Biomarkers of environmental manganese exposure

Abstract

We conducted a critical review on biomarkers of environmental manganese (Mn) exposure to answer the following questions: 1) are there reliable biomarkers of internal Mn exposure (Mn in biological matrices) associated with external metrics of Mn exposure (Mn in environmental media)? and 2) are there accurate reference values (RVs) for Mn in biological matrices? Three bibliographic databases were searched for relevant references and identified references were screened by two independent reviewers. Of the 6342 unique references identified, 86 articles were retained for data abstraction. Our analysis of currently available evidence suggests that Mn levels in blood and urine are not useful biomarkers of Mn exposure in non-occupational settings. The strength of the association between Mn in environmental media and saliva was variable. Findings regarding the utility of hair Mn as a biomarker of environmental Mn exposure are inconsistent. Measurements of Mn in teeth are technically challenging and findings on Mn in tooth components are scarce. In non-occupationally exposed individuals, bone Mn measurements using in vivo neutron activation analysis (IVNAA) are associated with large uncertainties. Findings suggest that Mn in nails may reflect Mn in environmental media and discriminate between groups of individuals exposed to different environmental Mn levels, although more research is needed. Currently, there is no strong evidence for any biological matrix as a valid biomarker of Mn exposure in non-occupational settings. Because of methodological limitations in studies aimed at derivation of RVs for Mn in biological materials, accurate RVs are scarce.

Keywords:

• Manganese
• biomarker
• reference value
• blood
• plasma
• serum
• hair
• urine
• saliva
• environmental exposure
• fingernails
• toenails
• review

1. Introduction

Although manganese (Mn) is essential for human health, excessive exposure to Mn is associated with neurotoxicity (Dobson et al. 2004). Earlier research focused on health effects in adults exposed to elevated levels of Mn via inhalation in occupational settings (Roels et al. 1987, 1992). In recent years, the focus of epidemiological studies has shifted to health effects of low-level Mn exposures in environmental settings. In particular, neurodevelopmental effects in children exposed to Mn via ambient air and drinking water have been extensively examined (Leonhard et al. 2019).

In this article, we provide a critical review of the literature on biomarkers of environmental Mn exposure. Biomarkers are an important component of the exposure to dose to biological effects continuum shown in Figure 1 and are becoming increasingly important in contemporary risk assessment frameworks (Krewski et al. 2014, 2020). Building on a similar continuum proposed by McClellan (1999), McClellan (2021) further discusses the role of biomarkers within an expanded risk assessment paradigm. When available, sensitive and specific biomarkers of exposure are useful surrogates of external exposure from a specific source: a validated biomarker of exposure can be used in the absence of direct measurements of external exposure. When there are multiple sources and pathways of exposure, a biomarker may also represent total exposure from multiple exposure sources and routes, which is particularly useful in risk assessments seeking to characterize dose-response incorporating all relevant exposure sources and routes. In human biological monitoring, biomarkers can be useful for risk assessment in establishing reference ranges and baseline concentrations of chemicals or metabolites in populations, as well as in tracking temporal trends in exposure levels (Gurusankar et al. 2017). Biomarkers of exposure can also be useful in disease diagnosis through the establishment of the fact of exposure to a particular causative agent, as has been recommended by the Institut de recherche Robert Sauvé en santé et en sécurité du travail (IRRST) in the clinical diagnosis of manganism (Ostiguy et al. 2005). Biomarkers are often expressed as tissue concentrations of the agent of interest, or its metabolites, in body tissues, such as blood, urine, and hair: such tissue concentrations represent a useful measure of internal dose, which can be used for more accurate dose-response analyses, taking into account the absorption, distribution, metabolism, and elimination of the agent of interest (Paustenbach and Galbraith 2006).

Figure 1. The role of exposure biomarkers in biomonitoring within the source to exposure to dose to health effects continuum for human health risk assessment [Reproduced from Gurusankar et al. (2017)].

Although outside the scope of this article, biomarkers of biological effect can be used to represent biological responses along adverse outcome pathways that ultimately lead to an undesired apical outcome, corresponding to a clinically demonstrable adverse health outcome (Biomarkers Definitions Working Group 2001). Farrell et al. (2022) discuss novel dose-response modeling techniques based on categorial regression that can be used to combine data on early biomarkers of effect with data on apical outcomes within a single categorical regression model based on the severity of different biological effects observed along the dose to effects continuum. The reader is referred to a detailed review by Gurusankar et al. (2017) for further discussion of reverse and forward dosimetry methods that can be used to describe linkages between exposure and dose and between dose and effect, respectively.

Also outside the scope of this article are biomarkers of susceptibility, including both biological and genetic markers (FDA-NIH Biomarker Working Group 2020). Validated biomarkers of susceptibility are particularly useful in identifying population groups at elevated risk that would benefit from targeted risk mitigation actions. This article provides a comprehensive review of biomarkers of exposure to environmental sources of Mn, including inhalation, dietary, and dermal exposure routes. These results will serve as a guide to the identification and use of biomarkers for Mn exposure assessment.

Validated biomarkers of Mn exposure, according to Hoet and Roels (2015), are sufficiently specific and sensitive to distinguish exposed from non-exposed subjects, display a dose-related association with the external level of exposure, and demonstrate a dose–response relationship with the development of adverse effects in the central nervous system, the critical target of Mn toxicity. This review focuses on examining relationships between Mn in biological matrices and Mn in environmental media.

The overall objective of this study is to summarize scientific evidence on biomarkers of environmental (non-occupational) Mn exposure. Specifically, the aim of this review was to answer the following questions: 1) are there reliable biomarkers of internal Mn exposure (Mn in biological matrices) associated with external metrics of Mn exposure (Mn in environmental media)? and 2) are there reliable reference values (RVs) for Mn in biological matrices?

2. Methods

2.1. Literature search

We searched three electronic literature databases, Medline, Embase, and Toxline, to identify relevant publications from inception to the search date. All searches were conducted on 4 June 2019, and updated on 19 January 2022. As the Toxline content was integrated into PubMed in the end of 2019, the search update was conducted in PubMed. A customized search strategy was developed for each database using combinations of MESH terms and keywords for “manganese” and “biomarkers”, as well as additional terms to exclude animal and in vitro studies. Detailed search strategies are described in Supplemental Material 1. References captured by the searches were imported into an EndNote database and duplicates were removed.

2.2. Eligibility criteria and study selection

References identified through the database searches and other sources were subject to Level 1 (title and abstract) and Level 2 (full text) screening based on predefined eligibility criteria. DistillerSR software (Evidence Partners, Ottawa, Canada) was used for screening by two reviewers, and discrepancies were resolved by consensus.

Studies eligible for inclusion had to be peer-reviewed primary studies published in English. Gray literature sources, news articles, literature reviews, conference abstracts, and editorials were excluded. There were no exclusions based on publication date. Studies had to have examined populations not occupationally exposed to Mn to be eligible. Populations including workers, those with liver diseases, iron deficiency, thyroid diseases, and other diseases were excluded. For inclusion, studies had to have examined exposure to Mn by inhalation, oral, or unspecified route. Studies assessing exposures to Mn from parenteral administration, pharmaceuticals, nutritional supplements, were excluded because these intentional exposures were not environmental in nature. Included studies assessed biomarkers of exposure in blood, urine, saliva, nails, hair, teeth, or bone. Studies examining magnetic resonance imaging (MRI) and other imaging as markers of exposure were not included because they were reviewed in an accompanying article. The eligibility criteria are summarized in Supplemental Material 1.

2.3. Data abstraction

Custom data abstraction forms were developed and then piloted. Collected information was qualitatively discussed and presented in detailed evidence tables (found in Supplemental Material 2). Only primary studies were eligible for inclusion, although systematic reviews were helpful for interpretation of data collected from the primary studies. Data from eligible studies were abstracted by one reviewer.

3. Results

The electronic database searches identified a total of 8892 references. Four additional references were identified from reference lists. Following the removal of duplicates, 6342 references were retained and were subject to Level 1 screening (title and abstract). A total of 481 potentially relevant articles progressed to Level 2 (full text) screening, of which 86 articles met the eligibility criteria (Figure 2).

Figure 2. Study flow diagram. [Adapted from Moher et al. (2009)].

Most articles included in this review reported on Mn in whole blood (N = 33) and hair (N = 31). The remaining articles reported on Mn in urine (N = 7), red blood cells (N = 1), saliva (N = 3), toenails (N = 7), fingernails (N = 4), tooth components (N = 4), and bone (N = 1).

Evidence for Mn in different biological materials as potential markers of environmental Mn exposure is summarized in Table 1. More details are provided in sections below and in the Supplemental Materials.

Table 1. Summary of evidence for Mn in biological materials as markers of environmental Mn exposure.

Biological material	Number of studies	Evidence for association with Mn in environmental media	Comments
Mn in whole blood	33	Little or no evidence	–
Mn in serum or plasma	0	No studies identified	–
Mn in erythrocytes	1	Insufficient	–
Mn in saliva	3	Insufficient/inconsistent	Significant but weak associations with Mn in some environmental media but not in others
Mn in urine	7	Little or no evidence
Mn in hair	31	Inconsistent	Further studies that follow a harmonized protocol are needed.
Mn in toenails	7	Some evidence for positive association	Further studies that follow a harmonized protocol are needed.
Mn in fingernails	4	Some evidence for positive association	Further studies that follow a harmonized protocol are needed.
Mn in tooth components	2 (described in 4 articles)	Insufficient	Technically challenging measurements
Mn in bone	1	Insufficient	Large uncertainties in measurements

3.1. Mn in whole blood as a biomarker of Mn exposure

3.1.1. Mn in whole blood and Mn in environmental media

Thirty-three articles eligible for this review reported on blood Mn as a biomarker of Mn exposure in various settings. These include 13 articles reporting on studies set in areas of active and/or historic ferromanganese alloy plants in Ohio, USA (Haynes et al. 2010; Kim et al. 2011; Bowler et al. 2012; Haynes et al. 2012; Rugless et al. 2014), in Brescia, Italy (Lucchini et al. 2012; Rentschler et al. 2012; Lucchini et al. 2014; Lucas et al. 2015; Butler et al. 2019), in Bahia, Brazil (Rodrigues, Araujo, et al. 2018), and in Cantabria, Spain (Ruiz-Azcona et al. 2021, 2022); seven articles reporting on studies conducted in Mn mining in Hidalgo, Mexico (Santos-Burgoa et al. 2001; Rodriguez-Agudelo et al. 2006; Solis-Vivanco et al. 2009; Riojas-Rodriguez et al. 2010; Montes et al. 2011; Torres-Agustin et al. 2013; Cortez-Lugo et al. 2018); one study in area of non-Mn mining in Oklahoma, USA (Zota et al. 2016); five articles describing studies of Mn in drinking water in Bangladesh (Wasserman et al. 2006; Khan et al. 2011; Parvez et al. 2011; Wasserman et al. 2011; Rodrigues et al. 2015) and one in Greece (Kondakis et al. 1989); and six studies of Mn exposure from other or unspecified sources in Australia (Gulson et al. 2006; Callan et al. 2013; Gulson et al. 2014), South Africa (Rollin et al. 2005), Canada (Bolte et al. 2004), and in Brazil (Nascimento et al. 2016).

Most studies reviewed suggest that, in non-occupational settings, blood Mn levels do not reflect external Mn exposures (see Supplemental Material 2, Table S1). These studies found no difference in blood Mn among groups with different levels of Mn exposure, and/or no positive associations between Mn in blood and Mn in environmental media including indoor and outdoor air, soil, and water. Moreover, three studies reported an inverse association of blood Mn levels in elderly subjects with Mn levels in soil (Rentschler et al. 2012), blood Mn levels in children with Mn in outdoor dust (Lucas et al. 2015), and blood Mn levels in infants and Mn in tap water (Zota et al. 2016). Positive associations with Mn in environmental media were reported; most were either weak (despite being statistically significant) or lost statistical significance after inclusion of covariates in the statistical models. Three articles describing the same population of approximately 170 children in Hidalgo, Mexico, reported that blood Mn was significantly higher among children residing close to Mn mining activities compared to controls (Riojas-Rodriguez et al. 2010; Montes et al. 2011; Torres-Agustin et al. 2013). Prior investigations among 288 adults in the same region of Mexico demonstrated a significant positive correlation (r = 0.22, p < 0.05) between blood Mn and Mn in 24 h stationary air samples of particles < 10 µm (Rodriguez-Agudelo et al. 2006; Solis-Vivanco et al. 2009). The model that best explained variations in blood Mn (R² = 0.08) included air Mn, age, and gender.

Cortez-Lugo et al. (2018) demonstrated that, after implementation of the Environmental Management Program (EMP) in three mining communities of Hidalgo, blood Mn in adult subjects significantly decreased. However, a significant decrease in blood Mn also occurred in a non-mining community, where no interventions were implemented over the same time period. Rodrigues et al. (2015) reported a weak but statistically significant Spearman correlation coefficient (rs = 0.13) between cord blood Mn and Mn in drinking water collected one-month post-partum in Bangladesh. Levels of Mn in drinking water and in blood samples collected from rural children after application of pesticides were significantly higher than water Mn and blood Mn levels of urban children; blood Mn was significantly correlated with drinking water Mn (r = 0.343, p < 0.01) (Nascimento et al. 2016). Gulson et al. (2006) collected blood samples from 113 Australian children and analyzed relationships between Mn in blood and several metrics of external Mn exposure, specifically Mn in handwipes (before and after playing outdoors), house dust accumulation (measured by the petri dish method), dust sweepings, soil, and diet. The only significant predictor for blood Mn was Mn in handwipes prior to playing outdoors. Gulson et al. (2014) reported a significant Pearson correlation (r = 0.190) of blood Mn with estimated dietary intake of Mn, as well as with Mn in handwipes (r = 0.262) in 108 Australian children. Reported associations lost statistical significance after inclusion of other potential predictors of blood Mn (child’s age, season of blood sample collection) into the statistical model. Rollin et al. (2005) studied associations between blood Mn and Mn in environmental samples collected from schools in a relatively larger study of 814 South African children. No significant associations were detected.

3.1.2. Summary and conclusion: blood Mn as a biomarker of Mn exposure

Most studies identified for this review suggest that, in non-occupational settings, blood Mn levels do not reflect Mn in environmental media. This conclusion applies to the literature that was reviewed, which is limited to environmental exposure studies in human populations, as opposed to occupational exposure studies or animal studies involving controlled, inhaled Mn exposures. Baker et al. (2014) summarized the relationships between Mn in air and blood Mn across worker studies and found that there may be a threshold below which inhalation exposure to Mn is not reflected in blood Mn. This threshold may be at air Mn level of about 10 µg/m³. The authors of this review noted that typical environmental exposures to Mn were well below this threshold. Studies that form the basis for the pharmacokinetic models for Mn indicate that blood is a reasonable biological indicator for inhalation exposures under controlled experimental conditions (Ramoju et al. 2017).

The main source of Mn for the general population is the diet (Rodrigues, Araujo, et al. 2018). Mn levels in blood are tightly controlled by homeostatic mechanisms regulating the intestinal absorption of Mn and its biliary excretion. The levels of Mn in whole blood are maintained in a relatively narrow concentration range (Butler et al. 2019) and are, therefore, not expected to be sensitive markers of external Mn exposure (Hoet and Roels 2015).

The relationship between Mn exposure and blood Mn is complicated by inter-individual variability of blood Mn due to differences in diet, iron status, and genes involved in iron metabolism (Haynes et al. 2012).

Few studies reported on steps taken to prevent external contamination of blood samples (e.g. use of metal-free tubes for blood sample collection). Overall, it appears that blood Mn is not a good biomarker of Mn exposure in environmental settings.

3.2. Mn in serum or plasma as a biomarker of Mn exposure

None of the studies eligible for inclusion assessed Mn in serum or plasma in relation to Mn in environmental media.

3.3. Mn in erythrocytes as a biomarker of Mn exposure

One study examining Mn in erythrocytes was identified. This study in 408 pregnant Bangladeshi women exposed to high levels of Mn in tube well water (10–6336, mean 720 µg/L) found no correlation between Mn in well-water and Mn in erythrocytes (Ljung et al. 2009).

3.4. Mn in saliva as a biomarker of Mn exposure

3.4.1. Mn in saliva and Mn in environmental media

Three studies included in this review examined associations between Mn in children’s saliva and Mn in environmental media (Supplemental Material 2, Table S2).

Butler et al. (2019) found a significant positive correlation (rs = 0.23) and a positive association in regression analysis between Mn levels in children’s saliva and Mn in the PM₁₀ fraction of 24 h personal air samples among 717 children in Brescia, Italy. No correlations were reported between saliva Mn levels and Mn in soil nearby the children’s homes or Mn in indoor or outdoor dust. In a weighted quantile sum regression (WQS), an approach to estimate the cumulative impact of all environmental media, increased weighted environmental exposure index was significantly associated with increased saliva Mn. Air Mn accounted for 65% and soil Mn accounted for 29% of the weight.

Lucas et al. (2015) enrolled 249 children from three locations: two locations near historic (n = 71 children) or active (n = 96 children) ferromanganese alloy plants in Brescia, Italy, and a location with no history of plant activity (n = 82 children). Saliva Mn concentrations were significantly lower in the site with no plant activity compared to the other two sites. Statistically significant weak correlations were reported between saliva Mn concentrations and Mn concentrations in outdoor dust (rs = 0.15), soil (rs = 0.14), and in PM₁₀ fraction of 24-h personal air samples (rs = 0.20). There was no correlation with Mn concentrations in indoor dust, indoor dust Mn load, or outdoor dust Mn load. In both studies, saliva was centrifuged before analysis.

Ntihabose et al. (2018) measured Mn in whole saliva and saliva supernatant in a study of 274 children living in New Brunswick, Canada, who were exposed to relatively low levels of Mn in drinking water. Mn in saliva supernatant was weakly but statistically significantly correlated with Mn in water (rs = 0.14) and with estimated Mn intake from water (rs = 0.18) in all children. Similar correlations were also reported among boys (rs = 0.19 for both relationships) but not among girls. When examining Mn levels in whole saliva, correlations with Mn levels in water or with estimated Mn intake from water were not significant.

3.4.2. Summary and conclusion: Mn in saliva as a biomarker of Mn exposure

The limited available information does not allow a conclusion regarding the performance of saliva Mn as a marker of environmental Mn exposure. Only three studies, all in children, analyzed associations between Mn in saliva and Mn in environmental media. Although some significant correlations/associations were reported, most were weak. Mn in whole saliva and Mn in supernatant may exhibit somewhat different patterns of association with environmental Mn. Wang et al. (2008) demonstrated that, in welders, saliva Mn significantly correlated with serum Mn. However, the exposure-associated change in serum Mn (72% increase vs. controls) was larger than that in saliva Mn (46%). These findings, in the authors’ view, indicate that changes in saliva Mn mirrored those in serum Mn; therefore, Mn in saliva may not be a more sensitive indication of Mn exposure than Mn in serum. The advantage of Mn in saliva is the noninvasive nature of this bioindicator.

3.5. Mn in urine as a biomarker of Mn exposure

3.5.1. Mn in urine and Mn in environmental media

Seven articles describing studies of Mn in urine are included in this review. These studies were conducted among Italian children and elderly subjects living in the area affected by historic or active ferromanganese alloy plants (Lucchini et al. 2012; Rentschler et al. 2012; Lucchini et al. 2014; Butler et al. 2019), among pregnant Bangladeshi women exposed to Mn from drinking water (Ljung et al. 2009), Australian nonsmoking pregnant women (Callan et al. 2013), and Korean adults (Choi and Bae 2020) exposed to Mn from unspecified sources (Supplemental Material 2, Table S3).

Ljung et al. (2009) found a weak positive correlation (rs = 0.19) between Mn in drinking water and Mn in urine of pregnant Bangladeshi women. A similar weak correlation was reported between urine Mn and soil Mn (rs = 0.14) but not with Mn in air, indoor dust, or outdoor dust in Italian children living near ferromanganese alloy plants (Butler et al. 2019). However, no associations between urine Mn and Mn in any environmental media were found in linear regression analyses. Other studies provide no evidence for a relationship between Mn in environmental media and Mn in urine in various populations (children, elderly, and pregnant women) (Rentschler et al. 2012; Callan et al. 2013) and do not suggest that urine Mn may distinguish between groups of subjects exposed to Mn at different levels (Lucchini et al. 2012, 2014). Choi and Bae (2020) found no correlation between daily Mn excretion in urine and daily Mn intake from diet among adults in Korea.

In summary, included studies suggest that urinary Mn does not reflect Mn in environmental media.

3.5.2. Summary and conclusion: Mn in urine as a biomarker of Mn exposure

Mn in urine was not frequently used as a marker of Mn exposure in the reviewed epidemiological studies. Most studies measure Mn in spot urine samples or in first-morning void samples. Measurement of chemicals in urine collected over 24 h is proposed as the gold standard to evaluate daily exposures (Chen et al. 2019). Chen et al. (2019) demonstrated that Mn measurements in spot samples and in first-morning void urine are not good predictors of Mn in 24-h samples. Moreover, serial measurements of Mn in spot samples over three months showed poor reproducibility. Reproducibility over time was better for Mn measured in first-morning void and in 24-h urine samples. However, even for these measurements, within-individual variance was larger than the between individual variance (Chen et al. 2019).

It is believed that Mn in urine may reflect recent short-term exposures (Leonhard et al. 2019). However, because more than 95% of Mn is excreted via the bile to feces (Aschner et al. 2007), Mn concentrations in urine are not expected to be sensitive markers of external Mn exposure (Zheng et al. 2011; Hoet and Roels 2015). Some researchers recommend abandoning Mn in urine as a biomarker (Zheng et al. 2011). The few studies eligible for inclusion in this review were generally in line with this conclusion. These studies suggest that urinary Mn concentrations do not distinguish between groups of individuals exposed to different Mn levels and do not reflect Mn concentrations in environmental media. An additional drawback is that urine is inconvenient to collect compared to other biological materials, such as hair (Standridge et al. 2008).

3.6. Mn in hair as a biomarker of Mn exposure

3.6.1. Mn in hair and Mn in environmental media

Data extracted from 31 articles examining associations between Mn levels in hair with Mn in environmental media are summarized in Supplemental Material 2 (Table S4).

3.6.1.1. Mn in drinking water

Nine studies, one of them described in two articles, examined hair Mn in relation to Mn levels in drinking water.

A pilot study conducted in Quebec, Canada (Bouchard M. et al. 2007) included 46 children living in a community with high Mn levels in the public water system. Hair Mn levels were significantly higher in children from houses connected to a well with mean water Mn level 500 µg/L, compared to children from households connected to a well with mean water Mn 160 µg/L. In multiple regression analyses, the supplying well was a significant determinant of hair Mn. A major limitation of this study is that hair samples were not washed before chemical analyses.

A later study conducted by Bouchard M. et al. (2011) included a larger population of children (n = 362) from municipalities in Quebec supplied by groundwater. Unadjusted analysis demonstrated an increase in hair Mn with increasing water Mn. In multiple regression analyses, hair Mn was significantly associated with estimated Mn intake from water consumption; there was no association between hair Mn and estimated dietary intake of Mn. A more recent study in 274 children from New Brunswick, Canada (Bouchard M. et al. 2018; Ntihabose et al. 2018) also demonstrated significant correlation of hair Mn with water Mn and with estimated Mn intake from water. Levels of Mn in water were lower in this area (mean = 62.1 µg/L) compared to those reported by Bouchard et al. (2011) (mean = 98 µg/L). In another study, Ntihabose et al. (2018) classified children into five approximately equal-sized groups corresponding to the quintiles of water Mn concentrations or to Mn intake from water. Hair Mn levels were higher in higher quintiles of water Mn concentrations or Mn intake.

A small pilot study that included 25 mother–infant pairs from Minnesota (USA) found no significant correlation between Mn in tap water (ranging from <10 to 290 µg/L) and Mn in 7-month-old infants’ hair (Cigan et al. 2018). Water Mn, however, was significantly correlated with Mn in maternal hair. A larger study conducted in Bangladesh included 302 mother-infant pairs exposed to higher levels of Mn in drinking water (mean ∼700 µg/L) (Rodrigues et al. 2015). Results of this study are in line with those reported by Cigan et al. (2018). Specifically, Mn in one-month-old infants’ hair was not correlated with water Mn (correlation coefficient 0.04). A weak but statistically significant correlation (r = 0.15) between water Mn and Mn in maternal hair was found.

In covariate-adjusted analyses, Zota et al. (2016) found no significant association between Mn in infant hair and relatively low levels of Mn in tap water. Ninety-fifth percentile of Mn in water collected before the first-morning use was 13.2 and 16.7 µg/L measured when the participants were 0–6 and 6–12-month old, respectively. Ninety-fifth percentile of Mn in water collected after running the tap for 3 min was ∼7 µg/L at both sampling sessions.

Skroder et al. (2017) collected hair from 207 Bangladeshi children exposed to high levels of water Mn (median = 410 µg/L) and found no correlation between the concentrations of Mn in hair and Mn in water. This study demonstrated that the concentration of Mn in hair depends on the distance of hair sampling from the scalp, as the geometric mean hair Mn concentration increased almost five times from 6.9 µg/g in the 2 cm of hair closest to the scalp to 32 µg/g in hair 7–8 cm from the scalp. This increase, in the authors’ view, reflects a longer exposure of distant hair to Mn from external sources and a greater extent of Mn binding from external sources over time (Skroder et al. 2017).

Kondakis et al. (1989) compared hair Mn concentrations in older individuals (50 years or older) living in three areas in Greece with different water Mn levels recorded between 1981 and 1986. Difference in hair Mn between the areas with highest and lowest Mn levels was statistically significant for both sexes combined, as well as for males and females separately.

Kousa et al. (2021) collected samples of hair and household well water from 92 adult residents of two municipalities in Finland, Sotkamo, and Kaavi. Mean concentrations of Mn in private well water were 85.59 µg/L in Sotkamo and 29.40 µg/L in Kaavi. Mean hair Mn levels were 3.19 µg/g in Sotkamo and 1.68 µg/g in Kaavi. The study areas were divided into black shale areas representing sulfide- and graphite-rich metasedimentary rocks, and reference areas representing other rock types with low concentrations of sulfur and graphitic carbon. There were no statistically significant differences between Mn concentrations in water or in hair between the black shale and reference areas. Hair Mn levels positively correlated with Mn levels in water, with Spearman correlation coefficients 0.569 (p < 0.01) in Sotkamo black shale area, 0.450 (p > 0.05) in Sotkamo reference area, 0.452 (p > 0.05) in Kaavi black shale area, and 0.856 (p < 0.01) in Kaavi reference area. Cleaning of hair samples before the chemical analysis is not described, which is a major limitation of reporting of this study.

In summary, data regarding the performance of hair Mn as an indicator of exposure to Mn from drinking water were inconsistent. Seven of the nine studies found at least some evidence of an association between Mn in hair and Mn in drinking water. Two studies suggest that Mn in hair may be useful for discriminating between high- and low-exposed subjects on a group level. Such a discriminating capacity was demonstrated in studies among children (Bouchard M. et al. 2007) and elderly individuals (Kondakis et al. 1989). Studies in children in Canada showed a significant positive correlation of hair Mn with levels of Mn in water and with estimated Mn intake from water even at relatively low levels of water Mn (Bouchard M. et al. 2018; Ntihabose et al. 2018). In contrast, two studies failed to detect such an association at low levels of water Mn in a covariate-adjusted analysis (Zota et al. 2016) or at high levels of Mn in water in a correlation analysis (Skroder et al. 2017). Two studies in mother–infant pairs which involved varying levels of Mn in drinking water suggest that levels of Mn in maternal hair – but not in infants’ hair – may be associated with concentrations of Mn in drinking water (Rodrigues et al. 2015; Cigan et al. 2018).

3.6.1.2. Mn emitted from ferromanganese facilities

Twelve studies (one study described in two articles) examined hair Mn as a biomarker of Mn exposure in individuals living in the vicinity of ferromanganese facilities.

Three studies with overlapping populations were conducted in Brescia, Italy, and included 10–14-year-old children living close to ferromanganese alloy plants (Lucchini et al. 2012; Lucas et al. 2015; Butler et al. 2019). Hair Mn levels were compared to those from controls living in areas with no history of ferromanganese activity. There were no significant differences in hair Mn across study sites (Lucchini et al. 2012; Lucas et al. 2015). Hair Mn levels were compared with levels in environmental samples. Significant correlations were reported with Mn in PM₁₀ fraction of 24 h personal air samples (median air Mn level 26.02 µg/m³) (Butler et al. 2019), with Mn concentrations in indoor dust (Lucas et al. 2015; Butler et al. 2019), and with Mn concentrations in outdoor dust and outdoor dust Mn loading levels (Lucas et al. 2015). Lucas et al. (2015) and Butler et al. (2019) analyzed Mn in soils and did not report any correlations with Mn in hair. Hair Mn was not associated with Mn in any of the environmental samples when using multiple regression analysis adjusted for Mn in all other environmental samples and for child’s sex and age. However, using WQS, an approach to estimate the cumulative impact of all environmental media, an increase in the weighted environmental exposure index was associated with increased hair Mn (Butler et al. 2019).

Lucchini et al. (2019) found a significant decrease in levels of Mn in 6–-12-year-old children’s hair with increasing residential distance from iron- and steel-making operations in Taranto, Italy. The median hair Mn level measured in hair was 135.3 ng/g (0.1353 µg/g). Menezes-Filho et al. (2009) examined hair Mn in 1–10-year-old children in relation to Mn in PM_2.5 samples (0.011–0.439, mean 0.15 µg/m³) from four different areas around a ferro-manganese alloy production plant and near a school and daycare center in Bahia, Brazil. Hair Mn levels were found to be the highest in areas of residences closest to the plant and those directly downwind from the plant. In multiple regression analyses including only data for exposed children, area of residence was a significant variable. Two other studies among older children (7–12 years) residing in the same area demonstrated that hair Mn could discriminate between children attending a school with the lowest Mn dust deposition rates from children in schools with higher Mn dust deposition rates (Rodrigues, Araujo, et al. 2018; Rodrigues, Bandeira, et al. 2018). In correlation analyses and in multiple regression analyses, hair Mn was significantly associated with interior and exterior Mn dust loading rates.

Ruiz-Azcona et al. (2021) and Ruiz-Azcona et al. (2022) collected hair samples from 130 adults residing in the vicinity of a ferromanganese allow plant in Cantabria, Spain. Mn levels measured in PM2.5 (0.00079–0.937, mean 0.079 µg/m³) were somewhat lower than those reported by Menezes-Filho et al. (2009) in the Brazilian study. Mn in PM10 ranged from 0.042 to 3.145 (mean 0.152) µg/m³. Hair Mn levels were higher for participants living at a shorter distance (≤1.5 km) from the industrial emission source (mean 0.530 and median 0.235 µg/g), compared to those living at a greater distance (>1.5–10 km) from the source (mean 0.230 and median 0.133 µg/g); only the difference between the medians reached statistical significance (p = 0.005). Mn in hair was negatively correlated with the distance from the emission source (Spearman’s r = −0.17, p = 0.059).

In a pilot study, Haynes et al. (2010) measured hair Mn concentrations in individuals of all ages residing in the Marietta area in Ohio (USA), a location of a ferromanganese refinery. Hair Mn was significantly correlated with air Mn, estimated by an air dispersion model (annual average concentrations 0.01–18.13 µg/m³; mean 0.13 µg/m³) only when the model included genes for iron metabolism. In a later study, Haynes et al. (2012) measured Mn in hair of 38 children from nonsmoking households located in the same area. In a multiple regression analysis, hair Mn was not associated with Mn in PM_2.5 fraction of stationary air samples (range 0.00453–0.03454 µg/m³; geometric mean 0.011 µg/m³), in PM_2.5 fraction of personal air samples (0.00147–0.05448 µg/m³; geometric mean 0.0081 µg/m³), or with time-weighted distance (TWD) from the refinery. The authors note the small sample size and that participants’ residences were not near the point source of Mn emissions as limitations of their study. Rugless et al. (2014) found no correlation between levels of Mn in hair and time weighted distance (TWD) from the Marietta refinery in a somewhat larger sample of children (n = 55). The authors point out, however, that their measure of exposure, TWD, did not account for wind direction and wind speed, which were significant predictors of personal air Mn in the Marietta area.

Using a structural equation modeling (SEM) approach, Fulk et al. (2017) evaluated a theoretical inhalation exposure pathway for children’s exposure to Mn in the Marietta area and found Mn household dust as significant contributor to hair Mn. Significant contributors to household dust Mn were modeled annual air Mn concentrations, Mn concentrations in soil (collected in the yard), and time spent outside. These results, in the authors view, suggest that significant pathways include direct exposure to Mn through inhalation of household dust and indirect exposure to Mn in ambient air through Mn deposited in house dust.

In summary, it is not clear whether hair Mn may be useful for discriminating between groups of children with different levels of environmental exposure to Mn from ferromanganese activities. Studies in Brazil demonstrated such a discriminating capacity of hair Mn while studies in Italy that presumably involved comparable levels of Mn exposure, did not. A study in Spain with somewhat lower levels of Mn in PM2.5 than those reported in the Brazilian study, found higher levels of hair Mn in those living at 1.5 km or less from the ferromanganese alloy plant compared to those living at greater distances. Hair Mn was negatively correlated with the distance from the emission site, although the correlation was weak. Hair Mn in the Italian and Brazilian children was positively correlated with Mn in dust and Mn in dust loading levels. Whereas in the Italian study, the association was no longer significant after adjustment for covariates, it remained significant in the Brazilian study. The Italian study demonstrated that hair Mn and Mn in personal air samples were weakly but significantly correlated. However, no significant association between Mn in hair and Mn in personal air was detected in multiple regression analyses performed within this study, as well as in a smaller study of children exposed to presumably lower levels of Mn from a ferromanganese refinery in Ohio. The SEM approach used in the Ohio study suggests that Mn in household dust contributes to Mn in hair directly, whereas the contributions of Mn in ambient air and Mn soil to hair Mn are indirect, through deposition of Mn from these sources into household dust.

3.6.1.3. Mn from mining activities

Six articles report on hair Mn as a biomarker of Mn exposure in areas of active or historical Mn or other metal mining activities. Of them, three articles (Riojas-Rodriguez et al. 2010; Montes et al. 2011; Torres-Agustin et al. 2013) describe findings from essentially the same population of 7–11-year-old children (n = 170–174) from Hidalgo (Mexico). Hernandez-Bonilla et al. (2016) describe a larger population of children (n = 267) of the same age from the same area. Two studies (Zota et al. 2016; Semenova et al. 2018) were conducted among children living in areas of non-Mn mining.

Hair Mn levels in children living in the Molango Mn-mining area in Hidalgo, Mexico, were significantly higher than in children living in the non-mining area (Riojas-Rodriguez et al. 2010; Montes et al. 2011; Torres-Agustin et al. 2013; Hernandez-Bonilla et al. 2016). Semenova et al. (2018) conducted a study among children living in one of three rural settlements near abandoned nonferrous metal mines in Bashkortostan, Russia. Hair Mn concentrations in residents of Tubinsk were higher than those in two other settlements; this pattern was consistent with that seen for Mn in soils. However, the differences in hair Mn between participants from the three areas were not significant and there were no significant correlations between Mn in soil and Mn in hair. (Soil samples were taken from four sites at each settlement, and it is unclear whether the sampling sites were near the children’s homes.) Zota et al. (2016) collected hair samples from 12-month-old infants living near a site impacted by Pb and Zn mining. Environmental samples were collected in two sessions: when the infants were 0–6 months old and when they were 6–12 months old. Mn levels in biological and residential samples were similar to those of the general population. A positive association was found between hair Mn and Mn concentration in house dust and Mn in house dust loadings but not with Mn in other environmental samples (soil, indoor PM_2.5, and tap water).

In summary, four studies of the same or overlapping populations of children living in an area impacted by Mn mining activities suggest that hair Mn may be useful for discriminating exposed from unexposed subjects on a group level. Two studies among children residing in areas potentially impacted by non-Mn mining activities provide ambiguous results.

3.6.1.4. Mn from other/unspecified sources

Mn in hair collected from 498 residents of New York City (all ages from 0 to 51 + years) was not associated with Mn in dust fall (data from air monitoring sites) or with Mn in house dust and Mn in soil collected from participants’ residences (Creason et al. 1975).

After a period of pesticides application, Nascimento et al. (2016) collected hair samples from 43 children living in a rural area and from 20 children living in an urban area. Tap water samples were collected from the kitchen of each child’s household. Levels of Mn in tap water from the rural area (mean 18.5 µg/L) were significantly (p < 0.001) higher than those in tap water from the urban area (mean 1.3 µg/L). Likewise, Mn concentrations in hair of the rural children (mean 2.17 µg/g) were significantly higher (p < 0.001) than those in hair of the urban children (mean 0.47 µg/g). Hair Mn was significantly correlated with drinking water Mn (r = 0.434, p < 0.01).

Soetrisno and Delgado-Saborit (2020) measured Mn in hair of 19 children residing near e-waste recycling and in 22 referent children residing far from any e-waste recycling activities. Samples of soil and tap water were collected from five households near the e-waste recycling and from five households in the reference area. Levels of Mn in soils (p = 0.008) and in water (p = 0.095) of the e-waste recycling area were higher than those in the reference area. Hair Mn levels were also higher in the “exposed” children (mean 130 µg/g) than in the referent children (mean 18 µg/g), although the difference was not statistically significant (p = 0.132). The authors of this study did not report whether hair samples were washed before chemical analysis.

3.6.1.5. Summary: association between Mn in hair and Mn in environmental media

Most studies reviewed in this section were conducted among children. These studies do not provide a consistent pattern of performance of hair Mn as a marker of Mn exposure. Some studies show that hair Mn may help discriminate, on a group level, exposed from unexposed subjects, or highly exposed from those exposed to lower environmental levels of Mn. Other studies do not show such a discriminating capacity of hair Mn. There is no consistency regarding the existence of associations between hair Mn and Mn in environmental media. For example, some studies reported a positive association while others did not. Other studies reported significant correlations, but associations were not found in analyses adjusted for covariates. Lack of correlation of hair Mn with Mn in some environmental media (such as air and soil) may be because Mn in these media contribute to hair Mn indirectly, through other media (such as household dust).

3.6.2. Summary and conclusion: Mn in hair as a biomarker of Mn exposure

Using metal concentrations in hair as potential biomarker of exposure offers several advantages for biomonitoring, including high mineralization of samples and irreversibility of incorporation of trace elements into hair matrix (Skalny et al. 2015). Hair sampling is simple, noninvasive and, therefore, does not require a medical clinic. Additionally, samples can be easily stored and transported (Skroder et al. 2017).

At the same time, there are limitations associated with the use of Mn in hair as a marker of exposure. The level of Mn in hair may be influenced by hair treatment (dye, bleach, or other topical treatment), pigmentation (may be higher in darker hair), and growth rate (Coetzee et al. 2016; Cigan et al. 2018). Regarding hair dyeing, it has been reported that the structure and composition of hair depend on whether the dye is temporary, semi-permanent, or permanent (Chojnacka et al. 2010). Also, depending on techniques used for cleaning hair prior to analysis, the extent to which Mn measured in hair reflects Mn body burden and external contamination may vary from study to study. For example, some studies introduced an additional step in the procedure of hair sample cleaning, an ultra-sonic bath in nitric acid solution, whereas most studies did not include this step. There is no consensus on the optimal method of sample preparation. Inadequate washing may not fully eliminate external contamination. On the other hand, excessive washing may damage hair and remove endogenous Mn (Skroder et al. 2017; Ntihabose et al. 2018). Because washing procedures differ from study to study, comparison of hair Mn levels from different studies may be challenging (Chojnacka et al. 2010). External Mn deposited on hair, although considered as a contaminant, is likely to correlate with cumulative ambient Mn exposure (Haynes et al. 2018). If so, hair Mn concentrations may be higher due to external contamination, but the associations would be similar (Menezes-Filho et al. 2014).

As demonstrated by Skroder et al. (2017), distance from the scalp at which the sample was collected is important. Because older parts of hair that are more distant from the scalp are exposed to external Mn from water and air for longer time than the first 1–2 cm closest to the scalp, the extent of binding from external sources over time is greater. The geometric mean Mn concentration increased almost five-fold from the first 2 cm sample to the fourth (most distant) 2 cm sample, supporting that the most distant 2 cm sample does not reflect the internal dose of Mn (Skroder et al. 2017). Most studies included in this review collected hair samples as close as possible to the scalp.

Both the choice of analytical technique and calibration were listed by Dipietro et al. (1989) as “difficulties in interpretation of trace element data for hair”. This difficulty appears to still exist, as reviewed studies used different techniques to measure Mn in hair. In addition, there are no well-defined reference ranges for Mn levels in hair (Llorente Ballesteros et al. 2017).

Studies summarized in this section do not provide a consistent pattern of performance of hair Mn as a marker of external Mn exposure. Several studies showed that hair Mn may help discriminate, on a group level, exposed from unexposed subjects, or highly exposed from those exposed to lower environmental levels of Mn. Other studies did not confirm such a discriminating capacity of hair Mn. Furthermore, findings were inconsistent regarding associations between hair Mn and Mn in environmental media as some studies detected such associations while others did not. Different scenarios of Mn exposure (route, levels, and gradient) may, at least partly, be responsible for these inconsistencies.

In some studies, the association detected in correlation analyses was no longer evident in analysis adjusted for covariates. This may be a result of a lack of a true association or lack of statistical power in adjusted analyses. Covariates most commonly adjusted for were age and sex. Higher levels of Mn in female hair were reported in several studies (Chojnacka et al. 2006; Bouchard M. et al. 2007; Menezes-Filho et al. 2009). Nevertheless, findings from other reviewed studies found no significant sex differences (Riojas-Rodriguez et al. 2010; Menezes-Filho et al. 2014, 2018; Rodrigues, Bandeira, et al. 2018). Ntihabose et al. (2018) found that, although there was no significant difference in hair Mn levels between boys and girls, the variation was greater in girls. In a pilot study, Haynes et al. (2010) found no significant differences in hair Mn by gender in adults or children. With respect to age, Riojas-Rodriguez et al. (2010) report increasing levels of hair Mn with age.

For the general population, the main source of exposure to Mn is through food (ATSDR 2012). As discussed by Leonhard et al. (2019), “estimating Mn exposure based on a single environmental source, such as drinking water or ambient air, may omit a meaningful proportion of total exposure”. Bouchard M. et al. (2011) examined an association between hair Mn and estimated intake of Mn from water and dietary sources. Although the median of estimated dietary manganese intakes was about two orders of magnitude higher than the median intake from water consumption, hair Mn increased with estimated intake of Mn from water, but no association was observed between hair Mn and the estimated dietary Mn intake.

Overall, there is no consensus in the literature regarding the validity of hair Mn as a marker of exposure. Some investigators claim that hair Mn is “a strong, noninvasive indicator for exposure” (Rink et al. 2014) and “the most consistent and valid biomarker for Mn to date for children in population studies” (Coetzee et al. 2016). Others question its utility on grounds that the mechanism by which Mn is incorporated into hair is not clear, and the uncertainty regarding the removal of external contamination (Riojas-Rodriguez et al. 2010). Our analysis of literature suggests that, currently, there is no strong evidence for hair Mn as a valid biomarker of Mn exposure. Standardization of methods for sample collection (e.g. location and length), cleaning, and analytical methods, would solve at least some of the problems associated with the use of hair Mn as a marker of exposure and would facilitate the comparison of findings across studies.

3.7. Mn in toenails as a biomarker of Mn exposure

The growth of toenails is slower than that of fingernails (Ntihabose et al. 2018). Due to their different pattern of growth, the toenails and fingernails are distinct substrates, and may reflect different exposure windows estimated at six months before sample collection for fingernails and 7–12 months prior to sample collection for toenails (Rodrigues, Bandeira, et al. 2018).

3.7.1. Mn in toenails and Mn in environmental media

Seven studies included in this review examined associations between Mn in toenails and various measures of environmental exposure to Mn (Supplemental Material 2, Table S5). Five studies included children, pregnant women, or mother-infant pairs exposed to Mn from drinking water. They were conducted in the USA (Cigan et al. 2018; Signes-Pastor et al. 2019), Canada (Ntihabose et al. 2018) and Bangladesh (Parvez et al. 2011; Rodrigues et al. 2015). Two studies (Rodrigues, Araujo, et al. 2018; Rodrigues, Bandeira, et al. 2018) were conducted in overlapping populations of children exposed to Mn due to residence near a ferromanganese alloy plant in Brazil.

Rodrigues et al. (2015) reported that Mn in maternal toenails collected one month postpartum was weakly but significantly correlated (rs=0.17) with Mn in concurrently collected drinking water samples. Mn in one-month-old infants’ toenails was not correlated with Mn in water. This study was conducted in two areas of Bangladesh with mean water Mn level of 752.3 and 731.5 µg/L. Another study (Parvez et al. 2011) included children living in Bangladesh exposed to similar mean Mn concentrations. The study assessed children’s motor function and Mn in blood as a biomarker. Reporting of findings on toenail Mn is limited and unclear. However, it appears that children exposed to higher Mn levels in drinking water (>500 μg/L) had higher concentrations of Mn in toenails. Although correlation findings were not reported, the study indicates that measures of Mn exposure were not well correlated with each other (r between 0.05 and 0.15) (Supplemental Material 2, Table S5).

A study in New Brunswick, Canada, included children exposed to Mn from water at lower levels than those reported in the Bangladeshi studies (mean = 5.96 µg/L) (Ntihabose et al. 2018). Toenail Mn was significantly correlated with water Mn and with estimated Mn intake from water among both sexes, as well as among girls and boys separately (rs range: 0.28–0.40). Children exposed to the highest water Mn and highest Mn intake quintiles had significantly greater toenail Mn concentrations than children exposed to lower Mn from drinking water. The authors concluded that Mn concentrations in toenails “are promising biomarkers, even at low-level chronic environmental Mn exposure from drinking water in children.”

Signes-Pastor et al. (2019) collected toenail samples from pregnant women at enrollment and at two weeks postpartum in New Hampshire, USA. Levels of Mn in drinking water collected at enrollment were relatively low (median = 2.9 µg/L). A significant positive correlation was found between Mn in drinking water and Mn in toenails collected at the two instances. The association was nonlinear, and the correlations were statistically significant only at water Mn concentrations higher than 9.8 µg/L. The correlations were close to null at water Mn concentrations ≤ 9.8 µg/L.

Cigan et al. (2018) conducted a pilot study in Minnesota, USA, among 25 infant–mother pairs exposed to relatively low Mn levels in drinking water ranging from <10 to 290 µg/L. Reported correlations were positive but not significant between water Mn and Mn in maternal or 7-month-old infants’ toenails. The strengths of the reported correlations are similar to those reported in other studies; hence, the lack of statistical significance may be due to a very small sample size in this pilot study.

Toenail Mn levels in Brazilian children attending schools with higher Mn deposition rates in exterior and interior environments were higher than those in children attending a school with the lower exposure to Mn from a ferromanganese alloy production plant. Toenail Mn was significantly correlated with Mn in exterior dust and with Mn in interior environment (r ∼ 0.5). Further, toenail Mn was significantly positively associated with Mn dust loading rates in multiple regression analyses such that Mn dust loading accounted for 26.5% of the variance in children’s toenail Mn levels (Rodrigues, Bandeira, et al. 2018) (Supplemental Material 2, Table S5).

3.7.2. Summary and conclusion: Mn in toenails as a biomarker of Mn exposure

Most data reviewed in this section suggest that toenail Mn levels reflect Mn levels in the environmental media and are higher in groups of children exposed to higher environmental Mn levels. It should be noted, however, that studies meeting the inclusion criteria for this review are relatively few, and most of them examined correlations. Analyses adjusted for potential confounders were conducted in only one eligible study.

In the VA Normative Aging Study that included older men, Wu et al. (2019) demonstrated a reasonably good intra-individual correlation between toenail Mn levels measured on up to six occasions over a period of about ten years. For subjects who did not take Mn supplements, the correlation coefficients ranged from 0.50 for measurements taken one year apart to 0.40 for measurements taken seven years apart. This, in the authors’ view, suggests that Mn levels in toenails “can reasonably reflect exposures over several years”.

There is a concern regarding possible contamination of toenails from external Mn sources and the inter-individual differences in nail characteristics and personal care habits (Ntihabose et al. 2018; Leonhard et al. 2019). It is thought that toenails are less susceptible to effects of external contamination than hair because, for a given sample weight, toenail clippings have a smaller surface-to-volume ratio than hair strands. Toenails may also be less susceptible to external contamination than fingernails (Ntihabose et al. 2018). Unlike hair, toenails are not altered by cosmetic procedures (dyeing, bleaching, and permanent waving) and are not affected by melanin content (nails are melanin-free) (Goulle et al. 2009).

The procedure of toenail collection is noninvasive, which is an important factor, especially when studying children (Bouchard M. et al. 2018). Ease of sample collection, transportation and storage represent advantages of this biomonitoring method (Ghazali et al. 2013).

Variations in sample cleaning procedures complicate comparisons of toenail Mn levels across studies. For example, the Brazilian studies introduced an additional step in the procedure of toenail cleaning, an ultra-sonic bath in nitric acid solution (Menezes-Filho et al. 2018; Rodrigues, Araujo, et al. 2018; Rodrigues, Bandeira, et al. 2018). The need for homogeneously performed cleaning procedures was acknowledged in a systematic review of toenails as biomarker of exposure to essential trace metals (Gutierrez-Gonzalez et al. 2019). In addition, the absence of certified reference material also represents a challenge associated with the use of this substrate for metal measurements (Gutierrez-Gonzalez et al. 2019).

Overall, toenails have practical advantages over other biological substrates for measurement of Mn as a marker of exposure. The limited available information suggests that Mn in toenails may reflect Mn in environmental media and may discriminate, on a group level, between children exposed to higher and lower environmental Mn. To validate toenail Mn as a marker of environmental exposure, more studies in environmental setting conducted according to a standardized protocol for sample preparation and analysis are needed.

3.8. Mn in fingernails as a biomarker of Mn exposure

3.8.1. Mn in fingernails and Mn in environmental media

Four articles describing three studies of Mn in fingernails were eligible for inclusion in this review.

Two studies among children living in Brescia, Italy, exposed to Mn from historic or current ferromanganese alloy plant activities, examined association between Mn in fingernails and Mn in environmental media (Lucas et al. 2015; Butler et al. 2019). Fingernail Mn concentrations varied significantly across three study sites and were significantly higher in subregions influenced by current or historical ferromanganese alloy plant activity, compared to the site with no history of ferromanganese alloy plant activity (Lucas et al. 2015). Correlation analyses suggest that fingernail Mn concentrations were significantly positively correlated with Mn concentrations in indoor and outdoor dust, and with Mn in soil collected near children’s homes (Lucas et al. 2015; Butler et al. 2019). There was no significant correlation with Mn levels in PM₁₀ fraction of 24 h personal air samples (Lucas et al. 2015). However, in a regression analysis of Mn in fingernails against Mn in each environmental sample adjusting for sex, age, and Mn in all other environmental samples, fingernail Mn was significantly associated only with Mn in soil (Butler et al. 2019). Using WQS regression, increased weighted environmental exposure index was associated with increased fingernail Mn (Butler et al. 2019).

Ruiz-Azcona et al. (2021) and Ruiz-Azcona et al. (2022) collected samples from 130 adults residing in the vicinity of a ferromanganese allow plant in Cantabria, Spain. Mn levels were significantly higher in fingernails of residents at a shorter distance (≤1.5 km) from the industrial emission source (mean 1.440, median 0.918 µg/g), compared to those living at a distance >1.5–10 km from the source (mean 0.433, median 0.331 µg/g). Mn in fingernails was negatively correlated with the distance from the emission source (Spearman’s r = −0.54, p < 0.001).

In these studies, fingernail samples were cleaned with Triton and nitric acid (Supplemental Material 2, Table S6).

3.8.2. Summary and conclusion: Mn in fingernails as a biomarker of Mn exposure

Few studies examined associations between Mn in fingernails and metrics of environmental Mn exposure. Specifically, two studies in children and one study in adults living in areas potentially influenced by historic or current ferromanganese alloy plant activities indicated that fingernail Mn levels may differ between Mn-exposed and unexposed individuals, may increase with increasing environmental exposure levels, and decrease with increasing distance from the emission source. Overall, more studies are needed to conclude regarding the utility of Mn in fingernails as a marker of environmental exposure.

3.9. Mn in tooth components as a biomarker of Mn exposure

It is believed that concentrations of Mn in the dentin and enamel of deciduous teeth reflect Mn exposures that occurred from second-trimester prenatal development through early postnatal development (Leonhard et al. 2019). In order to use tooth Mn concentrations as a biomarker of environmental exposure to Mn, it is necessary to know its distribution in different tooth compartments. Arora et al. (2011) demonstrated that, in enamel of teeth shed by children with no known elevated environmental or dietary exposure to Mn, the Mn:Ca ratios were highest on the outer edge of enamel for approximately 20–40 μm from the surface. (Mn concentrations are normalized to Ca to account for variations in bone mineral density within and between teeth, and expressed as 55 Mn:43 Ca ratio (e.g. Bauer et al. 2017).) Mn:Ca ratios deeper into enamel were low. Mn levels in enamel were generally lower than in dentin, except for surface enamel. The highest Mn levels were seen in dentin close to the pulp chamber. Mn distribution within this zone was not homogenous, showing localized spots of higher Mn:Ca levels. A high Mn zone was also found in the dentin near the incisal tip. It was shown that this zone of high Mn corresponds to the prenatally formed dentin in incisors. Using confocal laser scanning microscopy to visualize the neonatal line, it was confirmed that Mn:Ca levels in prenatally formed dentin were higher than those in postnatally formed dentin. Data obtained from analyses of a standard reference material (SRM) with known elemental composition were used to convert the Mn:Ca counts to Mn concentrations in μg/g. Mn concentrations quantified on the basis of analysis of the SRM were 0.1–0.2 μg/g in inner enamel (excluding the surface), 0.5–0.7 μg/g in the dentin close to the pulp, and 1.0–2.0 μg/g in discreet spots near the pulp chamber (Arora et al. 2011). Because there is no active metabolism of elements after the completion of dentin, concentrations of metals in dentin of permanent teeth may not be useful markers of exposure (Kumagai et al. 2012). In addition, the availability of permanent teeth for analysis may represent a problem (Malara et al. 2016).

3.9.1. Mn in tooth components and Mn in environmental media

Bauer et al. (2017) found that Mn concentrations in prenatal and postnatal dentin were not significantly different among Italian children living in three different areas: near an active ferromanganese alloy plant, in the vicinity of historical ferromanganese alloy plant operations, and in a tourist area with no history of ferromanganese alloy plant activities.

Arora et al. (2012), Gunier et al. (2013), and Gunier et al. (2014) enrolled pregnant women who were farmworkers themselves or lived with farmworkers and were exposed to Mn-containing fungicides. Mn levels were measured in house dust collected during pregnancy, and were estimated in outdoor air, diet, and drinking water. Mn was measured in shed teeth of their children when they were approximately seven years old. Mn in dentin was not correlated with estimated levels of Mn in outdoor air, diet or drinking water. However, Mn levels in the second and third-trimester prenatal dentin, and Mn in postnatal dentin, were significantly correlated with Mn in house dust concentrations and Mn in house dust loading (correlation coefficients ranging from 0.21 to 0.40). Mn dust load was a significant predictor of Mn levels in prenatal dentin in a multivariate modeling, with an estimated increase of 3.3% (95% CI: 0.3–6.4%) in prenatal dentin Mn levels per an interquartile range increase in Mn dust load (1465 µg/m²). Mn in prenatal enamel was not associated with Mn in house dust.

More details on these studies can be found in Supplemental Material 2, Table S7.

3.9.2. Summary and conclusion: Mn in tooth components as a biomarker of Mn exposure

Overall, the available data are limited. The analytical methods for Mn measurements in teeth are technically challenging, require knowledge in tooth histology, and validation against other biological and environmental matrices before they can be used in health studies (Arora and Austin 2013; Gunier et al. 2014). Concentrations of metals in permanent teeth may not be useful markers of exposure (Kumagai et al. 2012).

3.10. Measurements of Mn in bone by in vivo neutron activation analysis (IVNAA)

Aslam et al. (2008) reported on the application of a noninvasive measurement of Mn in bone, using in vivo neutron activation analysis (IVNAA). The authors believe that Mn in bone may reflect cumulative Mn exposure and may be a better predictor of health risks from chronic exposure than the existing alternatives. The study was conducted using the McMaster University Tandetron facility for neutron activation. The facility was calibrated using a hand bone phantom with known concentrations of Mn. Ten healthy male volunteers ages 25.7–74.3 years (mean 51.4 ± 16.0 years), with no history of occupational exposure to Mn, participated in this study. The minimum detection level (MDL) estimated from the calibration of the facility using the hand bone phantoms was 1.6 μg Mn/g Ca. A comparable in vivo MDL was derived for the volunteers in this study: 2.1 ± 0.70 μg Mn/g Ca. There was substantial variability in the measurements of Mn/Ca obtained for the volunteers, with most data points below the MDL for this technique and negative results for very low concentrations of Mn in bone. The negative values have no physical meaning and are explained by “statistical processes associated with radiation detection and measurement”. The authors acknowledge the large uncertainties associated with measurements of low levels of Mn in non-occupational populations and believe that the technique may be “a suitable means of screening patients and people exposed to excessive amounts of Mn who could develop many-fold increased levels of Mn in bones”. The authors also note that the dose equivalent received by study subjects is an important constraint of these measurements.

3.11. Reference values for Mn in biological matrices

Accurate RVs are essential for interpretation of biomonitoring data. Manganese, as an essential element and a normal component of the diet, is present in all human tissues and fluids. Because levels of Mn in biological matrices are influenced by environmental, physiological, and lifestyle factors, they may differ between countries and regions. Therefore, RVs should be established at a national/regional level. Mn levels may also change over time, so RVs should be established periodically (Bocca et al. 2010; Hoet and Roels 2015).

Derivation of accurate RVs requires measurements to be performed in a representative sample of a well-characterized population; the risk of external contamination during sampling or from the laboratory environment should be minimized; intra-laboratory quality assurance (validation and uncertainty estimation of measurements), as well as successful participation in external quality assessment schemes are essential. Because these are frequently disregarded or partially explored, accurate RVs are scarce (Bocca et al. 2010; Hoet and Roels 2015). RVs are derived from statistical calculations and are not based on adverse health effects or toxicological data. Because RVs are not health-based, observed values above them do not necessarily indicate a health risk (Hoet and Roels 2015; Da Silva et al. 2017) and should not be interpreted as a threshold for action either at an individual or population level (Stojsavljevic et al. 2020). RVs may be useful, for example, for identifications of individuals or groups with unusually high exposures and for assessment of the effectiveness of measures to reduce exposures, for tracking time-trends and regional differences in exposure levels (Bocca et al. 2010).

Thirty studies aimed specifically at derivation of RVs were identified for this review. RVs were derived for Mn in blood, serum/plasma, urine, and hair from studies conducted in various countries. Most studies deriving RVs were based on samples less than 500 subjects, although two studies included close to 7500 individuals. Study subjects varied among studies, including adults, young children and adolescents, and lactating women. The definition of RVs varied substantially among studies, with the most common being the 5–95th percentile. RVs derived for the same media varied across studies and geographic locations. These RVs are useful in comparing Mn levels found in different biological matrices in different populations. Table 2 summarizes the findings displaying the range of RVs reported for total and subpopulations in the studies along with the corresponding definition. Detailed characteristics and findings from these 30 studies are presented in Supplemental Material 3, Table S8.

Table 2. Summary of reference values (RVs) reported for Mn in biological matrices.

Matrix	RV definition	Range of reported RVs	References
Blood	5–95th percentiles	Total: 4.7–29.3 µg/L	White and Sabbioni (1998); Goulle et al. (2005); Heitland and Koster (2006b); Bocca et al. (2011); Freire et al. (2015); Jain and Choi (2015); Da Silva et al. (2017); Pino et al. (2017); Saravanabhavan et al. (2017); Yedomon et al. (2017); Snoj Tratnik et al. (2019); Stojsavljevic et al. (2020); Luan et al. (2022)
		Males: 4.6–29 µg/L
		Females: 5.7–31 µg/L
	2.5 − 97.5th percentiles	Total: 4.2–18.8 µg/L
		Total: 5.6–33.4 ng/g (5.89–35.16 µg/L)
		Males: 3.8–22.2 µg/L
		Females: 4.0–20.2 µg/L
		Lactating women: 9.0–30 µg/L
		<12 years: 5.9–19.5 µg/L
		12–19 years: 5.6–18.5 µg/L
		20–64 years: 4.9–18.9 µg/L
		65+ years: 4.3–6.5 µg/L
	95th percentile	Children and adolescents: 16–18 µg/L
		Total: 29.2 ng/g (30.7 µg/L)
Serum or plasma	95th percentile	Serum	Heese et al. (1988); Minoia et al. (1990); Rukgauer et al. (1997); Muniz et al. (2001); Goulle et al. (2005); Bocca et al. (2010); Stojsavljevic et al. (2020); Li et al. (2021)
		Total: <0.37 ng/g
		Total: 5.12 µg/L
	5–95th percentiles	Plasma
		Total: 0.63–2.26 µg/L
	Median [2.5 − 97.5th percentiles]	Serum
		Males: 0.40 [0.00, 1.40] µg/L
		Females: 0.39 [0.00, 1.20] µg/L
		<1 year: 0.48 [0.00, 1.63] µg/L
		1–3 years: 0.43 [0.00, 1.22] µg/L
		3–6 years: 0.38 [0.00, 1.36] µg/L
		6–12 years: 0.36 [0.00, 1.25] µg/L
	Mean	Serum
		Total: 0.5–1.4 µg/L
		Females: 1.2 µg/L
		22–75 years: 14.3 nmol/L (0.78561 µg/L)
		<2 years: 30–39 nmol/L (1.6481–2.1426 µg/L)
		2–10 years: 23–26 nmol/L (1.2636–1.4284 µg/L)
		10–18 years: 17.8–22 nmol/L (0.97790–1.2086 µg/L)
Urine	Mean	Adults: 0.003–1.9 µg/L	Minoia et al. (1990); Paschal et al. (1998); White and Sabbioni (1998); Goulle et al. (2005); Heitland and Koster (2006a); Hoet et al. (2013); Zhang et al. (2017); Zeng et al. (2019); Shi et al. (2020))
		Children: 0.1 µg/L
	5–95th percentiles	Total: 0.11–1.32 µg/L
	95th percentile	Adults: 0.02–3.33 µg/L
		Children: 0.25 µg/L
	97th percentile	Adults: 0.06–0.1.89 µg/L
	2.5–97.5th percentiles	Adults 0.1–1.9 µg/L
		Children: 1.4–25.3 ng/mg creatinine
Hair	5–95th percentiles	Total: 0.02–0.6 µg/g	Goulle et al. (2005); Chojnacka et al. (2010); Llorente Ballesteros et al. (2017))
		Children and adolescents: 115–937 µg/g
	10–90th percentiles	Young adults: 0.5–1.0 µg/g

We found that reporting of findings in many studies was inadequate. Specifically, reporting was poor regarding the study populations, sampling strategies, biological sample collection and preparation, validation of the analytical method, and internal and external quality control procedures. Although there exist recommendations on derivation of RVs, such as those developed by the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC), the International Union of Pure and Applied Chemistry (IUPAC), and the German Human Biomonitoring Commission (Saravanabhavan et al. 2017), methodologies used by studies included in this review are diverse. Methods for selection of study subjects, collection of biological samples, their preparation and analysis, differ from study to study. Definitions of RVs and statistical methods for their derivation are not harmonized either. It appears that relatively few studies used a sampling procedure to create a sample representative of the population of the country or region of interest. Inductively coupled plasma mass spectrometry (ICP-MS) and atomic absorption spectrophotometry (AAS) were most commonly used for measurements of Mn in biological matrices. These methods may produce somewhat different results. For example, as demonstrated by White and Sabbioni (1998), the values of Mn in urine by ICP-MS showed a significant positive bias compared with those measured by AAS. Given the wide variation in the methods used to derive RVs, standardization of methodologies and definitions would substantially facilitate their interpretation (Ichihara et al. 2017).

4. Conclusion

This review summarized findings from studies that assessed the utility of Mn measurements in biological materials as potential biomarkers of Mn exposure, as well as findings on the derivation of RVs for Mn in biological materials.

Performance of potential exposure biomarkers was analyzed in terms of the relationships between Mn in biological matrices and Mn in environmental media. Our analysis of the literature suggests that Mn levels in blood, saliva, and urine, may not be useful as markers of Mn exposure in non-occupational settings. Measurements of Mn concentrations in the dentin and enamel of deciduous teeth are technically challenging. These measurements are thought to characterize exposures that occurred during different developmental windows. Application of IVNAAs for noninvasive measurements of Mn in bone is associated with large uncertainties and is constrained by the dose equivalent received during the irradiation procedure. No consistency was revealed across studies examining associations of hair Mn with Mn in environmental media. The few studies that evaluated Mn in fingernails and toenails, suggest that Mn in nails may reflect Mn levels in environmental media and may discriminate between groups of children exposed to different environmental Mn levels. The absence of a certified reference material for nails represents a challenge associated with the use of this biological material for Mn measurements. We found that studies aimed at derivation of RVs for Mn in biological materials have methodological limitations and accurate RVs are scarce.

Comparisons across studies and synthesis of their findings were challenging. These challenges include differences in levels and gradients of external Mn exposure, use of different analytical techniques to measure Mn in biological materials, and variations in procedures used for cleaning of hair and nail samples prior to analysis. Depending on techniques used for cleaning, the extent to which Mn measured in hair or nails reflects Mn body burden and external contamination may vary from study to study. In addition, not all studies reported on measures taken to prevent external contamination of biological samples by Mn.

Interpretation of associations between Mn in biological materials and measures of external Mn exposure was complicated by multiplicity of Mn sources in the environment. Few studies assessed Mn intake from diet, the main source of Mn in the general population. Estimates of external Mn exposure based on a single source, such as air or drinking water, may capture only a relatively small proportion of the total exposure. Also, Mn exposure measurements based on stationary air sampling or modeled air Mn concentrations, as well as measurements of Mn in environmental samples at a single time point, may not characterize external exposure sufficiently well.

This review identified biomarkers of exposure to Mn that demonstrated varying degrees of correlation between external measures of Mn exposure and internal concentrations of Mn. In interpreting these findings, it should be noted that not all studies reviewed considered all possible sources and routes of exposure, so that Mn concentrations measured in body tissues might reflect contributions from unmeasured exposure sources and routes, resulting in an apparently weaker association than might otherwise be the case. Assuming that the concentration of Mn measured in a particular tissue accurately reflects total exposure from all exposure sources and routes, this internal dose metric would still represent a useful internal dose metric for risk assessment purposes.

In conclusion, our analysis of currently available evidence suggests that there is no strong evidence for measurements of Mn in any biological matrix as an ideal biomarker of Mn exposure in non-occupational settings. Mn levels in hair and nails may be considered for further evaluation as potential biomarkers of environmental Mn exposure. More studies conducted according to a harmonized protocol may be required to assess the performance of Mn in these biological substrates as markers of environmental exposure. Because of methodological limitations in studies aimed at derivation of RVs for Mn in biological materials, accurate reference values are scarce.

Acknowledgments

The authors thank Dr. Athena Keene from Afton Chemical, a member company of the International Manganese Institute (IMnI), for helpful comments on preliminary drafts of this article; although Dr. Keene also serves as the Afton Chemical representative on the IMnI Health, Safety, and Environment Committee, her comments were provided as an independent scientist with expertise in manganese toxicity. The authors gratefully acknowledge the comments of the three external reviewers selected by the editor and anonymous to the authors. As a result of their comments, we limited the scope of the article and made a number of changes that improved the original draft of this article.

Declaration of interest

This work was conducted under a research contract to review the recent scientific literature on manganese biomarkers between the International Manganese Institute (IMnI, www.manganese.org), an industry association representing the international manganese industry, and Risk Sciences International (RSI, www.risksciences.com), a Canadian company established in 2006 in partnership with the University of Ottawa. Additional financial support was provided by the Natural Sciences and Engineering Research Council of Canada (NSERC, www.nserc.ca) to D. Krewski, who holds the NSERC Chair in Risk Science at the University of Ottawa. N. Farhat, D. Krewski, N. Karyakina, D. Mattison, F. Momoli, S. Ramoju, and N. Shilnikova were compensated by RSI for their contributions to the review. N. Farhat was also supported in part by postdoctoral fellowships from the McLaughlin Center for Population Health Risk Assessment at the University of Ottawa and from Carleton University under the Mitacs Accelerate program (www.mitacs.ca), a peer-reviewed university-industry partnership program with RSI as the industrial partner. B. Cline contributed to this article as a senior scientist at IMnI. The authors, whose affiliations are shown on the title page, had sole responsibility for preparation of this article, including determining the strategy for reviewing the scientific literature summarized in this article, synthesizing the findings, and drawing conclusions. Although Dr. Cline briefed the IMnI Health, Safety and Environment Committee on the results of this review, the Committee did not provide input to this work. None of the authors have appeared before regulatory agencies on behalf of the sponsors or appeared as experts in legal proceedings concerning matters reviewed in this article. The scientific opinions and conclusions expressed in the article are exclusively those of the authors and are independent of the sources of financial support.

Supplemental material

Supplemental data for this article is available online at https://doi.org/10.1080/10408444.2022.2095979