Environmental Research 184 (2020) 109266

Contents lists available at ScienceDirect

Environmental Research
journal homepage: www.elsevier.com/locate/envres

Infants and mothers levels of mercury in breast milk, urine and hair, data T
from an artisanal and small-scale gold mining area in Kadoma / Zimbabwe
Stephan Bose-O’Reillya,b,∗, Beate Lettmeierb,d, Dennis Shokoc, Gabriele Roiderd, Gustav Draschc,
Uwe Siebertb,e,f
a
Institute and Outpatient Clinic for Occupational, Social and Environmental Medicine, University Hospital Munich, LMU Munich, Ziemssenstr. 1, D-80336, Munich,
Germany
b
Institute of Public Health, Medical Decision Making and Health Technology Assessment, Department of Public Health, Health Services Research and Health Technology
Assessment, UMIT (Private University for Health Sciences, Medical Informatics and Technology), Eduard Wallnoefer Zentrum 1, A-6060, Hall i.T., Austria
c
Tailjet Consultancy Services, 4 Tor Road, Vainona, Borrowdale, Harare, Zimbabwe
d
Institute of Forensic Medicine, Ludwig-Maximilians-University (LMU), Nussbaumstraße 26, D-80336, Munich, Germany
e
Cardiovascular Research Program, Institute for Technology Assessment and Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston,
MA, USA
f
Center for Health Decision Science, Department of Health Policy and Management, Harvard School of Public Health, Boston, MA, USA

A R T I C LE I N FO A B S T R A C T

Keywords: In Kadoma (Zimbabwe) gold is extracted from ore based on a simple technology using mercury. Women are
Mercury vapor exposed whilst living in an exposed area, or even more working directly with mercury. Breast fed infants are
Breast milk exposed via mercury contaminated milk and the general environmental mercury exposure. The aim of the study
Infant was to measure infants and mothers levels of mercury in urine and hair, as well as in breast milk. In 2006, an
Artisanal and small-scale gold mining
environmental epidemiological field study with 120 mother – child pairs was conducted. A non exposed control
Zimbabwe
group (n = 42) was compared with a medium exposed group (n = 51) and a high exposed group (occupational
exposure, n = 27). Urine and hair samples from mother and infants plus breast milk samples were analyzed for
total mercury. 120 breast milk samples were analyzed, median (maximum) concentrations [μg Hg/l] were (i)
control group < 0.50 (1.55), (ii) medium exposed group 1.10 (10.48), (iii) high exposed group 1.20 (24.80)
(p < 0,001). Urine and hair results were distributed similarly for infants and mothers, with higher levels
according to the exposure subgroup. All specimen results correlated very well with another, indicating there is a
pathway between breast milk and infant's level of mercury. The daily uptake of mercury via breast milk was
calculated. The reference dose of 0.3°[μg Hg/kg BW/d] was used to compare the burden of the infants. No infant
from the control group, but 17.6% of the medium and 18.5% of the high exposed group were above the reference
dose. Mercury in breast milk is generally higher in artisanal and small scale gold mining areas. Breast fed infants
were burdened with toxic mercury via breast milk and via the general environment, both are important public
health issues.

1. Introduction
In gold mining areas all over the word, artisanal and small-scale
gold miners extract the gold from ore using mercury (United Nations
Environment Programme Chemicals, 2013). Mainly the smelting of
amalgam (gold – mercury complex), exposes the miners and the com-
munities to very high levels of mercury vapor. Environmental and
health assessments as part of the UNIDO “Global Mercury Project” were
performed in several countries. LMU conducted environmental epide-
miological studies with over 2100 exposed miners and controls in the

Abbreviations: Crea, creatinine; CV-AAS, Cold-Vapor Atomic-Absorption-Spectrometry; GerES, German Environmental Survey; GMP, Global Mercury Project; HBM,
Human Bio-Monitoring; Hg, mercury; IKKF, Institute for Clinical Cardiovascular Research; LOD, limit of detection; LMU, Ludwig-Maximilians-University; NOAEL, no
observed adverse effect level; RfD, Reference dose; UMIT, University for Health Sciences Medical Informatics and Technology; UNIDO, United Nations Industrial
Development Organization; US-EPA, U.S. Environmental Protection Agency
∗
Corresponding author. Institute and Outpatient Clinic for Occupational, Social and Environmental Medicine; University Hospital Munich, LMU Munich,
Ziemssenstr.1, D-80336, Munich, Germany.
E-mail addresses: stephan.boeseoreilly@med.uni-muenchen.de (S. Bose-O’Reilly), Beate.Lettmeier@web.de (B. Lettmeier), dennis.shoko@gmail.com (D. Shoko),
Gabriele.Roider@med.uni-muenchen.de (G. Roider), gustav.drasch@gmx.de (G. Drasch), Uwe.Siebert@umit.at (U. Siebert).

https://doi.org/10.1016/j.envres.2020.109266
Received 25 May 2019; Received in revised form 10 February 2020; Accepted 16 February 2020
Available online 20 February 2020
0013-9351/ © 2020 Elsevier Inc. All rights reserved.
S. Bose-O’Reilly, et al.
Philippines, Indonesia, Tanzania, Zimbabwe, Mongolia and Ecuador
(Bose-O'Reilly et al., 2010a; Bose-O'Reilly et al., 2010b; Bose-O'Reilly
et al., 2016; Drasch et al., 2001; Schutzmeier et al., 2016; Steckling
et al., 2011; Steckling et al., 2014). One result was that not only the
amalgam smelters showed typical signs of a chronic mercury intoxica-
tion, but also some participants in surrounding communities, including
children were considered as being intoxicated (Bose-O'Reilly et al.,
2017; Bose-O'Reilly et al., 2008a; Doering et al., 2016).
Zimbabwe is in Southern Africa and Kadoma mining district is si-
tuated in Mashonaland West province. The Kadoma mining area is
characteristically spotted with evidence of mining activities which in-
clude abandoned shafts and excavations, tailings dumps as well as
milling centers. Mining operations have been carried out in this area
since the turn of the twentieth century by white settlers and for several
centuries before by the ‘ancients’. After 1980, small-scale mining ac-
tivities began; from about year 2000 onwards, the mining activities
have increased sharply. In Kadoma mining district approximately half
of the population were directly and/or indirectly involved in the mining
activities. Data derived from the Global Mercury Project (Metcalf and
Spiegel, 2007) and interviewing members of the miners association in
Kadoma resulted in an estimation of approximately 30,000 miners and
120,000 dependents in the Kadoma mining district in 2006. Only a few
of the miners had a licensed small-scale mine while most worked as
artisanal small scale miners on other people's claims (Metcalf and
Spiegel, 2007; Mudzwiti et al., 2015; Shoko, 2014).
An environmental and health assessment of Kadoma area was con-
ducted in 2004 by the French Bureau de Recherches Géologiques et
Minières and the German Institute of Forensic Medicine of the Ludwig-
Maximilians-University of Munich (LMU), under the general coordina-
tion of the United Nations Industrial Development Organization
(UNIDO). These studies showed that in Kadoma mining district, liquid
mercury coming from the panning processes is released into the en-
vironment. During the smelting, the amalgam is heated up, the gold
remains, and the mercury evaporates. These mercury fumes were
especially hazardous to the health of the local people (Billaud et al.,
2004; Lettmeier et al., 2010). Of major concern is the developing brain
before and after birth (Grandjean and White, 2002).
Breast-feeding has to be considered as a possible source of mercury
exposure. Methylmercury is excreted into breast milk (Grandjean et al.,
1994). Less is known about the excretion of inorganic mercury but
animal studies have demonstrated that mercury from mercury vapor
exposure can also be excreted into breast milk (Vimy et al., 1997;
Yoshida et al., 1992). Organ distribution of suckling babies suggested
that they were exposed to mercury vapor via breast milk (Vimy et al.,
1997; Yoshida et al., 1992). Analysis of breast milk gives information
on mercury exposures for a breastfed child (Amin-Zaki et al., 1981;
Bansa et al., 2017; Bose-O'Reilly et al., 2008b; Grandjean et al., 1995;
Molina et al., 2018; Pajewska-Szmyt, 2019; Sharma et al., 2019).
Mercury background exposures are relatively low, and the benefits
outweigh the risks of breast feeding (Anderson and Wolff, 2000; Bose-
O'Reilly et al., 2009; Bose-O'Reilly et al., 2008a; Bose-O'Reilly et al.,
2008b; Dorea, 2008; Gundacker et al., 2002; Marques et al., 2007;
Pronczuk et al., 2002; Solomon and Weiss, 2002).
Studies in fish eating communities showed a significant adverse
effect of methyl-mercury to the regular development of infants exposed
in utero (Clarkson et al., 2003; Grandjean et al., 1994, 1997; Marques
et al., 2016a, 2016b, 2016c; Myers and Davidson, 2000; Myers et al.,
2003; Tavares et al., 2005; van Wijngaarden et al., 2006). The redis-
tribution and excretion of mercury into breast milk from mothers ex-
posed to lipophilic mercury vapor (Hg0) is not well examined (Sundberg
et al., 1999a, 1999b). The Iraqi case with severe exposure to methyl-
mercury showed a long term effect for breast-fed infants. The main
route of exposure was mercury containing breast milk leading to ab-
normal neurological signs and neurodevelopmental delays in infants
(Amin-Zaki et al., 1981). Children in small scale gold mining areas are
exposed by inhaling mercury vapor from smelting, but also endangered
2
Environmental Research 184 (2020) 109266
by ingestion of mercury from breast milk (Bose-O'Reilly et al., 2009;
Bose-O'Reilly et al., 2008b; Molina et al., 2018).
Based on this knowledge and the previous studies, an environmental
epidemiological cross sectional study was conducted in 2006 by LMU
and UMIT. The University of Zimbabwe was the national partner during
the field project.
The aim of study was to investigate to analyze the levels of mercury
in breast milk, urine and hair of mothers and infants. Secondly whether
mercury in breast milk can be considered as an appropriate biomarker
to reflect the exposure of infants due to environmental exposure with
inorganic mercury of mother and child.
2. Participants and methods
2.1. Participants
In 2006, participants from the exposed Kadoma area and from a
non-exposed control area in Chikwaka were selected at random, using
the snowball sampling approach, to participate voluntarily in the study.
Only mother-infant pairs were considered, with breast fed children
between 3 and 9 months of age. 130 mother-infant pairs were ex-
amined, 85 pairs in Kadoma and 45 pairs in Chikwaka. 10 mother-child
pairs were excluded either because: 3 children over 9 months of age; 3
children under 3 months of age; 3 due to a severe confounding acute or
chronic disease of the child and 1 due to incomplete data mainly no
specimens. In total 120 mother –child pairs were statistically further
analyzed.
The population was divided into three subgroups: (i) a non exposed
control group at Chikwaka, where no gold mining activities occur (42
participants); (ii) a medium exposed group, mothers living in Kadoma
but not working with mercury (51 participants); (iii) a high exposure
group, mothers living in Kadoma and working with mercury, e.g.
panning the ore or smelting amalgam (27 participants). Half of this
group was performing panning with mercury, the other half of the
women was involved in amalgam smelting.
2.2. Questionnaire
A questionnaire was used to assess the exposure situation, possible
confounders, the general health and nutrition status. The anamnestic
part of the questionnaire was translated from English into the national
language “Shona” in Munich before the project started. The Shona
questionnaire was retranslated into English in Zimbabwe before it was
printed to control the quality of translation. Study nurses filled in each
questionnaire asking the mothers the questions in Shona. The study was
conducted in accordance to national regulations for health studies.
2.3. Laboratory methods
2.3.1. Collection of urine, breast milk and hair samples
From 120 participants in Kadoma and Chikwaka 120 breast milk
samples, 117 hair samples from mothers and 81 hair samples from
children, 120 urine samples from mothers and 120 urine samples from
children were selected. All samples were on the spot samples at the day
of the examination, children's urine was collected with a special urine
bag for infants. To avoid degradation, all breast milk and urine samples
(acidified) were transported in an electric cooling box and stored per-
manently at 4 °C until the analyses were performed.
2.3.2. Mercury analysis
Urine and hair samples of mother and child as well as breast milk
were analyzed for total mercury with CV-AAS (PerkinElmer, 1100B
Spectrometer with MHS 20 and an amalgamation unit), using sodium-
borohydride (NaBH4) for reduction. Internal and external quality con-
trol was performed.
S. Bose-O’Reilly, et al.
2.3.3. Preparation and limit of detection of breast milk samples
1 ml of the breast milk sample was digested in closed PTFE-lined
Parr pressure digesters (Kuerner, Rosenheim, Germany) with 1,0 ml of
65% nitric acid (Suprapur Grade, Merck, Darmstadt, Germany) for
6 h at 140 °C. After cooling, the sample was diluted to 5.0 ml with
redistilled water.
The limit of detection for breast milk is 1.0°[μg/l], the results below
the limit of detection were set to 0.5°[μg/l] for further statistical ana-
lysis.
2.3.4. Preparation and limit of detection of hair samples
When available, 20 mg–200 mg hair was cut in small pieces accu-
rately weighed and dissolved in nitric acid in a closed pressure vessel at
120 °C. Parts of the solution were diluted and analyzed by CV-AAS.
Initial washing steps with water, detergents or organic solvents like
acetone, were not performed although this procedure is frequently
applied before hair analysis with the aim to remove air-borne heavy
metal pollution from the outer layer of hair. But as shown in literature,
a distinct differentiation between exogenous and endogenous mercury
cannot be achieved (Kijewski, 1993). Preliminary tests supported this
assumption. During these tests, hair samples of mercury burdened vo-
lunteers were washed before analysis. The obtained results even from
the same strand of hair were not reproducible. The limit of detection
(LOD) for hair in the laboratory was 0.02°[μg/g], the results below the
limit of detection were set to 0.01°[μg/g] for further statistical analysis.
2.3.5. Preparation and limit of detection of urine samples
The urine samples were analyzed without any further pre-proces-
sing. The limit of detection for urine is 0.5°[μg/l], the results below the
limit of detection were set to 0.25°[μg/l] for further statistical analysis.
2.4. Statistical methods
The available data was descriptively analyzed including tables and
box-plots (with minimum/maximum levels and 95%, 75%, 50% 25%,
5% percentiles). The daily uptake of breast milk of an infant is esti-
mated as on average 0.120 [l/d * kg BW] (Butte et al., 2002). Mercury
in breast milk was analyzed, the weight of the child is known, therefore
the daily uptake was calculated as body weight°[kg] * 0.120°[l] * Hg in
breast milk°[μg/l]. Chi-square (Pearson) was used to test for any dif-
ference between all different exposure-groups for clinical ordinal data.
The statistical analysis of metric data like age, height and weight was
performed with ANOVA and T-test. Mercury levels had a tail to the
right. Therefore were log transformed for the figures. Kruskal Wallis
and Mann-Whitney test were used to analyze the non-parametric data.
Spearman's rank correlation coefficient test was used to analyze for
possible correlations between the different human specimens. Logistic
regression was performed using the backwards stepwise Wald method,
to identify significant factors that influence the body burden with
mercury of the infants. The significance level used was α = 0.05. The
confidence interval used was 95%. For statistical analysis, SPSS® 17.0
program was used.
3. Results
There were no significant differences between the three subgroups
relating to age, height and weight of children and mothers (see
Table 1).
The mothers’ health was generally poor and most of them reported
having suffered from malaria prior to the sampling exercise. Artisanal
and small-scale miners are mobile men with money, and they form a
high-risk group for spreading the HIV-virus in the community and into
other areas. Most women in Kadoma area had performed a HIV test
during pregnancy, if found positive the women were given an antiviral
treatment. The local health experts reported that more than 50% of the
women were found to be HIV positive.
3
Environmental Research 184 (2020) 109266
The health status of the infants in the examined areas was de-
termined by poverty. Many children suffered from skin problems,
diarrhea and upper respiratory tract infections. Malaria and severe
lower respiratory tract infections were the most serious health hazard
for infants in this area. The infants examined were well nourished, well
developed and healthy. The hygienic condition was poor in the villages
where ASGM occurs.
3.1. Exposure and confounder assessment
Only one of the women in the control group had ever worked with
mercury, but since the exposure was low and long ago, she was not
excluded from further analysis. The women in the medium exposed
group had only lived in the area where environmental exposure is
common. The women in the high exposure group were either panners
or amalgam smelters. Other family members might been working with
mercury, which was the case in 29% families in the medium exposed
group and in 59% families in the high exposure group. Mercury was
stored at home, which meant a potential local source of mercury vapor
exposure.
Possible confounders were analyzed like smoking and alcohol con-
sumption, there were no significant differences (p > 0.05) between
the three subgroups. These variables did not have an influence on the
amount of mercury in the human specimens. Besides, none of the vo-
lunteers had any dental amalgam fillings. Some of the variables differed
significantly between the groups: e.g. eating fish was more frequent in
group of medium and especially high exposure participants
(p < 0.001), use of pesticides in both exposed groups (p < 0.001),
use of kerosene for cooking in the medium exposed group (p = 0.038).
The financial status was considered to be better in the exposed groups
(p = 0.022), which might explain the difference in fish consumption,
use of kerosene and pesticides due to more disposable cash/income.
Fish consumption influences mercury levels in human specimens,
and if compared over all groups it looked, as if it would have been
significant, but one reason might have been, that in the non-exposed
groups participants only rarely consumed fish, different to both ex-
posure groups with higher fish consumption. Analyzing the three sub-
groups separately no significant differences due to fish consumption
were identified (see supplement Table S1). There is no significant cor-
relation between fish consumption frequency and mercury levels in
breast milk (p = 0.589).
3.2. Human-bio-monitoring
The main results from the laboratory analyses are shown in Fig. 1
(details in the supplement Table S2). A total of 120 breast milk samples
were analyzed. The participants from Kadoma, which had contact with
mercury, showed significantly higher levels of mercury in breast milk
than the two other groups (p < 0.001) (see Fig. 1g).
In urine of the mothers, the median level of total Hg was below LOD
for the control group, for the medium exposure group the median was
4.15°[μg/l], and for the high exposure group 8.45°[μg/l]; maximum
level was 441°[μg/l] (see Fig. 1a). For children the median level was
below LOD for the control group, for the medium exposure group the
median was 1.41°[μg/l], and for the high exposure group it was
4.24°[μg/l]; the maximum level was 86.9°[μg/l] (see Fig. 1b).
The results for mercury in urine were adjusted to creatinine in urine,
and they were for mothers as follows: The median level was 0.03°[μg
Hg/g° creatinine] for the control group, for the medium exposure group
the median was 2.58°[μg Hg/g° creatinine], and for the high exposure
group it was 6.60°[μg Hg/g° creatinine]; the maximum level was
286°[μg Hg/g° creatinine] (see Fig. 1c). In the urine of the children the
median level was 1.43°[μg Hg/g° creatinine] for the control group, for
the medium exposure group the median level was 9.83°[μg Hg/g°
creatinine], and for the high exposure group it was 25.0°[μg Hg/g°
creatinine]; the maximum level was 258°[μg Hg/g° creatinine] (see
S. Bose-O’Reilly, et al.
Table 1
Age, weight and height (mother and children), no statistical significant difference between the groups (p > 0.05) (ANOVA test.
Fig. 1d).
The median level for mercury in hair of mothers was in the control
group below LOD in the medium exposure group it was 1.10°[μg/g], in
the high exposure group it was 1.47°[μg/g], and the maximum level
was 27.8°[μg/g] (see Fig. 1e). The median level for mercury in chil-
dren's hair was in the control group below LOD, in the medium ex-
posure group it was 0.94°[μg/g], in the high exposure group it was
1.24°[μg/g], and the maximum level was 9.70°[μg/g] (see Fig. 1f).
The median of the breast milk samples was 0.50°[μg/l] in the con-
trol group, 1.10°[μg/l] in the medium exposure group and 1.20°[μg/l]
in the high exposure group, the maximum level was 24.80°[μg/l] in the
high exposed group (see Fig. 1g).
A Spearman's rank correlation coefficient test was performed to find
possible correlations between the different human specimens. All
human specimens were correlating significantly with one another (see
supplement Table S2).
The daily uptake of mercury from breast milk for the children was
calculated. The reference dose of 0.3°[μg°Hg/kg BW/d] was used to
compare the uptake of the infants in this study. No infant from the
control area, nine infants or 17.6% in the medium exposure group and
five infants or 18.5% in the high exposure group had a daily uptake of
mercury via breast milk above the RfD (see Fig. 2). The differences
compared with the control group were significant (Mann-Whitney test:
not exposed versus medium exposed p < 0.001, not exposed versus
high exposed p < 0.001, medium exposed versus high exposed not
significant).
The reference value 0.70°[μg/l] for mercury in infant's urine was
exceeded in 11 of 42 (26.2%) infants from the control group, in 39 of 51
(76.5%) from the medium exposed group and in 23 of 27 (85.2%) from
the high exposed group (Mann-Whitney test: not exposed versus
medium exposed p < 0.001, not exposed versus high exposed
p < 0.001, medium exposed versus high exposed not significant).
3.3. Transfer parameters
A logistic regression was performed using the backwards stepwise
Wald method, to identify the factors that significantly influence the
body burden with mercury of the infants.
The reference value of 0.70°[μg/l] for mercury in infant's urine was
taken as decision point. Several factors were stepwise excluded; such as
mercury in mother's breast milk, hair and urine, consumption of fish by
mothers and other food, age of infant and mother, food supplements of
the infant. The regression coefficient of the logistic regression was
0.473 and is significant (p = 0.015). The factors with a significant
influence were the human biomonitoring levels of mercury in urine of
4
Environmental Research 184 (2020) 109266
the mother, fruit consumption of the infant, chicken consumption of the
mother (see Table 3). But mainly the body burden of the mother,
measured by mercury in urine, is of importance for the infant.
Secondly the daily uptake of mercury via breast milk with the re-
ference dose of 0.3 [μg/kg bw/d] was considered as decision point and
similar factors were included, plus the mercury in urine of the child.
The regression coefficient of the logistic regression was −2.262 which
is highly significant (p < 0.001). The factors with a significant influ-
ence were the human biomonitoring levels of mercury in urine in [μg/l]
and hair [μg/g] of the mother and fruit consumption of the infant (see
Table 4).
The body burden of the infant depends therefore significantly on the
body burden of the mother, as indicated by increased mother's mercury
levels. Living in an exposed area can be interpreted in both ways as
direct burden of the infant and as indirect burden via mother. The
consumption of egg, fruit and the age of the child indicated that as soon
as breast milk is not the only source of food the body burden of child
decreased (Table 4).
4. Discussion
4.1. Human biomonitoring results
In order to evaluate variations of the levels of mercury in breast
milk, urine and hair between the various groups, the Kruskal-Wallis-
Test was used and a pattern was clearly recognizable. The group which
lived and/or worked in the Kadoma mining areas had a significantly
higher amount of mercury in breast milk, hair and urine. The Mann-
Whitney test was used to investigate the differences between the in-
dividual groups (see supplement Table S2).
The amount of mercury in urine of children showed a significant
difference between the contaminated groups and the control group. The
highest level of mercury was found in the group from Kadoma mining
areas which was highly exposed to mercury.
To discuss the risk, a comparison of the results from this study with
threshold values available from the German Human Biomonitoring
Commission were taken for urine (Commission Human Biomonitoring
of the Federal Environmental Agency, 1999; Ewers et al., 1999). These
threshold levels (called Human Bio-Monitoring levels - HBM) were used
to describe the health risks from mercury exposure. Results below the
first threshold HBM°I are considered as “safe” levels (Schulz et al.,
2007). Results between HBM°I and HBM°II should be taken as “alert”
levels and reducing the level of exposure as reasonably as is achievable
is indicated (Ewers et al., 1999), adverse health effects are possible
(Doering et al., 2016). Mercury levels above HBM°II are seen as “action”
S. Bose-O’Reilly, et al. Environmental Research 184 (2020) 109266

Fig. 1. Hg in urine, hair and breast milk, for children and mothers, for exposure groups.

levels, negative health effects are likely, and exposure reduction is es-
sential (Drasch et al., 2002; Ewers et al., 1999). For mercury in urine
the HBM°I level is 7°[μg/l] and the HBM°II level is 25°[μg/l]. For
mercury in urine in relation to the excretion of creatinine the HBM°I
level is 5°[μg Hg/g° creatinine] and the HBM°II level is 20°[μg Hg/g°
creatinine]. The reference value derived for mercury in urine of adults
aged 18–69 using data from GerES 1998 for mercury in urine is
1.0°[μg/l] (Commission “Human-Biomonitoring” German Federal
Environmental Agency, 2003; Schulz et al., 2007). From the Seychelles
study a no observed adverse effect level (NOAEL) was derived for hair
below 5°[μg/g] (Agency for Toxic Substances and Disease Registry,
2007; Davidson et al., 1998). Taking uncertainty into account the US

5
S. Bose-O’Reilly, et al.
Fig. 1. (continued)
EPA calculated the “benchmark limit” for total Hg in hair to 1°[μg/g]
(Rice, 2004; Rice et al., 2003; U.S. Environmental Protection Agency,
1997). The tolerable limits of mercury concentrations in breast milk
samples, as e.g. recommended by the US – EPA Reference Dose” for
inorganic mercury is 0.3°[μg/kg BW/d] (U.S. Environmental Protection
Agency, 1997). The calculated daily uptake of mercury of each child
was compared to the reference dose 0.3°[μg Hg/kg BW/d] and a deci-
sion was made for each infant if the daily uptake was above or below
the reference dose. A reference value of 0.7°[μg/l] for mercury in urine
of children aged 6–14 years was used (Wilhelm et al., 2006). The HBM
levels are the same as for adults. For mercury in urine the HBM°I level is
7°[μg/l] and the HBM°II level is 25°[μg/l]. For mercury in urine in re-
lation to the excretion of creatinine the HBM°I level is 5°[μg Hg/g°
creatinine] and the HBM°II level is 20°[μg Hg/g° creatinine]
(Commission Human Biomonitoring of the Federal Environmental
Agency, 2005; Schulz et al., 2007). No specific data is available for
mercury in hair for infants; therefore the same levels as for mothers are
used.
Comparing to threshold values, all children in the control group
6
Environmental Research 184 (2020) 109266
were below the German Reference value of 1.0 [μg/l] (Commission
“Human-Biomonitoring” German Federal Environmental Agency, 2003;
Schulz et al., 2007). Compared to threshold limits for mercury in urine
and hair, nobody in the control group exceeded the threshold limits (see
Table 5). The participants from Kadoma with medium exposure to
mercury and high exposure to mercury were in a high percentage above
the threshold limits.
Comparing the results for breast milk from this study with other
breast milk studies from Sweden (Oskarsson et al., 1996), Germany
(Drasch et al., 1998), Austria (Gundacker et al., 2002), Iran
(Dahmardeh Behrooz et al., 2012) and Slovenia showed that the control
group was a non-exposed group with background levels (see Table 6). A
study from mining areas in Colombia with 150 breast milk samples
came up with an average mercury level of 2.4 μg/l (Molina et al., 2018).
A study from a hospital in an mining area in Ghana with 114 breast
milk samples showed a geometric median of Hg of 8,62 μg/l (Bansa
et al., 2017).
The excellent overview paper of worldwide total mercury levels in
breast milk presented the data from 7815 breast milk studies,
S. Bose-O’Reilly, et al. Environmental Research 184 (2020) 109266

Fig. 2. Daily uptake of mercury via breast milk in μg/kg bodyweight/day (boxplots of subgroups, reference dose 0.3 μg/kg bodyweight/day for comparison, Mann-
Whitney test: not exposed versus medium exposed p < 0.001, not exposed versus high exposed p < 0.001, medium exposed versus high exposed not significant).

Table 2
Correlation between the different human biomonitoring results (Spearman's rank correlation coefficient).
Hg breast milk [μg/l] PC 1
Sig
N 120
Hg urine mother [μg/l] PC 0.928a 1
Sig 0.000
N 120 120
Hg urine child [μg/l] PC 0.831a 0.847a 1
Sig 0.000 0.000
N 120 120 120
Hg urine mother [μg/g PC 0.762a 0.846a 0.590a 1
creatinine] Sig 0.000 0.000 0.000
N 120 120 120 120
Hg urine child [μg/g PC 0.673a 0.682a 0.619a 0.826a 1
creatinine] Sig 0.000 0.000 0.000 0.000
N 120 120 120 120 120
Hg hair mother [μg/g] PC 0.551a 0.462a 0.295a 0.654a 0.572a 1
Sig 0.000 0.000 0.001 0.000 0.000
N 117 117 117 117 117 117
Hg hair child [μg/g] PC 0.356a 0.236b 0.278b 0.362a 0.446a 0.523a 1
Sig 0.001 0.034 0.012 0.001 0.000 0.000
N 81 81 81 81 81 79 81
Hg breast milk Hg urine mother Hg urine child Hg urine mother [μg/g Hg urine child [μg/g Hg hair mother Hg hair child
[μg/l] [μg/l] [μg/l] creatinine] creatinine] [μg/g] [μg/g]

Pearson correlation = PC; Significance two tailed = Sig, case number = N.

a
Correlation is significant at the 0.01 level (2-tailed).
b
Correlation is significant at the 0.05 level (2-tailed).

Table 3
Mercury in urine of infant below versus above reference dose 0.7 μg/l (logistic regression analysis (backwards stepwise Wald)).
B S.E. Wald df Sig. Exp(B) 95% C.I.for EXP(B)

Lower Upper

Hg in mother's urine −0.207 0.064 10.395 1 0.001 0.813 0.717 0.922

Hg in mother's urine/g creatinine 1.078 0.291 13.717 1 0.000 2.940 1.661 5.202
Chicken consumption of mother 0.629 0.350 3.222 1 0.073 1.875 0.944 3.725
Fruit consumption of child −0.371 0.225 2.722 1 0.099 0.690 0.444 1.072
Constant −1.160 0.572 4.120 1 0.042 0.313

Without influence: Fish consumption mother, Hg in breast milk, Hg in hair. chicken-. meat- and vegetables consumption of mother, years spend in the area,
occupation, age of mother or child, porridge-, egg-, milk- or vegetable consumption of child.

7
S. Bose-O’Reilly, et al. Environmental Research 184 (2020) 109266

Table 4 through the environment, and the special exposure risk of the breast fed
Daily uptake of mercury via breast milk below versus above reference dose infant.
0.3 μg/kg bodyweight/d (logistic regression analysis (backwards stepwise
Wald)).
4.2. Limitations of the study
B S.E. Wald df Sig. Exp(B)

Hg in mother's urine/g creatine 0.204 0.083 6.021 1 0.014
Hg in mother's hair 1.150 0.435 6.987 1 0.008
Fruit consumption of child −3.577 2.114 2.863 1 0.091
Constant −5.127 1.539 11.092 1 0.001
Without influence: Fish consumption mother, Hg in urine, chicken-. meat- and
vegetables consumption of mother, years spend in the area, occupation, age of
mother or child, porridge-, egg-, milk- or vegetable consumption of child.
extrapolating the levels Hg levels for the year 2015 based on the linear-
fit of Hg data. Those extrapolated average Hg breast milk levels where
for all 7815 samples 0.24 μg/l, for 226 samples Africa 3.94 μg/l and for
2937 background samples 0.10 μg/l (Sharma et al., 2019).
The exposed group from Kadoma had a median comparable with the
Faroe study (Grandjean et al., 1995), the first breast milk study from
GMP (Bose-O'Reilly et al., 2008b) and the study of a mining population
in Ghana (Molina et al., 2018). The study from Ghana of a heavily
polluted mining area even showed higher levels compared to our study
(Bansa et al., 2017).
The correlation between breast milk and urine from mother and
child was also analyzed and the result was that the amount of mercury
in breast milk correlated well with the mercury level in children's urine,
the correlation is 0.831 (see Table 2). Moreover mercury in breast milk
seemed to be a good indicator for exposure because the higher the le-
vels of mercury in breast milk the higher was the amount of mercury in
mother's and children's urine (see Fig. 1).
The exposure of the infants in gold mining areas with mercury oc-
curred directly via mercury vapor, as the correlation between mother's
and children's urine and hair results suggested (see Fig. 3). But there
was possibly an additional exposure pathway from mother to child via
breast milk in those areas, which added to the first exposure pathway.
Summing up these two main pathways resulted in a higher body burden
of infants compared to their mothers.
Breast milk is a good indicator for both the exposure of the mother
1.226 This study was based on a limited number of cases. It would be
3.159 important to perform a similar study with a higher case number, and as
0.028 well in different gold mining regions. There were not sufficient cases to
0.006
distinguish between the heavier exposed amalgam smelters and the
generally lesser exposed panners, so they were grouped together as one
group exposed by their working behaviors applying mercury. Only total
mercury was analyzed, due to limited funding, speciation of mercury in
hair, was not possible. Results of our previous study in 2004 in the same
areas showed very clearly, that the mercury in hair in the control area
was mainly methyl-mercury (70%), whereas the mercury in Kadoma
was in the general population and in miners mainly inorganic mercury
(75–85%) (Bose-O'Reilly et al., 2004). Due to the lack of speciation it is
unknown to what extend the mercury found in breast milk is methyl
mercury. The gastro-intestinal absorption of methyl mercury is much
higher compared to inorganic mercury.
A cross sectional study cannot answer the important question of
trans placental transfer of mercury from mother to fetus.
5. Conclusion
Breast fed children in the gold mining area of Kadoma were severely
exposed with mercury. The pathways seemed to be directly by living in
an exposed area, inhaling mercury vapor from smelting processes next
to the infant's home and as soon as they eat solid foods by contaminated
local food products. Another important pathway was the mother's
breast milk, mercury levels were above safety standards. The body
burden of infants did correlate well with the body burden of the mo-
ther, and with the mercury level in breast milk. Children were therefore
seriously exposed with the neurotoxic mercury.
This study was performed with very limited funding still it shows a
serious health hazard for children. Further research is needed, including
speciation of samples and neurodevelopmental studies, to assess the
public health dimension and relevance of this problem.

Table 5
Distribution of the results in the three exposure subgroups compared to threshold levels (HBM = human biomonitoring level).

8
S. Bose-O’Reilly, et al. Environmental Research 184 (2020) 109266

Table 6
Hg in breast milk [μg/l] – results from literature compared to study results.
Case number Minimum Median Arithmetic. mean 95% percentile Maximum Source

Sweden 1996 Background study 30 0.1 – 0.6 – 2 Oskarsson et al. (1996)

Germany 1998 Background study 70 < 0.2 0.37 – – 6.86 Drasch et al. (1998)
Austria 2002 Background study 116 – – 1.59 – 6.8 Gundacker et al. (2002)
Iran 2011 Background study 80 < LOD 0.08 0.39 — 5.86 Dahmardeh Behrooz et al.
(2012)
Slovenia Background study 470 < LOD 0.14 0.16 0.64 3.39 Snoj Tratnik, 2019
Global 1966–2015 Background study 2937 0.1 Sharma et al. (2019)
Controls Zimbabwe Background 42 < 0.5 < 0.5 0.54 < 0.5 1.55 This project
Faroe 1995 High consumption of 88 0.3 2.45 – – 8.7 Grandjean et al. (1995)
fish
Indonesia and Zimbabwe ASGM 46 < 1.0 1.87 8.11 46.7 149.6 Bose-O'Reilly et al. (2008b)
2004
Ghana ASGM 114 1.4 7.85 8.62 10.4 65.4 Bansa et al. (2017)
Colombia ASGM 150 2.11 2.5 Molina et al. (2018)
Exposed Zimbabwe ASGM 78 < 0.5 1.1 2.16 7.55 24.8 This project

References

Fig. 3. Scatterplot for correlation of mercury in urine of the child –mercury in
urine of the mother (confidence interval 95%).
Funding
This work was supported by from a personal grant (BL), which was
used for the cost of the field project. Laboratory work and statistical
analysis did not receive any specific grant from funding agencies in the
public, commercial, or not-for-profit sectors.
Declaration of competing interest
There are no competing financial interests.
Acknowledgements
We would like to thank very much the Local Health Office in
Kadoma, Dr Yama, Nyarai Tembo, Edith Mhike, Felistas Navhaya, Dr.
Edwin Muguti, Spencer Kahwai, the Mayflower School, the health units
in Brompton, Venice and Chikwaka; as well all participants of the
medical examinations. We acknowledge very much the support for this
work from Victoria Baskova-Schilling and Jennifer Bäuml.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.envres.2020.109266.
Agency for Toxic Substances and Disease Registry, 2007. Toxicol. Profile Mercury 1999.
Amin-Zaki, L., et al., 1981. Methylmercury poisoning in the Iraqi suckling infant: a
longitudinal study over five years. J. Appl. Toxicol. 1, 210–214.
Anderson, H.A., Wolff, M.S., 2000. Environmental contaminants in human milk. J. Expo.
Anal. Environ. Epidemiol. 10, 755–760.
Bansa, D.K., et al., 2017. Cross-sectional assessment of infants' exposure to toxic metals
through breast milk in a prospective cohort study of mining communities in Ghana.
BMC Publ. Health 17, 505.
Billaud, P., et al., 2004. Removal of Barriers to the Introduction of Cleaner Artisanal Gold
Mining and Extraction Technologies in Kadoma. Zimbabwe - Final report. part A
Environmental assessment, Orleans, France UNIDO Project EG/GLO/01/G34 No.03/
089. BRGM Project Nr 53320-FR. BRGM (Bureau de Recherches Géologiques et
Minières). www.unites.uqam.ca/gmf/intranet/gmp.
Bose-O'Reilly, S., et al., 2017. Signs and symptoms of mercury-exposed gold miners. Int. J.
Occup. Med. Environ. Health 30, 249–269.
Bose-O'Reilly, S., et al., 2004. Removal of Barriers to the Introduction of Cleaner Artisanal
Gold Mining and Extraction Technologies in Kadoma, Zimbabwe - Final Report, Part
B Health Assessment. UNIDO Project EG/GLO/01/G34 No.03/089. BRGM Project Nr
822657-3. BRGM (Bureau de Recherches Géologiques et Minières), Orleans, France,
pp. 133.
Bose-O'Reilly, S., et al., 2010a. Health assessment of artisanal gold miners in Indonesia.
Sci. Total Environ. 408, 713–725.
Bose-O'Reilly, S., et al., 2010b. Health assessment of artisanal gold miners in Tanzania.
Sci. Total Environ. 408, 796–805.
Bose-O'Reilly, S., et al., 2009. Reply to the letter to the editor by J.G. Dorea Health hazard
for infants: breast-milk mercury or non-breastfeeding? A dilemma for mothers ex-
posed to mercury vapours. Int. J. Hyg Environ. Health 212, 235–237.
Bose-O'Reilly, S., et al., 2008a. Mercury as a serious health hazard for children in gold
mining areas. Environ. Res. 107, 89–97.
Bose-O'Reilly, S., et al., 2008b. Mercury in breast milk - a health hazard for infants in gold
mining areas? Int. J. Hyg Environ. Health 211, 615–623.
Bose-O'Reilly, S., et al., 2016. A preliminary study on health effects in villagers exposed to
mercury in a small-scale artisanal gold mining area in Indonesia. Environ. Res. 149,
274–281.
Butte, N.F., et al., 2002. Nutrient Adequacy of Exclusive Breastfeeding for the Term Infant
during the First Six Months of Life. WHO, Geneva.
Clarkson, T.W., et al., 2003. The toxicology of mercury–current exposures and clinical
manifestations. N. Engl. J. Med. 349, 1731–1737.
Commission “Human-Biomonitoring” German Federal Environmental Agency, 2003.
Update of refence values for lead, cadmium and mercury in blood and urine of adults.
Bundesgesundheitsblatt - Gesundheitsforsch. - Gesundheitsschutz 46, 1112–1113.
Commission Human Biomonitoring of the Federal Environmental Agency, 1999.
Monograph mercury - reference values and human biomonitoring values.
Bundesgesundheitsblatt 42.
Commission Human Biomonitoring of the Federal Environmental Agency, 2005. New and
updated reference values of contaminants in blood and urine of children - arsenic,
lead, cadmium and mercury. Bundesgesundheitsblatt 11, 8–16.
Dahmardeh Behrooz, R., et al., 2012. Mercury concentration in the breast milk of Iranian
women. Biol. Trace Elem. Res. 147, 36–43.
Davidson, P.W., et al., 1998. Effects of prenatal and postnatal methylmercury exposure
from fish consumption on neurodevelopment: outcomes at 66 months of age in the
Seychelles Child Development Study. Jama 280, 701–707.
Doering, S., et al., 2016. Essential indicators identifying chronic inorganic mercury in-
toxication: pooled analysis across multiple cross-sectional studies. PloS One 11,
e0160323.
Dorea, J.G., 2008. Health hazard for infants: breast-milk mercury or non-breastfeeding? A
dilemma for mothers exposed to mercury vapours. Int. J. Hyg Environ. Health 212,
233–234.
Drasch, G., et al., 1998. Mercury in human colostrum and early breast milk. Its

9
S. Bose-O’Reilly, et al.
dependence on dental amalgam and other factors. J. Trace Elem. Med. Biol. 12,
23–27.
Drasch, G., et al., 2001. The Mt. Diwata study on the Philippines 1999 - assessing mercury
intoxication of the population by small scale gold mining. Sci. Total Environ. 267,
151–168.
Drasch, G., et al., 2002. Scientific comment on the German human biological monitoring
values (HBM values) for mercury. Int. J. Hyg Environ. Health 205, 509–512.
Ewers, U., et al., 1999. Reference values and human biological monitoring values for
environmental toxins. Report on the work and recommendations of the Commission
on Human Biological Monitoring of the German Federal Environmental Agency. Int.
Arch. Occup. Environ. Health 72, 255–260.
Grandjean, P., et al., 1994. Human milk as a source of methylmercury exposure in infants.
Environ. Health Perspect. 102, 74–77.
Grandjean, P., et al., 1995. Milestone development in infants exposed to methylmercury
from human milk. Neurotoxicology 16, 27–34.
Grandjean, P., et al., 1997. Cognitive deficit in 7-year-old children with prenatal exposure
to methylmercury. Neurotoxicol. Teratol. 19, 417–428.
Grandjean, P., White, R., 2002. Neurodevelopmental disorders. In: Tamburlini, G. (Ed.),
Children' S Health and Environment: a Review of Evidence. WHO - EEA, Kopenhagen.
Gundacker, C., et al., 2002. Lead and mercury in breast milk. Pediatrics 110, 873–878.
Kijewski, H. (Ed.), 1993. [The Forensic Impact of the Concentration of Mineral Nutrients
in Human Scalp Hair]. Verlag schmidt-Römhild, Lübeck.
Lettmeier, B., et al., 2010. Proposal for a revised Reference Concentration (RfC) for
mercury vapour in adults. Sci. Total Environ. 408, 3530–3535.
Marques, R.C., et al., 2016a. Traditional living in the Amazon: extended breastfeeding,
fish consumption, mercury exposure and neurodevelopment. Ann. Hum. Biol. 43,
360–370.
Marques, R.C., et al., 2016b. Neurodevelopment of Amazonian children exposed to
ethylmercury (from Thimerosal in vaccines) and methylmercury (from fish). Environ.
Res. 149, 259–265.
Marques, R.C., et al., 2016c. Impact of organic mercury exposure and home delivery on
neurodevelopment of Amazonian children. Int. J. Hyg Environ. Health 219, 498–502.
Marques, R.C., et al., 2007. Maternal mercury exposure and neuro-motor development in
breastfed infants from Porto Velho (Amazon), Brazil. Int. J. Hyg Environ. Health 210,
51–60.
Metcalf, S., Spiegel, S., 2007. Global Mercury Project Activities in Zimbabwe 2002-2007.
UNIDO.
Molina, C.F., et al., 2018. Mercury contamination in breast milk of nursing mothers in
gold mining municipalities of Antioquia, Colombia. Biomedica 38, 19–29.
Mudzwiti, P., et al., 2015. A golden opportunity: scoping study of artisanal and small
scale mining in Zimbabwe. In: PACT (Ed.), Harare, Zimbabwe.
Myers, G.J., Davidson, P.W., 2000. Does methylmercury have a role in causing devel-
opmental disabilities in children? Environ. Health Perspect. 108 (Suppl. 3), 413–420.
Myers, G.J., et al., 2003. Prenatal methylmercury exposure from ocean fish consumption
in the Seychelles child development study. Lancet 361, 1686–1692.
Oskarsson, A., et al., 1996. Total and inorganic mercury in breast milk in relation to fish
consumption and amalgam in lactating women. Arch. Environ. Health 51, 234–241.
Pajewska-Szmyt, M., et al., 2019. The impact of environmental pollution on the quality of
mother’s milk. Environ. Sci. Pollut. Res. Int. 26, 7405–7427.
Pronczuk, J., et al., 2002. Global perspectives in breast milk contamination: infectious
10
Environmental Research 184 (2020) 109266
and toxic hazards. Environ. Health Perspect. 110, A349–A351.
Rice, D.C., 2004. The US EPA reference dose for methylmercury: sources of uncertainty.
Environ. Res. 95, 406–413.
Rice, D.C., et al., 2003. Methods and rationale for derivation of a reference dose for
methylmercury by the U.S. EPA. Risk Anal. 23, 107–115.
Schulz, C., et al., 2007. The German human biomonitoring commission. Int. J. Hyg
Environ. Health 210, 373–382.
Schutzmeier, P., et al., 2016. Gold mining in Ecuador: a cross-sectional assessment of
mercury in urine and medical symptoms in miners from portovelo/zaruma. Int. J.
Environ. Res. Publ. Health 14.
Sharma, B.M., et al., 2019. An overview of worldwide and regional time trends in total
mercury levels in human blood and breast milk from 1966 to 2015 and their asso-
ciations with health effects. Environ. Int. 125, 300–319.
Shoko, D., 2014. Country diagnostic report on environmental health risk of mercury in
artisanal and small scale mining in Zimbabwe (unpublished draft). In: Srivastava, S.,
Hirji, R. (Eds.), Worldbank, pp. 64 Washington D.C.
Snoj Tratnik, J., et al., 2019. Results of the first national human biomonitoring in
Slovenia: trace elements in men and lactating women, predictors of exposure and
reference values. Int. J. Hyg Environ. Health 222, 563–582.
Solomon, G.M., Weiss, P.M., 2002. Chemical contaminants in breast milk: time trends and
regional variability. Environ. Health Perspect. 110, A339–A347.
Steckling, N., et al., 2011. Mercury exposure in female artisanal small-scale gold miners
(ASGM) in Mongolia: an analysis of human biomonitoring (HBM) data from 2008.
Sci. Total Environ. 409, 994–1000.
Steckling, N., et al., 2014. The burden of chronic mercury intoxication in artisanal small-
scale gold mining in Zimbabwe: data availability and preliminary estimates. Environ.
Health 13, 111.
Sundberg, J., et al., 1999a. Protein binding of mercury in milk and plasma from mice and
man–a comparison between methylmercury and inorganic mercury. Toxicology 137,
169–184.
Sundberg, J., et al., 1999b. Lactational exposure and neonatal kinetics of methylmercury
and inorganic mercury in mice. Toxicol. Appl. Pharmacol. 154, 160–169.
Tavares, L.M., et al., 2005. Performance on neurological development tests by riverine
children with moderate mercury exposure in Amazonia, Brazil. Cad. Saúde Pública
21, 1160–1167.
U.S. Environmental Protection Agency, 1997. Volume V: health effects of mercury and
mercury compounds. Mercury Study. Rep. Congress 349.
United Nations Environment Programme Chemicals, 2013. Global mercury assessment
2013. In: UNEP-Chemicals (Ed.), pp. 44 Geneva.
van Wijngaarden, E., et al., 2006. Benchmark concentrations for methyl mercury obtained
from the 9-year follow-up of the Seychelles Child Development Study.
Neurotoxicology 27, 702–709.
Vimy, M.J., et al., 1997. Mercury from maternal “silver” tooth fillings in sheep and human
breast milk. A source of neonatal exposure. Biol. Trace Elem. Res. 56, 143–152.
Wilhelm, M., et al., 2006. Revised and new reference values for arsenic, cadmium, lead,
and mercury in blood or urine of children: basis for validation of human biomoni-
toring data in environmental medicine. Int. J. Hyg Environ. Health 209, 301–305.
Yoshida, M., et al., 1992. Exposure to mercury via breast milk in suckling offspring of
maternal Guinea pigs exposed to mercury vapor after parturition. J. Toxicol. Environ.
Health 35, 135–139.