breast milk samples were similar for groups 1 and 2 mothers (1.7 and 1.2%, respectively) at baseline and at time point 3 (10 and 7.2%, respectively). Likewise, the percentages of positive urine samples were similar for groups 1 and 2 infants at baseline (10.0 and 9.5%, respectively), and at time point 3 (21.7% for both).

Table 1.

Range, median and mean AFM₁ levels in breast milk of mothers and urine from infants in groups 1 and 2 at baseline and time point 3^*

	AFM1 in breast milk (ng/ml)	AFM1 in breast milk (ng/ml)	AFM1 in infants’ urine (ng/mg creatinine)	AFM1 in infants’ urine (ng/mg creatinine)
Group 1 (0–6 months)	Baseline n = 60	Time point 3 n = 60	Baseline n = 60	Time point 3 n = 60
Number and percent positive	1/60 (1.7%)	6/60 (10.0%)	6/60 (10.0%)	13/60 (21.7%)
Range	0.013	0.004–0.333	0.039–0.281	0.016–0.51
Median	0.013	0.023	0.074	0.081
Mean	0.013	0.077	0.217	0.134
Group 2 (7–12 months)	n = 83	n = 83	n = 84	n = 83
Number and percent positive	1/83 (1.2%)	6/83 (7.2%)	8/84 (9.5%)	18/83 (21.7%)
Range	0.027	0.008–0.227	0.013–0.178	0.018–0.964
Median	0.027	0.015	0.058	0.209
Mean	0.027	0.151	0.065	0.016

^*9.8% (28/287) of samples were from time point 2

Matching of Mother-Infant AFM₁

For both groups of mothers, only one breast milk sample was AFM₁-positive at baseline. At time point 3, infants of 5 of the 6 mothers in both groups who had AFM₁-positive breast milk were also AFM₁-positive in their urine (83.3% for both). However, the mothers of eight group 1 infants whose urine were AFM₁-positive had AFM₁-negative breast milk and mothers of 13 group 2 infants who had AFM₁-positive urine had AFM₁-negative breast milk. Although the number of AFM₁-positive breast milk and urine samples at time point 3 was insufficient for correlation analyses, it is important to note that these cases of aflatoxin exposure in infants are of substantial concern.

Sociodemographic Characteristics, Diet and Infant Factors by AFM₁ in Infant Urine

Significant differences were observed between mothers’ age (p = 0.043), religion (p = 0.030), number of pregnancies (p = 0.027), number of children born alive (p = 0.027), and whether mothers reported purifying water at home (p = 0.001; Table 2).

Table 2.

Sociodemographic characteristics of the study participants by aflatoxin M₁ detection in urine of infants

	Totala N = 144 (%)	Infant positive N = 40 (%)	Infant negative N = 104 (%)	P-valueb
Mother’s age (years)
≤ 25	53 (46.1)	6 (24.0)	47 (52.2)	0.043*
26–35	45 (39.1)	14 (56.0)	31 (34.4)
≥ 36	17 (14.8)	5 (20.0)	23 (13.3)
Marital status
Single/separated/widow	7 (6.1)	2 (8.0)	5 (5.6)	0.887
Living with partner	44 (38.6)	9 (36.0)	35 (39.3)
Married	63 (55.3)	14 (56.0)	49 (55.1)
Education level
Not-educated	44 (38.3)	9 (36.0)	35 (38.9)	0.943
Primary	46 (38.3)	10 (40.0)	36 (40.0)
Secondary or above	25 (21.7)	6 (24.0)	19 (21.1)
Religion
Evangelist	93 (80.9)	24 (96.0)	69 (76.7)	0.030*
Other	22 (19.1)	1 (4.0)	21 (23.3)
Employed	16 (11.1)	2 (5.0)	14 (13.5)	0.148
Depends economically on family	98 (68.1)	24 (60.0)	74 (71.2)	0.199
Number of children
1–2	25 (17.4)	3 (7.5)	22 (21.2)	0.053*
≥ 3	119 (82.6)	37 (92.5)	82 (78.9)
Number of pregnancies
1–2	53 (36.8)	9 (22.5)	44 (42.3)	0.027*
≥ 3	91 (63.2)	31 (77.5)	60 (45.7)
Number of children born alive
1–2	57 (39.6)	10 (25.0)	47 (45.2)	0.027*
≥ 3	87 (60.4)	30(75.0)	57 (54.8)
Person living in household
1–5	41 (35.7)	5 (20.0)	36 (40.0)	0.167
6–8	36 (31.3)	9 (36.0)	27 (30.0)
≥ 9	38 (33.0)	11 (44.0)	27 (30.0)
Source of water
Pipe in home or own well	99 (86.1)	24 (96.0)	75 (83.3)	0.106
Othersc	16 (13.9)	1 (4.0)	15 (16.7)
Purify the water at home	115 (79.9)	25 (62.5)	90 (86.5)	0.001*

^aNumbers may not always sum to total due to missing observations

^bP-value estimated from Chi-Square test, threshold of 0.1

^cPublic well, public faucet, water truck, and river or canal

^*P-value significant at < 0.1

Examination of mothers’ diet by AFM₁ in infant urine, showed a significant difference for mothers who ate cooked maize or maize dough ≥ 3 days per week (p = 0.022; Table 3).

Table 3.

Diet of mothers and infants by aflatoxin M₁ in urine

	AFM1 Positive N = 40 N (%)	Negative N = 104 N (%)	P-value$
Mother diet (days per week)€
Tortillas made from maize ≥ 3 days	34 (100.0)	94 (95.9)	0.232
Tamales made from maize ≥ 3 days	9 (26.5)	43 (43.9)	0.074*
Cooked maize or maize dough ≥ 3 days	26 (76.5)	53 (54.1)	0.022*
Coffee-like beverage made from burnt tortilla or toasted maize ≥ 1 day	5 (15.2)	11 (11.2)	0.551
Infant diet#
Breastfeeding	34 (100.0)	97 (99.0)	0.554
Number of times breastfed/day (mean ± SD)	12.21 ± 3.45	12.72 ± 4.14	0.516
Supplemented breastfeeding, other liquids or solid food	17 (50.0)	59 (60.2)	0.300
Age (months) when first began drinking other liquids (mean ± SD)+	6.18 ± 0.53	6.4 ± 0.65	0.183
Age (months) when first began eating solids (mean ± SD)	6.53 ± 0.72	6.67 ± 0.80	0.524
Taking vitamins	7 (41.2)	28 (48.3)	0.606
Number of times consumed solid food (mean ± SD)	3.12 ± 0.93	3.20 ± 1.39	0.819
Number of times consumed tortilla or Tamalito (mean ± SD)	2.18 ± 0.95	2.06 ± 0.93	0.679
Consumption of maize > 165 g per day	18 (45.0)	34 (32.7)	0.168

^$Computed from chi-square tests for categorical variables and t-tests for continuous variables

^€Diet within previous week of interview

^#Within day before interview?

⁺Including infant formula or liquid or powdered cow’s milk

^*P-value significant at < 0.1

Examination of infants’ age, gestation and mothers’ perception of infants’ health and growth by AFM₁ in urine showed significant differences for age (p = 0.010) and length of pregnancy (p = 0.022; Table 4).

Table 4.

Infants’ age, gestation and health and growth by aflatoxin M₁ in urine of infants

	Infant positive N = 40 (%)	Infant negative N = 104 (%)	P-valuea
Infant age (months)
≤ 6	28 (70.0)	48 (46.2)	0.010*
≥ 7	12 (30.9)	56 (53.9)
Length of the pregnancy (months)
≤ 8	16 (40.0)	22 (21.2)	0.022*
= 9	24 (60.0)	82 (78.9)
Premature	2 (8.0)	5 (5.6)	0.651
Small for gestational age	3 (12.0)	5 (5.6)	0.263
Mother perceives child is generally healthy	5 (20.0)	18 (20.2)	0.980
Mother perceives child’s growth is good/normal	23 (92.0)	83 (93.3)	0.828
Received medicine to treat infectionb	5 (13.9)	15 (7.3)	0.184

^aP-value estimated from Chi-Square test, threshold of 0.1

^bWithin the previous 2 weeks of the interview

^*P-value significant at < 0.1

Stunting, Malnourishment, and Wasting of Infants in Relation to AFM₁ in Urine

Overall, 16.7% of group 1 infants had ≤ 2 weight for height z-scores that indicated wasting, 41.7% were stunted (height for age z-score ≤ 2) and 21.7% were mal-nourished (weight for age z-score ≤ 2). Among group 2 infants, 25.0% showed wasting, 63.1% were stunted, and 35.7% were malnourished. However, the number of AFM₁-positive infants in each group was insufficient to conduct analysis for comparisons of these growth and nutritional variables.

Multivariable Logistic Regression Model (Table 5)

Table 5.

Crude and adjusted odds ratios (ORs)^a and 95% confidence intervals (CIs) for the association between AFM₁-positive infants and sociodemographic, reproductive, and dietary factors

	Crude odds ratio (95% CI)	Adjusted odds ratiob (95% CI)
Mother’s age (years)
16–25	Referent	–
26–35	3.54 (1.23–10.19)	–
≥ 36	3.26 (0.85–12.53)	–
Religion
Evangelist	7.30 (0.93–57.25)	–
Other	Referent	–
Number of children
≥ 3	3.31 (0.93–11.75)	–
1–2	Referent	–
Number of pregnancies
≥ 3	2.53 (1.09–5.84)	0.95 (0.08–11.25)
1–2	Referent	Referent
Number of children born alive
≥ 3	2.47 (1.10–5.58)	1.93 (0.17–22.13)
1–2	Referent	Referent
Purify water at home
Yes	0.26 (0.11–0.61)	0.22 (0.02–2.42)
No	Referent	Referent
Mother consumed tamales made from maize*
≥ 3 days/week	0.46 (0.20–1.09)	0.28 (0.10–0.73)
< 3 days/week	Referent	Referent
Mother consumed cooked maize or dough*
≥ 3 days/week	2.76 (1.14–6.70)	2.96 (1.19–7.34)
< 3 days/week	Referent	Referent
Infant age
≥ 7 months	0.37 (0.17–0.80)	0.60 (0.23–1.56)
< 7 months	Referent	Referent
Length of pregnancy
≤ 8 months	2.49 (1.13–5.47)	0.55 (0.06–4.99)
> 8 months	Referent	Referent

^aEstimated using logistic regression for those variables with a p-value of < 0.1 in bivariate analyses

^bAdjusted for variables with a p-value of < 0.1 in bivariate analyses and with groups larger than N = 5, including number of pregnancies, number children born alive, purify water at home, mother consumption of tamales made from maize and cooked maize/maize dough, infant age, and length of pregnancy

^*Within previous 1 week of interview

The crude logistic model showed increased odds for AFM₁ among infants born to mothers 26–35 years old (OR 3.54 95% CI I1.23–10.19), with ≥ 3 pregnancies (OR 2.53, 95% CI = 1.09–5.84), and with ≥ 3 live births (OR 2.47, 95% CI = 1.10–5.58). Additionally, there was increased odds for AFM₁ among infants born ≤ 8 months of gestation (OR 2.49, 95%CI = 1.13–5.47), and infants whose mothers consumed cooked maize or maize dough ≥ 3 days per week (OR 2.76, 95% CI = 1.14–6.70; Table 5). Infants who were ≥ 7 months of age had decreased odds of being AFM₁-positive (OR 0.37, 95% CI = 0.17 – 0.80). In the adjusted multivariable model, two variables were statistically significant. Infants whose mothers consumed cooked maize or maize dough ≥ 3 days the week before the baseline survey had a threefold increased odds of AFM₁ positivity compared to those whose mothers consumed cooked maize or maize dough < 3 days (OR 2.96, 95% CI = 1.19–7.34). However, infants of mothers who consumed tamales ≥ 3 days the week preceding the baseline survey were estimated to be 72% less likely to be AFM₁-positive compared to those whose mothers’ consumed tamales < 3 days (OR 0.28, 95% CI = 0.10–0.73).

Discussion

Linear aflatoxin biomarkers like AFM₁ are predictive of AFB₁ intake and are used as the standards for assessment of population exposures. AFM₁ in milk and urine has been shown to be a reliable short-term bioassay of AFB₁ exposure, due to its linearity with dose and ease of detection (Groopman & Kensler, 1993). The liver is the primary site of metabolism of aflatoxins. AFB₁ is bioactivated by hepatic cytochrome P450 enzymes (CYP450) to a highly reactive electrophilic metabolite known as aflatoxin-8,9-exo-epoxide. This unstable metabolite reacts with cellular macromolecules such as proteins causing cytotoxicity and with DNA causing genotoxicity (Doi et al., 2002). Chronic exposure to AFB₁ and AFM1₁ contribute to immune suppression, impaired growth, and nutritional interference (Turner et al., 2003; Jiang et al., 2005; Gong et al., 2004; Shane, 1994).

In the current study, approximately 5% of breast milk and 16% of urine samples were AFM₁-positive. Studies from Brazil and Iran reported similar percentages of AFM₁-positive breast milk samples (5.3 and 6.0%, respectively; Ishikawa et al., 2016; Ghiasian & Maghsood, 2012). Other studies from Nigeria and Mexico reported higher percentages of AFM₁-positive breast milk (Adejumo et al., 2013; Cantú-Cornelio et al., 2016). A study from Nigeria that examined breast milk from 50 mothers and urine of their infants 0–6 months old reported that 96% breast milk and 88% of infant samples were AFM₁-positive (Gide et al., 2019). The differences in percentages of AFM₁-positive samples in these studies most likely reflect aflatoxin contamination levels of staple foods in areas of these countries.

We observed higher percentages of AFM₁-positive breast milk and urine samples for both groups 1 and 2 at time point 3 than at baseline. The difference is most likely due to contamination levels of the maize consumed during the different months due to seasonal factors and source of maize. Guatemala has a dry season that lasts from November to the beginning of May and a rainy season from May to November with the heaviest rain during September and October. Our study was conducted from June to October with baseline recruitment in the earlier months and the follow-up time-points in the latter months. At baseline, participants were most likely eating maize that they grew and harvested in the previous growing season and with less aflatoxin contamination; however, this needs further investigation. Time point 3 corresponds mostly to September and October which is before the maize harvest season in the study region, but just about the time of crop season of maize coming from the warm and humid lowlands of the South Coast and North. Although a significant proportion of families in our study region grow and consume their own maize, not all families produce enough maize to support their home consumption all year around, so they rely on the purchase of maize coming from the maize producing regions of the low lands of the South Coast and the North. Consumption of maize kept under poor post-harvest practices and with increased aflatoxin contamination in these hot and humid zones probably explains the increased percentages of AFM₁-positive breast milk and urine samples at time point 3 (Orozco, 2014). Previously, we observed higher aflatoxin levels when participants were eating maize harvested after the previous growing season that was insufficiently dried and stored under hot, humid conditions (Jolly et al., 2015).

In this study, the percent of AFM₁-positive infant urine samples is higher than AFM₁-positive breast milk samples (15.7 vs 4.9%, respectively). This is understandable for group 2 infants who were already on a mixed diet. Earlier studies showed higher aflatoxin levels in infants who had stopped breastfeeding and were correlated with maize consumption (Gong et al., 2003, 2004). However, for group 1 infants who were exclusively breastfed, the urine of five infants whose mothers’ breast milk was not positive, was AFM₁-positive at baseline and 7 infants whose mothers’ breast milk was not positive were AFM₁-positive at time point 3. It is possible that these infants, although reportedly exclusively breastfed, were given other foods contaminated with aflatoxin. Alternately, a portion of AFM₁ may have bound to casein in breast milk and result in under detection of AFM₁ (Brackett & Marth, 1982). However, when AFM₁-positive breast milk and urine samples for time point 3 were matched, infants of five of the six mothers in both groups 1 and 2 who were AFM₁-positive in breast milk were also AFM₁-positive in their urine (83.3%). This suggests a link between mother and infant aflatoxin exposure and indicates breast milk exposure among the infants, especially when exclusively breastfed.

The urine from infants of mothers who consumed cooked maize or maize dough ≥ 3 days the week before the baseline interview was 3 times more likely to be AFM₁-positive than those whose mothers consumed cooked maize or maize dough < 3 days. This finding is most likely a result of the high daily consumption of tortilla (12–30) made from aflatoxin contaminated maize dough by the mothers and transfer to the infants through breast milk for exclusively breastfed infants, and through breast milk and complementary foods for infants on mixed diets. A recent study conducted among adult Guatemalans from five geographic departments showed a statistically significant association between tortilla consumption and AFB₁-albumin adduct levels, with higher AFB₁ adducts in those in the highest quintile of tortilla consumption (Kroker-Lobos et al., 2019).

Infants with mothers who consumed tamales ≥ 3 days per week were 72% less likely to be AFM₁ positive compared to those whose mothers consumed tamales < 3 days per week. We cannot explain this finding with certitude. However, in Guatemala, different types of tamales are prepared from maize dough that is frequently combined with any of a variety of other ingredients such as seasonal green leafy vegetables, tomatoes, chopped garlic, onions, chilies, bell peppers olives, chicken or pork (if available) and other spices. As such, depending on the ingredients and size, a single tamale may be enough for a meal and would result in less maize intake compared to a meal with tortillas.

Interventions to control or mitigate aflatoxin contamination in maize in Guatemala should reduce the overall exposure to aflatoxin in this population, help to ameliorate mother and infant nutritional status, and improve infant growth and development. Due to the latent onset of aflatoxin-related health consequences, serum and urinary bio-markers, such as the AFB₁-albumin adduct and AFM₁ are important tools in determining the efficacy of intervention trials for the reduction of aflatoxin exposure in human populations. These biomarkers have been applied in multiple epidemiological studies and clinical intervention trials (Afriyie-Gyawu et al., 2008; Gong et al., 2003; Groopman & Kensler, 1993; Jolly et al., 2006; Turner et al., 2007). Post-harvest and educational interventions to decrease crop contamination of aflatoxin and human exposure should be conducted in this region of Guatemala using aflatoxin biomarkers as indicators of success of the interventions.

Limitations and Conclusions

Certain limitations should be considered in interpreting the results. The small sample and number of AFM₁-positive samples prevented correlation analyses between breast milk and urine and between AFM₁ and growth and nutritional indicators. Collection of the one breast milk sample at each time point is probably not enough to characterize aflatoxin exposure adequately. However, these assays are expensive and we had limited funding. The study sample represents mothers and infants from three municipalities of Quiché and so the results may not be generalizable to other regions of Guatemala. Data on the length of pregnancy were collected in months instead of more accurately in weeks. The baseline food frequency data were used in the statistical analyses instead of data collected at each time point. However, this is not a major concern considering the low diet diversity and relatively high contribution of maize in the diet. Additionally, social desirability bias is always a possibility in self-reported data.

Regardless of these limitations, the results document exposure of both mothers and infants to aflatoxin and show agreement between AFM₁ in infant urine and breast milk of mothers. Statistically significant associations were obtained for AFM₁ in infant urine with mothers’ consumption of cooked maize or maize dough ≥ 3 days per week. Further studies should be conducted to examine the association between aflatoxin levels in infants and nutritional growth indicators and on ways to mitigate dietary aflatoxin exposure in this population.

Significance.

There is a high rate of stunting and malnutrition in Guatemala that is partly due to the low diet diversity, high maize consumption, and high risk of aflatoxin exposure. Although AFM₁ has been demonstrated in breast milk and urine in different populations, this study is the first to document aflatoxin exposure in mother-infant pairs in Guatemala. The results show agreement between AFM₁ in infant urine and breast milk of mothers and significant associations between AFM₁ in infant urine with mothers’ consumption of cooked maize/maize dough ≥ 3 days per week.