Regional variation in pesticide concentrations in plasma of delivering women residing in rural Indian Ocean coastal regions of South Africa.

Exposure to pesticides places pregnant women and the developing foetus at the highest risk. The objective of this study is to obtain an exposure assessment by investigating levels of pesticides in blood plasma of delivering women. We report on the concentrations of α, ß, γ HCH, endosulfan, HCB and the pyrethroids: cis-permethrin, cyfluthrin, cypermethrin and deltamethrin found in the maternal blood plasma of delivering women (n = 241) in three coastal sites of KwaZulu Natal. γ-HCH and endosulfan 1 and 2 were the most dominant pesticides in all three sites. Significantly, higher levels of γ-HCH and endosulfan were found in site 3 (vicinity of Empangeni) compared to the other two sites (p < 0.05). The GM levels for γ-HCH, endosulfan 1 and 2 were 956, 141 and 21 ng g(-1) lipids in site 3, respectively. The pyrethroid pesticides, HCB, α-HCH and ß-HCH were detected in less than 31% of the samples in all sites. γ-HCH correlated positively and strongly to both endosulfan 1 and 2 (r > 0.47), indicating a common source of exposure. The high levels of γ-HCH and endosulfan in maternal plasma samples in site 3 indicate the current and on-going exposure, which is of great concern for reproductive health and prenatal exposure.

Introduction to the Special Issue of IJERPH Entitled "Prenatal Exposure to Environmental Pollutants and Other Stressors: Impacts on Fetal Development, Birth Outcomes, Children's Health and Beyond".

Environmental pollution is a major cause of global diseases, death and disability, with a toll greater than that caused by communicable diseases, such as HIV/AIDS, malaria and tuberculosis, combined [...].

A Children's Health Perspective on Nano- and Microplastics.

BACKGROUND: Pregnancy, infancy, and childhood are sensitive windows for environmental exposures. Yet the health effects of exposure to nano- and microplastics (NMPs) remain largely uninvestigated or unknown. Although plastic chemicals are a well-established research topic, the impacts of plastic particles are unexplored, especially with regard to early life exposures. OBJECTIVES: This commentary aims to summarize the knowns and unknowns around child- and pregnancy-relevant exposures to NMPs via inhalation, placental transfer, ingestion and breastmilk, and dermal absorption. METHODS: A comprehensive literature search to map the state of the science on NMPs found 37 primary research articles on the health relevance of NMPs during early life and revealed major knowledge gaps in the field. We discuss opportunities and challenges for quantifying child-specific exposures (e.g., NMPs in breastmilk or infant formula) and health effects, in light of global inequalities in baby bottle use, consumption of packaged foods, air pollution, hazardous plastic disposal, and regulatory safeguards. We also summarize research needs for linking child health and NMP exposures and address the unknowns in the context of public health action. DISCUSSION: Few studies have addressed child-specific sources of exposure, and exposure estimates currently rely on generic assumptions rather than empirical measurements. Furthermore, toxicological research on NMPs has not specifically focused on child health, yet children's immature defense mechanisms make them particularly vulnerable. Apart from few studies investigating the placental transfer of NMPs, the physicochemical properties (e.g., polymer, size, shape, charge) driving the absorption, biodistribution, and elimination in early life have yet to be benchmarked. Accordingly, the evidence base regarding the potential health impacts of NMPs in early life remains sparse. Based on the evidence to date, we provide recommendations to fill research gaps, stimulate policymakers and industry to address the safety of NMPs, and point to opportunities for families to reduce early life exposures to plastic. https://doi.org/10.1289/EHP9086.

Toxic and essential elements in blood from delivering women in selected areas of São Paulo State, Brazil.

This study was designed to evaluate the degree of environmental contamination and possible exposure of pregnant women to toxic elements in seven selected areas of São Paulo State, Brazil. The overall median concentration of Mo in maternal blood was 0.53 µg L⁻¹, highly significant differences found between sites (p < 0.0001). Cd was found to be low overall - 0.09 µg L⁻¹ (0.01-0.58 µg L⁻¹) - with mothers from the Coastal and Rural 1 sites having the highest levels (p < 0.016).Median Hg concentration was 0.60 µg L⁻¹ (0.06 µg L⁻¹-4.35 µg L⁻¹); median Pb level was 16.2 µg L⁻¹ (3.5-57.7 µg L⁻¹) and no differences between sites were observed for both metals. Median Mn level was 16.7 µg L⁻¹ (7.0-39.7 µg L⁻¹), being highest in Urban 2 site (p < 0.016). Concentrations of maternal Co were found to range between 0.06 µg L⁻¹ and 1.1 µg L⁻¹ (median 0.25 µg L⁻¹) and As level was 0.60 µg L⁻¹ (0.10-3.8 µg L⁻¹) overall, with no statistical significance between sites for Co and As. Median Se concentrations were found to be 64 µg L⁻¹ (36-233 µg L⁻¹), with the highest median levels found in Urban 3 site; site differences were statistically significant (p < 0.0001). Correlation for each element (between paired maternal and cord blood) was measured only in Rural site 1; significant correlation was shown for Hg, Pb, Mn and Co (p < 0.05). These findings may be interpreted as indicating low environmental contamination in São Paulo State, Brazil. These findings could also indicate that pregnant women have little or no contact with pollutants, possibly due to awareness campaigns carried out by public health practitioners.

Selenium Status, Its Interaction with Selected Essential and Toxic Elements, and a Possible Sex-Dependent Response In Utero, in a South African Birth Cohort.

Selenium (Se) is an essential trace element and its deficiency in utero may affect fetus development and birth outcomes. The current study aimed to assess serum Se status at delivery and examine the possible association between Se levels and birth outcomes. The interaction of Se with selected essential and toxic elements as well as possible sex-dependent responses in utero were also evaluated. The negative association between Se levels and head circumference of neonates was evident in the total cohort (ß = -0.164; p < 0.001) as well as in the pre-term and full-term cohorts. Significant positive correlations were found between maternal serum Se concentrations and zinc (Zn) and copper (Cu) in the total and regional cohorts. In the total cohort, the toxic elements lead (Pb) and arsenic (As) showed a negative correlation with Se levels, while mercury (Hg), aluminum (Al) and cadmium (Cd) showed a positive correlation. The study found a sex-dependent response in utero for Zn, Cu, Pb, Hg, and Al. The findings of the current study may inform reproductive health policy on Se status in South Africa and highlight the need for sensitive methods to measure Se intake during pregnancy and its complex interactions with other micronutrients and environmental pollutants.

In Utero Exposure to Aluminium and Other Neurotoxic Elements in Urban Coastal South African Women at Delivery: An Emerging Concern.

Aluminium (Al) is a non-essential neurotoxicant and there is limited information regarding exposure to Al in utero. This study sought to evaluate the in utero exposure to Al in urban South African women, its effects on birth outcomes and possible synergistic effects between Al, essential and neurotoxic elements such as lead (Pb), mercury (Hg) and arsenic (As), as well as a a potential sex-dependent response to these elements in neonates. This study has found elevated levels of Al in urban women at delivery. The Spearman's rank correlation coefficients (p-value) of the association between maternal serum Al and birth outcomes (gestational age and parity), and between maternal serum Al and Cu, Zn and Se, were statistically significant. However, in the general and the stratified models, no association was found between any of the birth outcomes and maternal serum Al. The association between maternal serum Al and neurotoxic elements at delivery showed a significant positive correlation for Pb only (rho = 0.361; p < 0.001) which was found to be sex-dependent in neonates (males, rho = 0.285; p < 0.004 and females, rho = 0.444, p < 0.001). Our preliminary findings indicate that in utero exposure to Al is an emerging concern requiring further research and directives from public health authorities.

Levels of toxic and essential metals in maternal and umbilical cord blood from selected areas of South Africa--results of a pilot study.

This pilot study uses concentrations of metals in maternal and cord blood at delivery, in seven selected geographical areas of South Africa, to determine prenatal environmental exposure to toxic metals. Samples of maternal and cord whole blood were analysed for levels of cadmium, mercury, lead, manganese, cobalt, copper, zinc, arsenic and selenium. Levels of some measured metals differed by site, indicating different environmental pollution levels in the regions selected for the study. Mercury levels were elevated in two coastal populations studied (Atlantic and Indian Ocean sites) with mothers from the Atlantic site having the highest median concentration of 1.78 microg/L ranging from 0.44 to 8.82 microg/L, which was found to be highly significant (p < 0.001) when compared to other sites, except the Indian Ocean site. The highest concentration of cadmium was measured in maternal blood from the Atlantic site with a median value of 0.25 microg/L (range 0.05-0.89 microg/L), and statistical significance of p < 0.032, when compared to all other sites studied, and p < 0.001 and p < 0.004 when compared to rural and industrial sites respectively, confounding factor for elevated cadmium levels was found to be cigarette smoking. Levels of lead were highest in the urban site, with a median value of 32.9 microg/L (range 16-81.5 microg/L), and statistically significant when compared with other sites (p < 0.003). Levels of selenium were highest in the Atlantic site reaching statistical significance (p < 0.001). All analysed metals were detected in umbilical cord blood samples and differed between sites, with mercury being highest in the Atlantic site (p < 0.001), lead being highest in the urban site (p < 0.004) and selenium in the Atlantic site (p < 0.001). To the best of our knowledge this pilot investigation is the first study performed in South Africa that measured multiple metals in delivering mothers and umbilical cord blood samples. These results will inform the selection of the geographical sites requiring further investigation in the main study.

The placenta as a barrier for toxic and essential elements in paired maternal and cord blood samples of South African delivering women.

Environmental toxicants such as metals may be detrimental to foetus and infant development and health because of their physiological immaturity, opportunistic and differential exposures, and a longer lifetime over which disease, initiated during pregnancy and in early life, can develop. The placental mechanisms responsible for regulation of absorption and excretion of elements during pregnancy are not fully understood. The aim of this paper is to assess the correlation for selected toxic and essential elements in paired whole blood samples of delivering women and cord blood, as well as to evaluate the placental permeability for selected elements. Regression analyses used to assess this correlation in 62-paired samples of maternal and cord whole blood of delivering women show that the concentrations of mercury, lead, cobalt, arsenic and selenium in maternal and cord blood differed statistically. Lead, cobalt, arsenic and selenium appear to pass the placental barrier by a diffusion mechanism. It was also found that the mercury levels in cord blood were almost double those of the mother, suggesting that the foetus may act as a filter for the maternal mercury levels during pregnancy. Transplacental transfer for arsenic and cobalt was 80% and 45%, respectively, suggesting that the placenta modulates the rate of transfer for these elements. Cadmium, manganese, copper and zinc levels did not show statistically significant correlations between two compartments (maternal versus cord whole blood). The study confirms that most of the toxic metals measured have an ability to cross the placental barrier.

Prenatal Exposure to Aluminum and Status of Selected Essential Trace Elements in Rural South African Women at Delivery.

This study sought to evaluate the in utero exposure to aluminum and status of selected trace elements in South African women at delivery since aluminum is known to be toxic in all developmental stages even at low concentrations. Serum aluminum was negatively correlated with aluminum in urine, both uncorrected and corrected for creatinine, which suggests the retention of aluminum in body stores. Serum copper and zinc levels were found to be high in this study population. Serum copper levels were negatively correlated with aluminum in serum (ß = -0.095; p = 0.05). There was a marginal negative correlation between aluminum levels in serum and manganese levels in whole blood (ß = -0.087; p = 0.08). Copper levels in maternal serum were negatively correlated with birth weight and the length of neonates. There were a number of positive correlations between maternal characteristics and birth outcomes. Mothers who consumed root vegetables frequently appeared to be protected from aluminum retention and increased body burden since their serum aluminum levels were found to be significantly lower. The findings of the current study can be used as a baseline for further research on aluminum exposure and its associated interactions and outcomes in vulnerable populations.

Evaluation of in utero exposure to arsenic in South Africa.

Early life exposure to inorganic arsenic (iAs) has been shown to interfere with foetal and early childhood development, and is associated with morbidity and mortality in adulthood. The purpose of this study was to evaluate exposure to As in utero, to determine the association between maternal and cord blood of As and birth outcomes in South African populations. Total arsenic was measured in maternal blood of a total cohort (n=650) and in paired cord blood and urine of a subset cohort (n=317). Overall, the geometric mean (GM) of As in maternal blood was 0.62µg/L (n=650; 95% CI, 0.58-0.66). In the subset cohort, the GM of maternal blood As was 0.96µg/L (n=350; 95% CI, 0.91-1.02); in paired cord blood, the GM was 0.78µg/L (n=317; 95% CI, 0.74-0.83); and in urine (creatinine-corrected), the GM was 14.26µg/g creatinine (n=317; 95% CI, 12.64-16.09). A linear correlation was found between log maternal blood As and log cord blood As (rho=0.80, p<0.001). Birth outcomes showed geographical differences. in gestational age (p<0.001), birth length (p=0.019), head circumference (p<0.001), Apgar score at 5min (p<0.001) and parity (p<0.002). In a multivariate analysis, no association between maternal blood (AsB) levels and birth outcomes were found. However, the lower the gestational age, the higher the levels of maternal AsB (ß=-0.054; 95% CI-0.087 to -0.020) and mothers who had had at least one child were less likely to have higher AsB if compared to those who had never had any child (ß=-0.177; 95CI-0.322 to 0.031). In both univariate and multivariate analyses, being single, and drinking water from communal outdoor taps, boreholes and rivers was associated with higher As levels. The findings suggest that more research is needed to evaluate the impact of low level As exposure on postnatal development.

Reduction of in utero lead exposures in South African populations: Positive impact of unleaded petrol.

BACKGROUND: Prenatal exposure to lead (Pb) has been shown to have negative and irreversible health impacts on foetal and early childhood development, affecting morbidity and mortality in adulthood. This study aimed to assess in utero Pb exposure, examine birth outcomes, and identify confounding factors in the large cohort of South African population, following the legislated removal of Pb from petrol. METHODS: Lead was measured in the maternal blood, urine and cord blood using Inductive Coupled Plasma Mass spectrometry (ICP-MS). The statistical analyses included Spearman's correlation, Wilcoxon rank sum (Mann Whitney), Kruskal-Wallis rank tests and multivariate linear regression. RESULTS: Overall, the geometric mean (GM) of Pb in maternal blood (PbB) was 1.32 µg/dL (n = 640; 95% CI, 1.24-1.40). In the subset cohort, the GM of paired maternal PbB and cord blood (PbC) was 1.73 µg/dL (n = 350; 95% CI, 1.60-1.86) and 1.26 µg/dL (n = 317; 95% CI, 1.18-1.35), respectively with a positive correlation between the log PbB and the log PbC (rho = 0.65, p = <0.001). Birth outcomes showed geographical differences in the gestational age (p<0.001), birth length (p = 0.028) and head circumference (p<0.001), Apgar score at 5 min (p<0.001) and parity (p<0.002). In female neonates, a positive association was found between PbC and head circumference (rho = 0.243; p<0.016). The maternal PbB levels were positively correlated with race, educational status, water sources, cooking fuels and use of pesticides at home. CONCLUSIONS: This study has demonstrated not only the positive impact that the introduction of unleaded petrol and lead-free paint has had on in utero exposure to Pb in South Africa, but has also contributed new data on the topic, in a region where such data and scientific investigations in this field are lacking. Future research should evaluate if similar effects can be detected in young children and the adult population.

Prenatal Exposure to Cadmium, Placental Permeability and Birth Outcomes in Coastal Populations of South Africa.

BACKGROUND: The impact of prenatal exposure to cadmium (Cd) on birth outcomes is an area of concern. This study aimed to assess an impact of prenatal Cd exposure on birth outcomes in distinct coastal populations of South Africa. METHODS: Cadmium was measured in maternal blood (CdB) (n = 641), cord blood and in maternal urine (n = 317). This investigation assessed the associations between CdB (non-transformed) and birth outcomes across the 25th, 50th, and 75th percentile for birth weight, birth length and head circumference, to test for a linear trend. Associations between natural log-transformed maternal CdB, size at birth and other factors were further evaluated using linear mixed-effects modelling with random intercepts. RESULTS: The average gestational age in the total sample was 38 weeks; 47% of neonates were female, average birth weight was 3065 g and 11% were of low birth weight (< 2500 g). The geometric mean (GM) of the maternal CdB level was 0.25 µg/L (n = 641; 95% CI, 0.23-0.27). The cord blood Cd level was 0.27 µg/L (n = 317; 95% CI, 0.26-0.29) and urine (creatinine-corrected) Cd level was 0.27 µg/L (n = 318; 95% CI, 0.24-0.29). The CdB cord:maternal ratio in the sub-cohort was 1, suggesting that the placenta offers no protective mechanism to the foetus. An inverse association was found between CdB and the lower birth weight percentile in female neonates only (ß = - 0.13, p = 0.047). Mothers who reported eating vine vegetables daily had lower levels of CdB (ß = - 0.55, p = 0.025). Maternal smoking was associated with an elevation in natural log-transformed CdB levels in both male and female cohorts. DISCUSSION: Significant inverse associations between prenatal Cd exposure and birth anthropometry were found in female neonates but not in male neonates, suggesting potential sex differences in the toxico-kinetics and toxico-dynamics of Cd.

Prenatal exposure to manganese in South African coastal communities.

Exposure to environmental sources and altered physiological processes of manganese uptake during pregnancy and its possible effect on prenatal and postnatal development are of concern. This study investigates manganese blood levels at the time of delivery across four cohorts of pregnant women residing in coastal communities of South Africa and examines birth outcomes and environmental factors that could influence manganese levels in the study population. The geometric mean (GM) manganese blood levels (MnB) for all women at delivery was 15.2 µg L(-1). Collectively, rural women reported higher MnB concentrations (GM, 16.1 µg L(-1)) than urban women (GM, 13.5 µg L(-1), p < 0.001). Of the 302 cord blood samples drawn from the study participants (rural women only), GM MnB levels reported for three rural sites were 25.8 µg L(-1) (Rural 1), 33.4 µg L(-1) (Rural 2) and 43.0 µg L(-1) (Rural 3) and were twice as high as their respective maternal levels. However, no significant correlations were found between maternal and cord MnB levels across the 3 study areas. Factors associated with elevated maternal MnB levels, after adjusting for gestational age were: women living in a rural area (Rural 2) (p = 0.021); women drinking potable water from an outdoor/communal tap sourced from municipality (p = 0.021); drinking water from river/stream (p = 0.036); younger maternal age (p = 0.026); consuming leafy vegetables once a week (p = 0.034); and elevated maternal blood lead concentrations (PbB) (p = 0.002). The results indicate that MnB concentration in rural women during pregnancy is higher compared to urban women and increases with manganese intake from food and water.

Differences in prenatal exposure to mercury in South African communities residing along the Indian Ocean.

Mercury is a persistent environmental pollutant that has the potential to adversely affect human health, particularly, foetal neurodevelopment. The purpose of the study was to investigate prenatal mercury (Hg) exposure in the population in three sites along the South Africa coast. Study subjects included women (n=350) who were admitted for delivery at the local hospitals. Maternal and cord blood samples were collected to measure total mercury and each participant was required to answer a questionnaire. The 90th percentile of mercury levels in maternal and cord blood of the total population was 1.15 µg/l and 1.67 µg/l, respectively. Site 1 (Manguzi) participants had the highest maternal geometric mean (GM) values of 0.93 µg/l, which was significantly different from Site 2 (Port Shepstone) (0.49 µg/l) and Site 3 (Empangeni) (0.56 µg/l) (ANOVA test, p<0.001). Umbilical cord blood GM Hg level for Site 1 (1.45 µg/l) was more than double the GM Hg level in Site 2 (0.70 µg/l) and Site 3 (0.73 µg/l). Univariate analysis indicated that the following maternal characteristics were positive predictors for elevated umbilical cord Hg levels: maternal blood Hg levels, living with a partner, residing in Site 1, living in informal housing, using wood and gas for cooking, borehole water as a drinking source, and a member of the household being involved in fishing. Maternal dietary predictors of elevated Hg levels in umbilical cord blood included consuming fresh fish, tinned fish, fruit or dairy products, daily. This study provides baseline data and reveals that 2% of the study population were above the EPA's reference value (5.8 µg/l) suggesting low level exposure to mercury in pregnant women and the developing foetus in South Africa. Further research is required to explore the sources of elevated Hg levels in Site 1.

Prenatal exposure to DDT in malaria endemic region following indoor residual spraying and in non-malaria coastal regions of South Africa.

Exemption was granted by the Stockholm Convention in 2004 for use of DDT by indoor residual spraying (IRS) as a malaria vector control. South Africa endorsed the use of DDT in its Malaria Control Programmes in malaria endemic regions and IRS remains a primary method of controlling malaria transmitting mosquitoes in this country. This study examines the impact of IRS on the levels of DDT and its metabolites in maternal blood of delivering women as a measure of prenatal exposure. We report on the concentrations of DDT and its metabolites (p,p'- DDE, p,p'-DDT, o,p'-DDE, o,p'-DDD, p,p'-DDD, and o,p'-DDT) in maternal plasma of 255 delivering women residing in three sites along Indian Ocean, namely in malaria endemic where IRS takes place, low risk and non-malaria sites. Concentrations of measured compounds were found to be significantly higher in the malaria endemic site (p=0.0001): the geometric mean concentration (95% confidence intervals; n=91) for o,p'-DDE was 9 ng/g lipids (7-10); for p,p'-DDE, 3840 ng/g lipids (3008-4902); for o,p'-DDD, 8 ng/g lipids (6-9); for p,p'-DDD, 26 ng/g lipids (20-32); for o,p'-DDT, 168 ng/g lipids (127-221) and for p,p'-DDT, 2194 ng/g lipids (1706-2823). These compounds were also detected in women residing in other sites but in lower concentrations. The maternal characteristics, age, IRS, number of children and breastfeeding were significantly associated for both p,p'-DDE and p,p'-DDT levels in the malaria area where exposure through IRS is predominant. There was no association between maternal characteristics and DDT levels in the low risk and non-malaria area. Results presented are of particular value to the policy decision makers and regulatory toxicology organizations as they characterise the extent of controlled exposure to DDT used exclusively for IRS purposes. Furthermore, findings of this study will form a base for further investigation of foetal exposure to pollutants.

Levels of selected persistent organic pollutants in blood from delivering women in seven selected areas of São Paulo State, Brazil.

Persistent organic pollutants (POPs) present in the living environment are thought to have detrimental health effects on the population, with pregnant women and the developing foetus being at highest risk. We report on the levels of selected POPs in maternal blood of 155 delivering women residing in seven regions within the São Paulo State, Brazil. The following selected POPs were measured in the maternal whole blood: 12 polychlorinated biphenyls (PCBs) congeners (IUPAC Nos. 99, 101, 118, 138, 153, 156, 163, 170, 180, 183, 187, 194); dichlordiphenyltrichloroethane p,p'-DDT, diphenyldichloroethylene p,p'-DDE and other pesticides such as hexachlorocyclohexanes (α-HCH, ß-HCH, γ-HCH), hexachlorobenzene (HCB), chlordane derivatives cis-chlordane, trans-chlordane, oxy-chlordane, cis-nonachlor and trans-nonachlor. Statistical comparisons between regions were performed only on compounds having concentrations above LOD in 70% of the samples. PCB118 congener was found to be highest in the industrial site (mean 4.97 ng/g lipids); PCB138 congener concentration was highest in the Urban 3 site (mean 4.27 ng/g lipids) and congener PCB153 was highest in the industrial and Urban 3 sites with mean concentration of 7.2 ng/g lipids and 5.89ng/g lipids respectively. Large differences in levels of p,p'-DDE between regions were observed with the Urban 3 and industrial sites having the highest concentrations of 645 ng/g lipids and 417 ng/g lipids, respectively; ß-HCH was found to be highest in the Rural 1 site; the γ-HCH in Rural 1 and industrial; the HCB in the Rural 1 and industrial sites and oxy-chlordane and t-NC in the Rural 2 sites. An association between levels of some contaminants and maternal age and parity was also found.

Examining the association between blood manganese and lead levels in schoolchildren in four selected regions of South Africa.

This study assessed the possibility of relationship between manganese and lead levels in blood in schoolchildren residing in different geographical regions of South Africa. A cross-sectional survey was carried out in schools of three cities: Cape Town (11 schools), Johannesburg (10 schools), Kimberley (six schools) and in the Northern Cape in three rural sites (four schools). A total of 1282 venous blood samples were collected from grade one children. The relationships between blood manganese and blood lead levels (treating each in turn as the response variable in order to adjust for the effect of confounding variables) were investigated by fitting mixed models. Mixed models were fitted with natural log (manganese concentration) as the response variable, and blood lead level as the principal explanatory factor. The model also included terms for centre and a centre by lead interaction and examined potential confounders. The important confounders were found to be gender, race and whether there was paint peeling from the outside walls. There was a significant centre by lead interaction (P<0.0001) with the effect of lead on ln(Mn) being different in the various centres. Mixed models fitted with blood lead level as the response variable and blood manganese as the principal explanatory factor, with terms for centre, a centre by manganese interaction and potential confounders, again found overwhelming evidence (P<0.0001) of a centre by manganese interaction. The study found great variability in both blood lead and manganese levels, both within and between sites, and there was not a consistent relationship between the two metals in the four sites.