Manganese Exposure and the Neuropsychological Effect on Children and Adolescents: a Review

Revisión bibliográfica / Literature review

Manganese exposure and the neuropsychological effect on children and adolescents: a review

José A. Menezes-Filho,1,2 Maryse Bouchard,3 Paula de N. Sarcinelli,2 and Josino C. Moreira 2

Suggested citation Menezes-Filho JA, Bouchard M, Sarcinelli PN, Moreira JC. Manganese exposure and the neuropsychological effect on children and adolescents: a review. Rev Panam Salud Publica. 2009;26(6):541–8.

ABSTRACT Objectives. Manganese (Mn) is an essential element, but overexposure can have neurotoxic effects. Methods. In this article, we review and summarize studies on exposure to Mn and nervous system impairments in children. Results. We identified 12 original articles published between 1977 and 2007. Overexposure to Mn was suspected to occur through diverse sources: infant milk formula, drinking water, industrial pollution, and mining wastes. The most common bioindicator of exposure to Mn was hair Mn content, but some studies measured Mn in blood, urine, or dentin; one study on prenatal exposure measured Mn content in cord blood. Most studies indicate that higher postnatal exposure to Mn is associated with poorer cognitive functions and hyperactive behavior. Conclusions. The limitations of the existing studies are numerous: most were cross- sectional, had a modest sample size, and lacked adjustment for important confounders. Future investigations should be performed on a larger sample size and include a more detailed exposure assessment, addressing multiple sources of exposure such as food, water, and airborne particulates.

Key words Manganese; neurobehavioral manifestations; cognition; neurotoxicity.

Manganese (Mn) is very common in tial nutrient, with an adequate daily in- ing concern for possible harmful effects the environment, being the 5th most take level of 1.8–2.6 milligrams (mg) for of environmental exposures to Mn, in- abundant metal and the 12th most abun- adults (2), but also has the potential to cluding the economic implications of in- dant element on earth (1). It is an essen- produce neurotoxic effects when, de- tellectual deficits due to Mn toxicity (5). pending on the route and dose of expo- Mn is an essential element, necessary 1 Laboratory of Toxicology, Faculty of Pharmacy, sure, it accumulates in an organism, es- for bone mineralization, energy and pro- Federal University of Bahia, Salvador, Bahia, pecially in the brain (3). The vast majority tein metabolism, regulation of cell me- Brazil. Send correspondence and reprint requests to: José A. Menezes-Filho, Faculdade de Farmácia, of studies on neurotoxic effects of Mn tabolism, and protection against oxida- Federal University of Bahia, Av. Barão de Jere- were conducted in occupational settings tive stress (6). With normal dietary moabo s/n Campus Universitário de Ondina, where exposure occurs mainly though consumption, systemic homeostasis of 41170-115 Salvador, Bahia, Brazil; telephone: 55-71-3283-6960; fax: 55-71-3283-6949; e-mail: inhalation of airborne particulates. Few Mn is maintained by both its rate of [email protected] studies have investigated possible over- transport across enterocytes lining the 2 Public and Environmental Health Program, Na- tional School of Public Health, Oswaldo Cruz exposure of children to Mn. It is gener- intestinal wall and its efficient removal Foundation, Av. Leopoldo Bulhões 1480 Mangui- ally accepted that children are at greater in the liver (7). Ingested Mn is subjected nhos, Rio de Janeiro, Rio de Janeiro, Brazil. risk than adults exposed to the same con- to homeostatic mechanisms that regulate 3 Université du Québec à Montréal, CINBIOSE, Suc- cursale Centre-ville, Montreal, Quebec H3C 3P8, taminants from the environment (4). Al- its concentration in the body, but expo- Canada. though data are sparse, there is increas- sure through inhalation bypasses most

Rev Panam Salud Publica/Pan Am J Public Health 26(6), 2009 541 Literature review Menezes-Filho et al. • Manganese exposure and effects in children of these protective mechanisms. There- sorption rate of ingested Mn in children lected studies for this review only if in- fore, inhaled Mn poses greater risks of was higher than in adults (19). Second, formation was presented that pertained toxicity, as seem to be supported by the high demand for Fe linked to growth specifically to assessment of the neu- preponderance of occupational reports could further enhance the absorp- ropsychological effects on children ex- of toxicity in environments with air- tion of ingested Mn (20). Third, a low exposed to Mn from all possible sources. borne exposure. cretion rate was observed in infants due We found six articles on exposed popu- Mn shares several characteristics with to the poorly developed biliary excretion lations in the United States, Canada, iron (Fe); both are transition metals with mechanism in neonatal animals (21). Spain, Bangladesh, and Malaysia. In the valences of 2+ and 3+ in physiological Data are lacking on the overall reten- reference list of those articles, we se- conditions and proximate ionic radius. tion of ingested Mn in infants and chil- lected six more investigation reports. In addition, as Mn and Fe both strongly dren, but in mice, rats, and kittens, there Only full-length research articles were bind to transferrin and accumulate in the is almost a complete absence of biliary reviewed. mitochondria, low Fe stores are associ- Mn excretion during the neonatal period ated with increased Mn uptake and re- (22). With inhaled Mn, the ratio of in- RESULTS tention in the blood (8). haled air/weight is much higher in chil- Neurotoxic effects resulting from ex- dren, leading to a higher dose of expo- Table 1 shows a schematic summary of cessive Mn exposure were first described sure for almost any contaminant in air. the studies with their general characteris- by Couper in 1837 in Scottish laborers Thus, exposure during this period may tics such as country of origin, type of de- grinding Mn black oxide in the chemical result in increased delivery of Mn to the sign, sample size, source of exposure, industry (9). Neurological symptoms of brain and other tissues. and outcome investigated. In Table 2, we manganism include decreased memory Mn neurotoxicity has been extensively summarize the main findings of the stud- and concentration, fatigue, headache, studied and a lot has been learned about ies. The first published article exploring a vertigo, equilibrium loss, insomnia, tinni- its mechanism of action at the cellular possible adverse neurobehavioral effect tus, trembling of fingers, muscle cramps, and molecular levels and the detection of of exposure to Mn was a case-control rigidity, alteration of libido, and sweat- subclinical effects at low exposures, al- study conducted in Canada on learning- ing (10). Many reports of neurotoxic ef- though there is a long way to go until its disabled children (27). The authors pos- fects in Mn-exposed workers were later toxicology can be fully understood. Re- tulated that nutritional deficiencies and published (for a review, see Zoni et al. cently, several literature reviews have exposure to toxic substances, using hair (11)), and the definition of Mn intoxica- been published on aspects such as neu- mineral content as an indicator of body tion has evolved to include subclinical rotoxic effects on exposed laborers (15, burden, would be associated with learn- signs of intoxication indicated by alter- 23), the application of magnetic reso- ing disabilities in children. A cohort of ations of neurobehavioral functions (12). nance imaging (24), neuropsychological third- and fourth-grade students was Mn can accumulate in the central ner- testing for the assessment of Mn neuro- evaluated on a series of cognitive scales. vous system, particularly the basal gan- toxicity (11), Mn neurotoxicity focused The children diagnosed with a learning glia but also the cortex (13). Exposure to on neonates (25), and neurotoxicology of disability were compared with children Mn was shown to interfere with several chronic Mn exposure in nonhuman pri- without such a diagnosis and were neurotransmitter systems, especially in mates (26). matched by frequency for school atten- the dopaminergic system in areas of the As far as we know, no study has ad- dance, grades, and gender; the groups brain responsible for motor coordina- dressed the issue of children’s exposure had similar socioeconomic status. Chil- tion, attention, and cognition (14, 15). and the effect on behavior and cognition. dren with learning disabilities had a sig- Mn is a potent dopamine oxidant, which Therefore, we propose here to review nificantly higher concentration of hair could explain the toxic lesions in certain and summarize the scientific literature Mn as well as six other elements (sodium, dopaminergic brain regions (16). Exces- on the associations between Mn expo- cadmium, copper, lead, chromium, and sive exposure could result in dopamine sure and effects on children’s neuropsy- lithium), which makes the findings diffi- receptor loss or inactivation through chological functions. cult to interpret. damage to the membrane mediated by The high Mn concentration in infant free radicals or cytotoxic quinones gen- METHODOLOGY milk formula drew the attention of re- erated by the Mn catalyzing effect on au- searchers (28), who conducted a two- tooxidation of this neurotransmitter (17). A systematic scientific literature part study in the United States. First, hair One hypothesis for the toxic mechanism search was carried out on the electronic Mn concentration was measured in chil- of Mn is the production of excess free database Medline (National Library of dren fed infant formula and in children radicals in the nerve cell, potentiating Medicine, Bethesda, Maryland, United exclusively breast-fed. Hair Mn levels lipid peroxidation and resulting in tissue States) for 1997–2007, using the key- were found to increase significantly destruction (18). words manganese, child, children, in- from 0.19 microgram per gram (µg/g) at Several factors could predispose chil- fant, childhood, adolescents, neurotoxic- birth to 0.69 µg/g at 4 months in the in- dren to Mn overexposure and subse- ity, neuropsychological effects, behavior, fant formula group; no significant in- quent toxic effects. Exposure to Mn by motor, cognition, cognitive, intellectual crease was observed in the breast-fed ingestion or inhalation could have differ- functioning, hyperactivity, ADHD, and group. Second, the metal levels in the ent consequences in children than in hyperactive behaviors. Upon reviewing hair of children with hyperactivity were adults and through different mecha- a large number of article abstracts compared with those in age-matched nisms. First, the observed intestinal ab- yielded by these search terms, we se- children without this disorder. Signifi-

542 Rev Panam Salud Publica/Pan Am J Public Health 26(6), 2009 Menezes-Filho et al. • Manganese exposure and effects in children Literature review

TABLE 1. Characteristics of studies on the association between children’s exposure to manganese and neuropsychological effects, published between 1977 and 2007

N Study Country Design Subpopulation Exp./Ref.a Exposure source Effect sought

Pihl and Parkes 1977 (27) Canada Case control Children in 3rd and 4th 31/22 Not reported Learning deficit grades (± 10 years) Collipp et al. 1983 (28) United Cross- Neonates and children 16/44 Baby formula Hyperactivity States sectional ≤ 10 years Learning deficit Barlow 1983 (29) United Cross- Children and adolescents 68/65 Unknown Hyperactivity Kingdom sectional ≤ 16 years He et al. 1994 (30) China Cross- Children 11–13 92/92 Drinking water Neurobehavioral sectional years Zhang et al. 1995 (31) China Cross- Children 11–13 years 92/92 Drinking water Learning deficit sectional Woolf et al. 2002 (32) United Case report Child 10 years 1 Well water General States neuroeffects Takser et al. 2003 (33) France Prospective Neonates 247 Mother’s blood Psychomotor (60% loss) development Torrente et al. 2005 (34) Spain Cross- Adolescents 12–14 years 54/45 Industrial pollution Cognitive deficit sectional Wasserman et al. 2006 (36) Bangladesh Cross- Children 10 ± 0.5 years 142 Well water Cognitive deficit sectional (intelligence quotient) Wright et al. 2006 (37) United Cross- Adolescents 11–13 years 32 Mining waste Neuropsychological States sectional and neurobehavioral Ericson et al. 2007 (39) United Cross- Adolescents 11–13 years 27 Mother’s blood Behavioral States sectional in disinhibition a follow-up cohort Bouchard et al. 2007 (38) Canada Cross- Children and adolescents 28/18 Well water Hyperactive sectional 6–15 years behaviors a Exp./Ref. = exposed/referents. N = one studied group (no controls). cantly higher levels of Mn were ob- The first publication addressed Mn lev- than controls on 5 of 12 neurobehavioral served in the hair of hyperactive chil- els in drinking water and children’s hair tests: digit span, Santa Ana manual dex- dren (0.43 µg/g) than in controls (0.27 and the associated neurobehavioral ef- terity, digit symbol, Benton visual re- µg/g). fects on children. The second article ad- tention test, and pursuit aiming test. In Barlow (29) investigated the associa- dressed the association of Mn hair and addition, exposed children had signifi- tion of Mn exposure and hyperactivity in blood levels with learning deficiencies cantly lower school performance than the United Kingdom. A hair sample was and the level of neurotransmitters mea- the control children. Multiple regression collected by the family and sent to the re- sured in peripheral blood. analysis showed that school grade for searchers by mail. Slightly higher Mn The exposed group was composed of language and mathematics was corre- concentrations were detected in children 92 students from a village where the con- lated to serum levels of the neurotrans- with hyperactivity (0.84 ± 0.64 µg/g) centration of Mn in drinking water was mitters 5-hydroxytryptamine, norepi- than in controls (0.68 ± 0.45 µg/g), between 0.24 and 0.35 mg per liter (L) for nephrine, and dopamine and to the although this difference was not signifi- many years. They were compared with activity of acetylcholinesterase. cant (p = 0.10). The author identified sev- children in another rural town with low Woolf and associates (32) reported a eral limitations—including the subjectiv- Mn in water (< 0.03 mg/L). case of suspected Mn intoxication in a ity of the diagnosis made by different The children in both groups were eval- 10-year-old child from a suburb of psychiatrists, psychologists, and social uated by neurobehavioral tests that as- Boston, Massachusetts (United States). workers—and the risk of contamination sessed emotional state, motor coordi- For more than 5 years, the family had of hair samples. In addition, the control nation, visual memory, and time of used water from a tube well with a very group was inappropriate, coming from a reaction. Children from the exposed vil- high Mn concentration (1.21 mg/L). The different area of residence, and socioeco- lage had a mean hair Mn concentration child had high Mn levels in serum (9.00 nomic status was not controlled. significantly higher than control children µg/L, reference normal < 0.27 µg/L), A research group in China published (1.25 ± 0.72 µg/g versus 0.96 ± 0.42 whole blood (38.2 µg/L, reference nor- two articles (30, 31) on the investigation µg/g); the same difference was observed mal < 14.0 µg/L), and hair (3.09 µg/g, of a rural community with Mn-contami- with Mn blood levels (33.9 µg/L versus reference normal < 1.20 µg/g). However, nated drinking water following the use 22.6 µg/L). Children from the exposed magnetic resonance imaging showed no of high-Mn sewage water for irrigation. village had lower performance (p < 0.01) changes in the globus pallidus and basal

Rev Panam Salud Publica/Pan Am J Public Health 26(6), 2009 543 Literature review Menezes-Filho et al. • Manganese exposure and effects in children

TABLE 2. Results of studies on the association between children’s exposure to manganese and neuropsychological effects, published between 1977 and 2007

Exposure Analytical Study index method Mean Main findings

Pihl and Parkes 1977 (27) Mna in hair AA-Flameb Case 0.83 µg/gc Children with learning disability had significantly higher Mn Control 0.58 µg/g levels in hair than controls (P < 0.001). Adjusted for sex, age, school, class, and language. Collipp et al. 1983 (28) Mn in hair GFAASd Case 0.43 µg/g Children with learning disability had significantly higher Mn Control 0.27 µg/g levels in hair than controls (P < 0.05). Adjusted for age. Barlow 1983 (29) Mn in hair GFAAS Exp.e 0.84 µg/g Higher Mn level in hair in hyperactive children (P < 0.10). Ref.f 0.68 µg/g He et al. 1994 (30) Mn in hair AA-Flame Exp. 1.25 µg/g Differences in Mn in hair (P < 0.001) exposed/referents. Ref. 0.96 µg/g Poorer neurobehavioral performance (P < 0.01) in exposed than in referents. Significant negative relation between Mn in hair and visuomotor coordination, executive function, and memory (r = –0.213 to r = –0.319). Zhang et al. 1995 (31) Mn in hair and AA-Flame Exp. 1.25 µg/g, Ref. 0.96 µg/g Higher Mn in hair and Mn in blood levels (P = 0.001) in Mn in blood Exp. 33.9 µg/Lg, Ref. 22.6 µg/L exposed children compared with referents. Exposed had lower grades for math and language. Woolf et al. 2002 (32) Mn in hair ICP-MSh 3.09 µg/g Case report on a child with high Mn exposure who had normal Mn in blood 38.2 µg/L cognition, except for impaired memory. Mn in urine 8.5 µg/L Takser et al. 2003 (33) Mn in hair GFAAS Initial, at 6 years Higher Mn in cord blood significantly associated with poorer Mn in umbilical 0.75, 0.77 µg/g (GM)i psychomotor indices (r = –0.33, P < 0.001) at 3 years. cord blood 38.5, 38.6 µg/L (GM) No correlation observed for postnatal Mn blood levels and Mn in placenta 0.1 µg/g (GM) later psychomotor development (assessed up to 6 years of age). Adjusted for child’s gender and maternal education. Torrente et al. 2005 (34) Mn in hair ICP-MS Exp. 0.18 µg/g No significant difference (P > 0.05) in Mn levels in hair in exposed and controls. Ref. 0.26 µg/g Attention negatively correlated (r = –225, P = 0.023) with lead in hair. No correlation observed with Mn in hair. Wasserman et al. Mn in blood ICP-MS 12.8 µg/L Mn in water negatively associated with intelligence quotient 2006 (36) Mn in water 795 µg/L (IQ) full-scale (β = –4.56) and IQ performance (β = –3.82). Adjusted for water arsenic levels and blood lead levels. Wright et al. 2006 (37) Mn in hair ICP-MS 0.47 µg/g Mn in hair inversely associated with full-scale (β = –0.01, P = 0.07) and verbal (β = –0.12, p = 0.002) IQ. Adjusted for sex and maternal education. Ericson et al. 2007 (39) Mn in dentin IMS j Not reported Mn in dentin (formed during 20th gestation week) was correlated with attention deficit hyperactivity disorder (r = 0.47, P = 0.001) and disruptive disorder (r = 0.41, P = 0.001) at school age. Adjusted for lead levels. Bouchard et al. 2007 (38) Mn in hair ICP-MS Exp. 6.2 µg/g Mn in hair levels significantly higher (P = 0.025) in the more Ref. 3.3 µg/g exposed group than in the less exposed group. Mn in hair was significantly associated with more problems with oppositional behaviors (P = 0.031) and hyperactive behaviors (P = 0.008). a Mn = manganese. b AA-Flame = atomic absorption spectroscopy, flame mode. c µg/g = micrograms per gram. d GFAAS = graphite furnace atomic absorption spectrometry. e Exp. = exposed. f Ref. = referents. g µg/L = micrograms per liter. h ICP-MS = inductively coupled plasma mass spectrometry. i GM = geometric mean. j IMS = ion mass spectrometry.

ganglia that indicated Mn accumulation. coordinate alternating movements. The tagmus, or fixed faces. The patient’s bal- The results of a battery of neuropsycho- child had no medical history that ex- ance was good, fine motor skills and re- logical tests (Table 2) showed that the plained these findings. The detailed neu- flexes were normal, and the sensory child had a normal intelligence quotient rological evaluation was otherwise nor- examination was unremarkable. The (IQ) but unexpectedly poor verbal and mal, with no tremors, normal gait and mother reported that teachers had no- visual memory as well as the ability to muscle tone, and no cog wheeling, nys- ticed the inattentiveness of the child. The

544 Rev Panam Salud Publica/Pan Am J Public Health 26(6), 2009 Menezes-Filho et al. • Manganese exposure and effects in children Literature review family discontinued all use of well water visuospatial abilities and abstract reason- lower intellectual function and verbal and the residence was connected to ing were used as a general intelligence memory scores. the municipal water system. Eighteen index. No significant correlations were Bouchard and collaborators (38) eval- months later, teachers continued to ob- found between chromium, Mn, nickel, uated children from a community in serve his difficulty in carrying out cer- and tin levels and cognitive measures. A Quebec, Canada, which was served by tain tasks due to attention deficit. It is negative correlation was found between municipal water supplied from two noteworthy that the 16-year-old brother hair lead level and the ability to concen- wells with different Mn concentrations of this child had normal psychometric trate, after adjustment for socioeconomic (0.61 mg/L versus 0.16 mg/L). Forty-six evaluation, with elevated Mn in hair but status, and a positive correlation was children participated in the study (me- not in blood. observed between mercury and visuo- dian age 11 years, range 6–15 years). Mn Takser and associates (33) conducted a spatial ability. The authors suggested levels were measured in children’s hair, prospective epidemiologic study in 247 that the mercury might come from con- and parents and teachers completed the healthy pregnant women and their babies sumption of fish, also associated with a Conner’s Rating Scale on hyperactive/ to investigate the long-term effect of in high intake of fatty acids with a beneficial attention deficit behaviors. The results utero Mn levels on their children’s psy- effect on brain function. showed that (1) children living in houses chomotor development. The population Wasserman and associates (35) inves- supplied with water at higher Mn con- was recruited from a maternity hospital tigated the association between Mn ex- centrations had significantly higher levin Paris (France). Mn levels were mea- posure from well water and children’s els of Mn in hair, and (2) hair Mn con- sured in the mothers’ blood and hair at IQ in Bangladesh. In a previous study, centrations were associated with higher delivery, in umbilical cord blood, and in Mn levels were found to be associated scores for hyperactive and oppositional placental tissue; children’s hair and with arsenic levels, and intellectual func- behavior in the classroom after adjusting monoamine (dopamine and serotonin) tion was negatively related to both met- for age, gender, and family income. In metabolite concentrations were assessed als. However, the association between addition, all children with high scores on in umbilical cord plasma. Children’s de- Mn and intellectual function was no these subscales, based on clinical cut-off velopment was assessed at 9 months with longer significant when the level of ar- provided by the test manual, had hair the Brunet-Lézine scales, and at 3 and 6 senic in water was included in the re- Mn levels higher than the upper limit of years of age with the McCarthy scales. gression model. Therefore, a study (36) the reference range (> 3.0 µg/g). An in- The results showed significant negative was designed to test the hypothesis that triguing result of this study is that 90% correlations between Mn levels in cord Mn had an independent effect on cogni- of the children did not drink water from blood and nonverbal scales (attention, tive function. In the new study, 54 chil- the tap, but elevated Mn levels were nonverbal memory) and boys’ manual dren were added who lived in houses found in hair of a large proportion of ability at 3 years, after adjusting for po- supplied by water with very low arsenic them; use of tap water in food prepara- tential confounders (sex and mother’s ed- levels (< 10 µg/L). All children received tion or in showering might contribute to ucational level). However, no relation be- a complete clinical examination and pro- the dose of exposure. This pilot study tween Mn and development subscales vided urine samples for determination was used to design a larger epidemio- was observed at the 6-year follow-up. The of arsenic and a blood sample for Mn, logic study on the effects of exposure to authors concluded that the exposure to lead, and arsenic determination. Intellec- Mn in tap water and food. high in utero Mn levels can affect chil- tual function was evaluated by the In another study in the same year, Er- dren’s psychomotor development but Wechsler Intelligence Scale for Children icson and associates (39) evaluated neu- that sociocultural factors might have version III (WISC-III), providing three robehavioral effects with scales that masked Mn effects in older children. Fi- IQ scores: verbal, performance, and full measure the degree of disinhibition. Un- nally, it was emphasized that fetal life can scale. After adjusting for covariables like previous investigations, the enamel be regarded as a period of great vulnera- (mother’s schooling, quality of housing, of shedding teeth was used as a matrix bility to Mn toxicity at low environmental access to television, and cranial circum- for Mn determination, with the objective levels. ference), increasing Mn water levels of measuring the previous exposure dur- A cross-sectional study was performed were associated with lower IQ on all ing the intrauterine stage when these tis- in Spain (34) with the objective of investi- three scales. Mn levels in water were sues were formed. The children of this gating correlations between hair metal not associated with Mn in whole blood, study were randomly selected from a concentrations and cognitive functions of and the latter were not associated with cohort of normal newborns from a adolescents living in areas with contrast- WISC-III scores. prospective developmental study started ing levels of industrial contamination. In the study conducted by Wright and in 1991 in the United States. Three psy- One hundred adolescents (12–14 years collaborators (37), the interaction of chil- chometric tests had been directly applied old) were selected from schools in urban dren with coexposure to Mn and arsenic to children in two phases of develop- areas and in the vicinity of the industrial was also evaluated. Hair was used as a ment: at 36 and 54 months. Scales that as- complex in the region of Tarragona, Cat- biomarker, psychometric scales were sessed behavior were applied to parents alonia. Hair samples were collected for used to assess cognitive functioning, and and teachers when children were in the analysis of cadmium, chromium, mer- a series of other scales were adminis- first and third grades. The authors con- cury, lead, Mn, nickel, and tin by induc- tered for behavioral evaluation (see cluded that, even after adjusting for lev- tively coupled plasma mass spectro- Table 2). The main finding of this study els of lead in the tooth enamel, measured scopy. The attention span test was used was that high levels of both Mn and ar- at the same time as Mn, children with as an indicator of attention, and tests of senic were significantly associated with high levels of this metal in the uterine

Rev Panam Salud Publica/Pan Am J Public Health 26(6), 2009 545 Literature review Menezes-Filho et al. • Manganese exposure and effects in children phase had higher scores on all scales of tions based on the level of Mn in differ- 0.15 µg/m3 (46). In general, studies that disinhibitory behavior: more children ent biological tissues. The lack of de- measure exposure from environmental played with forbidden toys at 36 months tailed exposure assessment to identify sources will always have weaker expo- of age, committed more errors by impul- Mn sources precludes the proper deter- sure levels and consequently the possi- siveness at 54 months, and were more mination of the risk factors and therefore bility of an association will be overshad- often evaluated by their mothers and does not allow for proposing solutions to owed by confounding variables. teachers as having externalizing and at- reduce exposure. One difficulty in the study of Mn ex- tention problems. Studies published so far have several posure is the lack of a well-recognized serious limitations, including sample bioindicator of exposure. Most studies DISCUSSION size, research design, adjustment for po- reviewed here used hair as an indicator tential confounding variables, and con- of Mn exposure, except for the Wasser- The literature on possible adverse ef- trol of coexposure to other neurotoxi- man et al. study (36), which measured fects of exposure to Mn on children’s cants. All reviewed studies except that of Mn in blood, and the Ericson et al. study health is relatively sparse, despite the Takser and colleagues (33) were cross- (39), which used Mn in the dentin of de- fact that Mn is acknowledged to be a sectional and had a modest sample size. ciduous teeth as a biomarker of in utero neurotoxin. There are some indications Cross-sectional studies provide less con- exposure. Use of hair has several advan- in the literature that Mn exposure might vincing evidence than cohort studies in tages over other biomarkers. Hair aver- be related to cognitive deficits and hy- showing a potentially harmful effect. ages off the variations of Mn levels peractive behaviors. Most studies attempted to control for found in blood or plasma, as it grows an Five studies reported indications of confounders by design, selecting a refer- average of 11 millimeters/month, thus adverse effects of exposure to Mn in ence group matched on some important representing a time-weighted average water on the central nervous system of variables (i.e., socioeconomic status, age, over the duration of exposure. In addi- children (29, 31, 32, 36, 38). Given that gender) or controlling confounders by tion, the sample collection procedure is high levels of Mn in well water are not restriction, as in the study of Wasserman simple, can be performed by minimally uncommon, this situation could pose a et al. (35), which selected households trained staff, and is not invasive. A major significant public health risk. Mn levels supplied with water containing low lev- drawback to the use of hair as a marker corresponding to the level found in the els of arsenic. of internal dose of exposure is exoge- Chinese study to have a significant effect The early studies had greater limita- nous deposition, which is particularly on children (around 0.3 mg/L) are found tions in this respect. For instance, in the problematic in the context of exposure to in 6% of household wells in the United study by Pihl and Parkes (27), the only airborne Mn particulates. Washing can States (40). controlled variable was age. The signifi- minimize the problem of external conta- Two studies (33, 39) addressed possi- cant difference observed in the levels of mination, but using a very abrasive ble adverse neurobehavioral effects from Mn in hair in children with and without reagent can erode the capillary structure, in utero exposure to Mn, although a learning disability could have been as- leading to loss of endogenous elements. sources of exposure were not specified. sociated with metal exposure from other Research would benefit from a standard- It is well documented that Mn levels rise sources—for example, due to airborne ized washing procedure. in the mother’s blood, where Mn has an particulate matter exposure from vehic- Others have hypothesized the possi- important role as a cofactor of several ular traffic and soil or dust around resi- bility of overexposure to Mn through enzymes that regulate metabolism and dences or different levels of exposure to ingestion of infant milk formula (47), bone growth (40, 41). Although new- metals from the diet. Although it is showering in water with a high Mn level borns’ exposure to Mn through infant known that maternal educational level (48), and adding methylcyclopentadi- milk formula is a cause for concern (see has a large influence on children’s intel- enyl manganese tricarbonyl (MMT) to the review by Ericson et al. (39)), only lectual development, only three of the gasoline (49). Whether these exposures one investigation addressed this possible reviewed studies controlled for this fac- represent a significant toxic risk remains source of exposure (29), despite the fact tor (33, 36, 37). to be investigated, but some exposure that baby formulas, especially those Studies conducted in the occupational data suggest that might be the case. In a based on soybean, have been shown to setting reported relationships between study in South Africa, levels of Mn in de- have considerably high Mn levels (42, neurological outcomes and concentra- posited dust and in blood of children 43). tion of airborne Mn. However, the levels were compared in Johannesburg, where Children’s exposure to Mn resulting of exposure in the general environment the gasoline additive MMT had been from environmental contamination was are several orders of magnitude lower used for 1 year, and in Cape Town, addressed in Spain (34), and no signifi- than in most occupational settings where where MMT is not used. The mean level cant association was found between Mn Mn is present in the process and there- of Mn in dust and blood was signifi- levels and cognitive deficits, and in the fore are more difficult to measure. For cantly higher in Johannesburg than in United States (37), where significant as- example, a concentration of 80.2 µg of Cape Town (blood Mn 9.8 ± 3.6 µg/L sociations with cognitive deficits have Mn per cubic meter (m3) was reported in versus 6.7 ± 3.5 µg/L). In addition, a sig- been observed in children living near a the air of a ferromanganese plant (44), nificant association was found between mining waste site. a Mn range of 1.5–450 mg/m3 was re- Mn in classroom dust and children’s Mn Another study (29) evaluated children ported in mines in the United States (45), blood levels (50). from the general population with no but the level in the air in nonindustrial- Although limited by poor study de- known exposure and reported associa- ized regions is expected to be at most sign and difficulties in exposure assess-

546 Rev Panam Salud Publica/Pan Am J Public Health 26(6), 2009 Menezes-Filho et al. • Manganese exposure and effects in children Literature review ment, the evidence of adverse effects comprehensive environmental assess- A. Menezes-Filho, who has a scholarship from Mn environmental exposure on ment in order not only to better under- from CAPES (Brazilian governmental children is compelling enough to war- stand the exposure pathways but also to agency). The project “Children’s expo- rant further research. Future investiga- provide reliable data for risk assess- sure to manganese from a ferro-man- tions based on a prospective design will ment, which can be used later to design ganese plant” was partially financed by shed more light on the relation of Mn ex- efficient interventions to abate exposure. the Fogarty International Center, Na- posure and neuropsychological effects tional Institutes of Health Grant 1 D43 on environmentally exposed children. TW000640. He is also a scholar of the Irv- Finally, it is of paramount importance Acknowledgments. This review is ing Selikoff International Foundation, that epidemiologic studies include a part of the doctoral thesis project of José Mount Sinai School of Medicine.

REFERENCES

1. Institute for Environment and Health/Insti- dence for cortical dysfunction and wide- the neonate. Pharmacol Ther. 2007;113(2): tute of Occupational Medicine. Occupational spread manganese accumulation in the non- 369–77. exposure limits: criteria document for man- human primate brain following chronic man- 26. Burton NC, Guilarte TR. Manganese neuro- ganese and inorganic manganese com- ganese exposure: a 1H-MRS and MRI study. toxicity: lessons learned from longitudinal pounds. Web report W17. Leicester, United Toxicol Sci. 2006;94(2):351–8. studies in nonhuman primates. Environ Kingdom: Medical Research Council, Insti- 14. Dobson AW, Erikson KM, Aschner M. Man- Health Perspect. 2009;117(3):325–32. tute for Environment and Health; 2004. Avail- ganese neurotoxicity. Ann N Y Acad Sci. 27. Pihl RO, Parkes M. Hair element content in able from: http://www.le.ac.uk/ieh. Ac- 2004;1012:115–28. learning disabled children. Science. 1977;198: cessed 25 March 2008. 15. Mergler D, Baldwin M. Early manifestations 204–6. 2. Institute of Medicine. Dietary reference in- of manganese neurotoxicity in humans: an 28. Collipp PJ, Chen SY, Maitinsky CS. Manga- takes for vitamin A, vitamin K, arsenic, update. Environ Res. 1997;73:92–100. nese in infant formulas and learning disabil- boron, chromium, copper, iodine, iron, man- 16. Pal PK, Samii A, Calne DB. Manganese neu- ity. Ann Nutr Metab. 1983;27:488–94. ganese, molybdenum, nickel, silicon, vana- rotoxicity: a review of clinical features, imag- 29. Barlow PJ. A pilot study on the metal levels in dium, and zinc (2000). Washington, DC: IOM; ing and pathology. Neurotoxicology. 1999;20 the hair of hyperactive children. Med Hy- 2001. (2–3):227–38. potheses. 1983;11:309–18. 3. World Health Organization. Manganese. En- 17. Fitsanakis VA, Au C, Erikson KM, Aschner 30. He P, Liu DH, Zhang GQ. Effects of high- vironmental health criteria 17. Geneva: WHO; M. The effects of manganese on glutamate, level-manganese sewage irrigation on chil- 1981. dopamine and gamma-aminobutyric acid dren’s neurobehavior. Zhonghua Yu Fang Yi 4. World Health Organization. Principles for regulation. Neurochem Int. 2006;48:426–33. Xue Za Zhi. 1994;28(4):216–8. evaluating health risks in children associated 18. Graham DG. Catecholamine toxicity: a pro- 31. Zhang G, Liu D, He P. A preliminary study of with exposure to chemicals. Environmental posal for the molecular pathogenesis of man- the effects of manganese on learning abilities health criteria 237. Geneva: WHO; 2006. ganese neurotoxicity and Parkinson’s disease. of primary school pupils. Zhonghua Yu Fang 5. Weiss, B. Economic implications of manga- Neurotoxicology. 1984;5:83–96. Yi Xue Za Zhi. 1995;29(3)157:156–8. nese neurotoxicity. Neurotoxicology. 2006;27: 19. Dorner K, Dziadzka S, Hohn A, Sievers E, 32. Woolf A, Wright R, Amarasiriwardena C, 362–8. Oldigs HD, Schulz-Lell G, et al. Longitudinal Bellinger D. A child with chronic manganese 6. Keen CL, Ensunsa JL, Clegg, MS. Manganese manganese and copper balances in young in- exposure from drinking water. Environ metabolism in animals and humans including fants and preterm infants fed on breast-milk Health Perspect. 2002;110(6):613–6. the toxicity of manganese. Met Ions Biol Syst. and adapted cow’s milk formulas. Br J Nutr. 33. Takser L, Mergler D, Hellier G, Sahuquillo 2000;37:89–121. 1989;61:559–72. J, Huel G. Manganese, monoamine metabo- 7. Papavasiliou PS, Miller ST, Cotzias GC. Role 20. Mena I, Horiuchi K, Burke K, Cotzias GC. lite levels at birth, and child psychomo- of liver in regulating distribution and excre- Chronic manganese poisoning: individual tor development. Neurotoxicology. 2003;24: tion of manganese. Am J Physiol. 1966;211(1): susceptibility and absorption of iron. Neurol- 667–74. 211–6. ogy. 1969;19:1000–6. 34. Torrente M, Colomina MT, Domingo JL. 8. Roth JA. Homeostatic and toxic mechanisms 21. Cotzias GC, Miller ST, Papavasiliou PS, Tang Metal concentrations in hair and cognitive as- regulating manganese uptake, retention, and LC. Interactions between manganese and sessment in an adolescent population. Biol elimination. Biol Res. 2006;39(1):45–57. brain dopamine. Med Clin North Am. 1976; Trace Elem Res. 2005;104:215–21. 9. Iregren A. Manganese neurotoxicity in indus- 60:729–38. 35. Wasserman GA, Liu X, Parvez F, Ahsan H, trial exposures: proof of effects, critical expo- 22. Cotzias GC, Horiuchi K, Fuenzalida S, Mena Factor-Litvak P, van Geen A, et al. Water ar- sure level, and sensitive tests. Neurotoxicol- I. Chronic manganese poisoning: clearance of senic exposure and children’s intellectual ogy. 1999;20:315–24. tissue manganese concentrations with persis- function in Araihazar, Bangladesh. Environ 10. Tanaka S. Manganese and its compounds. tence of the neurological picture. Neurology. Health Perspect. 2004;12:1329–33. In: Zenz C, ed. Occupational medicine: prin- 1968;18:376–82. 36. Wasserman GA, Liu X, Parvez F, Ahsan H, ciples and practical applications. Chicago: 23. Antonini JM, Santamaria AB, Jenkins NT, Al- Levy D, Factor-Litvak P, et al. Water man- Year Book Medical Publishers; 1988. Pp. bini E, Lucchini A. Fate of manganese associ- ganese exposure and children’s intellectual 583–9. ated with the inhalation of welding fumes: function in Araihazar, Bangladesh. Environ 11. Zoni S, Albini E, Lucchini R. Neuropsycho- potential neurological effects. Neurotoxicol- Health Perspect. 2006;114(1)124–9. logical testing for the assessment of man- ogy. 2006;27:304–10. 37. Wright RO, Amarasiriwardena C, Woolf AD, ganese neurotoxicity: a review and a pro- 24. Fitsanakis V, Zhang N, Avison MJ, Gore JC, Jim R, Bellinger DC. Neuropsychological cor- posal. Am J Ind Med. 2007;50(11):812–30. Aschner JL, Aschner M. The use of magnetic relates of hair arsenic, manganese, and cad- 12. Mergler D. Neurotoxic effects of low level ex- resonance imaging (MRI) in the study of mium levels in school-age children residing posure to manganese in human populations. manganese neurotoxicity. Neurotoxicology. near a hazardous waste site. Neurotoxicol- Environ Res. 1999;80:99–102. 2006;27(5):798–806. ogy. 2006;27:210–6. 13. Guilarte TR, McGlothan JL, Degaonkar M, 25. Erikson KM, Thompson K, Aschner J, Asch- 38. Bouchard M, Laforest F, Vandelac L, Bel- Chen M, Barker PB, Syversen T, et al. Evi- ner M. Manganese neurotoxicity: a focus on linger D, Mergler D. Hair manganese and hy-

Rev Panam Salud Publica/Pan Am J Public Health 26(6), 2009 547 Literature review Menezes-Filho et al. • Manganese exposure and effects in children

peractive behaviors: pilot study of school-age 43. Golub MS, Hogrefe CE, Germann SL, Tran Environmental health criteria 17. Geneva: children exposed through tap water. Environ TL, Beard JL, Crinella FM, et al. Neurobehav- WHO; 1981. Health Perspect. 2007;15(1):122–7. ioral evaluation of rhesus monkey infants fed 47. Aschner, M. Manganese: brain transport and 39. Ericson JE, Crinella FM, Clarke-Stewart KA, cow’s milk formula, soy formula, or soy for- emerging research needs. Environ Health Allhusen VD, Chan T, Robertson RT. Prenatal mula with added manganese. Neurotoxicol Perspect. 2000;108(3):429–32. manganese levels linked to childhood be- Teratol. 2005;27:615–27. 48. Elsner RJ, Spangler JG. Neurotoxicity of in- havioral disinhibition. Neurotoxicol Teratol. 44. VEEP (Veredas Estudos e Execução de Proje- haled manganese: public health danger in the 2007;29:181–7. tos Ltda). Technical Report: “Relatório Final shower? Med Hypotheses. 2005;65:607–16. 40. U.S. Geological Survey. National water- do Estudo Epidemiológico para Verificação 49. Lynam DR, Roos JW, Pfeifer GD, Fort BF, quality assessment program. Reston, VA: de Efeitos sobre a Saúde da População Cir- Pullin TG. Environmental effects and expo- USGS; 2005. Available from: http://water. cunvizinha à Empresa RDM”. Simões Filho: sures to manganese from use of methylcy- usgs. gov/nawqa/. Accessed 1 October 2005. Novembro 2005. clopentadienyl manganese tricarbonyl (MMT) 41. Krachler M, Rossipal E, Micetic-Turk D. Trace 45. Agency for Toxic Substances and Disease in gasoline. Neurotoxicology. 1999;20:145–50. element transfer from the mother to the new- Registry. Toxicological profile for manganese. 50. Rollin H, Mathe A, Levin J, Theodorou P, born—investigations on triplets of colo- Washington, DC: U.S. Department of Health Wewers F. Blood manganese concentrations strums, maternal and umbilical cord sera. Eur and Human Services, Public Health Service; among first-grade schoolchildren in two South J Clin Nutr. 1999;53:486–94. 2000. Available from: http://www.atsdr. African cities. Environ Res. 2005;97: 93–9. 42. Lonnerdal B, Keen CL, Ohtake M, Tamura cdc.gov/toxprofiles/tp151.html. Accessed T. Iron, zinc, copper and manganese in in- September 2000. fant formulae. Am J Dis Child. 1983;137: 46. World Health Organization. Manganese. In- Manuscript received on 18 March 2009. Revised version 433–7. ternational programme on chemical safety. accepted for publication on 28 July 2009.

RESUMEN Objetivo. El manganeso (Mn) es un elemento esencial, pero la sobreexposición puede tener efectos neurotóxicos. Métodos. En este artículo se hace una revisión y un compendio de los estudios pu- Exposición al manganeso blicados sobre la exposición al Mn y los trastornos del sistema nervioso en niños. y su efecto neurosicológico Resultados. Se identificaron 12 artículos originales publicados entre 1977 y 2007. La en niños y adolescentes: sobreexposición al Mn puede haber ocurrido a partir de diversas fuentes: leche en polvo o maternizada, agua de beber, polución industrial y desechos de la producción una revisión minera. El bioindicador de exposición utilizado con mayor frecuencia fue el contenido de Mn en el pelo, aunque algunos estudios lo midieron en la sangre, la orina o la dentina; un estudio sobre exposición prenatal midió su contenido en la sangre del cordón umbilical. La mayoría de los estudios indican que una mayor exposición pos- natal al Mn se asocia con deficiencias en las funciones cognitivas y el comportamiento hiperactivo. Conclusiones. Las limitaciones de los estudios publicados son numerosas: la mayo- ría de ellos eran transversales, se basaban en muestras pequeñas y en ellos no se ajus- taron los resultados por importantes factores de confusión. Se requieren investigacio- nes adicionales con muestras mayores y que hagan una evaluación más detallada de la exposición, tomando en cuenta múltiples fuentes, como los alimentos, el agua y las partículas suspendidas en el aire.

Palabras clave Manganeso; manifestaciones neurocomportamentales; cognición; intoxicación por manganeso.

548 Rev Panam Salud Publica/Pan Am J Public Health 26(6), 2009