3.2.1.1. Death

No studies were located regarding death in humans after inhalation of elemental selenium or selenium compounds.

In animals, the acute lethality of hydrogen selenide and elemental selenium dust when inhaled has been investigated. In guinea pigs exposed to hydrogen selenide for 2, 4, or 8 hours, 5/16 died within 10 days of exposure at 12 mg selenium/m³, 3/16 died at 6 mg selenium/m³, and 8/16 died at 6 mg selenium/m³, respectively (Dudley and Miller 1941).

No deaths were observed among rabbits or guinea pigs exposed to elemental selenium dust at levels of 31 mg selenium/m³ for 4 hours every other day for 8 exposure days (Hall et al. 1951). Higher levels were not tested.

Table 3-1

Levels of Significant Exposure to Selenium - Inhalation.

Figure 3-1

Levels of Significant Exposure to Selenium - Inhalation. Acute (≤14 days)

All LOAEL values from each reliable study for death in each species and duration category are recorded in Table 3-1 and plotted in Figure 3-1.

3.2.1.2. Systemic Effects

The selenium compounds that are most likely to be encountered in air in occupational settings are dusts of elemental selenium, hydrogen selenide, and selenium dioxide. Other volatile selenium compounds (e.g., dimethyl selenide, dimethyl diselenide) might be encountered in some naturally occurring situations. Because selenium is converted from one form to another, as in plant biosynthesis of selenoamino acids, it is not clear which specific forms may be encountered at hazardous waste sites. If a hazardous waste site specifically contains deposits of compounds of selenium, those compounds could be released off-site in dust or air. Toxicity data for exposures via inhalation are available for elemental selenium, selenium dioxide, selenium oxychloride, hydrogen selenide, and dimethyl selenide. Because there are few studies of inhalation of selenium of any single form, all available studies of inhalation exposures to selenium compounds will be included in this discussion.

In studies of human occupational exposures, it appears that the respiratory tract is the primary site of injury after inhalation of selenium dust or selenium compounds, but gastrointestinal (possibly due to swallowed selenium) and cardiovascular effects, as well as irritation of the skin and eyes, also occur. Little of the available information for humans, however, relates health effects exclusively to measured concentrations of the selenium dust or compounds because of the possibility of concurrent exposures to multiple substances in the workplace. In animals, the respiratory tract is also the primary site of injury following inhalation exposure to selenium dust and hydrogen selenide. Hematological and hepatic effects have also been noted in animals. Inhalation data from laboratory animal studies are available only for acute exposures.

No information was located regarding hematological, musculoskeletal, dermal, or ocular effects in humans or laboratory animals after inhalation exposure to selenium or selenium compounds. The systemic effects that have been observed after inhalation exposure are discussed below. The highest NOAEL values and all LOAEL values for each reliable study for systemic effects in each species and duration category are recorded in Table 3-1 and plotted in Figure 3-1.

Respiratory Effects. In humans, the respiratory system is the primary site of injury after inhalation of elemental selenium or selenium compounds. The largest number of reported human exposures occurred in occupational settings, especially in industries that extract, mine, treat, or process selenium-bearing minerals and in industries that use selenium or selenium compounds in manufacturing. The reports of occupational exposure do not link observed symptoms to specific air concentrations of elemental selenium or selenium compounds. Several reports, however, have noted common effects associated with inhalation exposure in occupational settings.

Selenium dioxide is formed when selenium is heated in air. Direct exposure to selenium dioxide is, therefore, primarily an occupational hazard and not likely to be a risk at hazardous waste sites. Selenium dioxide forms selenious acid on contact with water, including perspiration, and can cause severe irritation. Acute inhalation of large quantities of selenium dioxide powder can produce pulmonary edema as a result of the local irritant effect on alveoli (Glover 1970). Bronchial spasms, symptoms of asphyxiation, and persistent bronchitis have been noted in workers briefly exposed to high concentrations of selenium dioxide (Wilson 1962). Kinnigkeit (1962) reported that selenium dioxide concentrations of 0.007–0.05 mg selenium/m³ in a selenium rectifier plant produced slight tracheobronchitis in 9 of 62 exposed workers.

Hydrogen selenide, a highly poisonous selenium compound, is a gas at room temperature, with a density much higher than air. Selenium oxychloride, also highly toxic, is more irritating and corrosive to the human respiratory tract than are other forms of selenium because the compound hydrolyzes to hydrogen chloride (HCl), which can then form hydrochloric acid in humid air and in the respiratory tract (Dudley 1938). Hydrogen selenide and selenium oxychloride are occupational exposure hazards that are not expected to be much of a concern at hazardous waste sites.

Acute inhalation exposure to elemental selenium dust, possibly including some selenium dioxide, in occupational settings has been shown to irritate mucous membranes in the nose and throat and produce coughing, nosebleed, loss of olfaction, and in heavily exposed workers, dyspnea, bronchial spasms, bronchitis, and chemical pneumonia (Clinton 1947; Hamilton 1949). Chronic exposure of 40 workers at a copper refinery produced increased nose irritation and sputum (Holness et al. 1989). The exact concentration of selenium was not given, but the concentration was reported to exceed 0.2 mg selenium/m³. Confounding variables in this study include concurrent exposure to several other metals including copper, nickel, silver, lead, arsenic, and tellurium.

In experimental animals, the respiratory tract is the primary site of injury following acute inhalation exposure to elemental selenium and selenium compounds. Rats exposed to selenium fumes (selenium concentration and particle size were not reported) for 2–16 minutes experienced moderate to severe respiratory effects, including hemorrhage and edema of the lungs (Hall et al. 1951). Rats exposed to selenium dust (average particle diameter, 1.2 µm) at levels of 33 mg selenium/m³ for 8 hours experienced severe respiratory effects, including hemorrhage and edema of the lungs, and several animals died (Hall et al. 1951). Histopathological examinations of surviving animals revealed chronic interstitial pneumonitis. Acute exposure of rabbits and guinea pigs to selenium dust (average particle diameter, 1.2 µm) at a concentration of 33 mg selenium/m³ resulted in mild interstitial pneumonitis or congestion, and slight emphysema in both species (Hall et al. 1951). Other histological findings included vascular lymphocytic infiltration and intra-alveolar foci of large macrophages.

Acute inhalation exposure of guinea pigs to 8 mg selenium/m³ as hydrogen selenide for 4 hours produced diffuse bronchopneumonia and pneumonitis (Dudley and Miller 1941). The investigators do not indicate if any of these guinea pigs died as a result of the exposure. Histologic examination of animals that died following exposure to higher concentrations revealed thickening of the alveolar walls and congestion of alveolar capillaries (Dudley and Miller 1937). In contrast, 1-hour exposure of rats to 25,958 mg selenium/m³ as dimethyl selenide produced only minor effects (increased weight of lung and liver) 1 day postexposure. These changes disappeared by 7 days postexposure (Al-Bayati et al. 1992). Enzymatic methylation of selenium compounds is the primary route of detoxification and may explain the low toxicity of dimethyl selenide (Al-Bayati et al. 1992). Although this form of selenium is environmentally relevant since it is formed in soil, plants, and microorganisms, dimethyl selenide appears to be relatively nontoxic in comparison to occupational exposure to hydrogen selenide.

The effects of intratracheal instillation of selenium on pulmonary function may be dependent on the form in which it is supplied (Nonavinakere et al. 1999). Instillation of 0.06 mg selenium/100 g body weight as selenium dioxide produced a significant decrease in respiratory rate and a significant increase in lung resistance compared with controls. Instillation with 0.06 mg selenium/110 g body weight as seleno-L-methionine also produced a decrease in respiratory rate and an increase in lung resistance, but the values were not significantly different from controls.

Intratracheal instillation of 0.3 mg selenium as sodium selenite in male Hartley-guinea pigs decreased dynamic-lung-compliance and increased pulmonary resistance compared with control animals instilled with saline (Bell et al. 1997). Analysis of bronchoalveolar-lavage fluid showed increased activities of lactate dehydrogenase, β-glucuronidase, alkaline phosphatase, and protein, suggesting damage to lung tissue.

Histological analysis of guinea pigs that received single intratracheal instillations of 0.3 mg selenium as sodium selenite found mild acute inflammation in approximately one-third of the lung tissue and a noticeable amount of sloughed epithelium and mucus within the bronchi (Bell et al. 2000). Lungs of animals treated with 0.06 mg selenium showed neutrophils aggregated in the alveoli and some dilation of the alveoli suggestive of emphysema. Relative lung weights and the ratio of wet/dry lung weight were increased in the selenium-treated animals compared with controls; the increase was only significant for those receiving the higher dose of selenium. Leukocyte counts in bronchoalveolar-lavage fluid were decreased for selenium-treated animals compared with controls, and the difference was significant for the animals receiving 0.3 mg selenium, but not the 0.06 mg dosage.

No studies were located regarding respiratory effects in animals after intermediate or chronic inhalation of selenium or selenium compounds.

Cardiovascular Effects. Several workers experienced symptoms of shock, including lower blood pressure and elevated pulse rates, following an acute exposure (at most 20 minutes) to selenium dioxide fumes resulting from a fire (Wilson 1962). The subjects were treated with oxygen and inhalation of ammonia vapor, and pulse rates were normalized within 3 hours.

Cardiovascular effects were not observed in guinea pigs exposed to hydrogen selenide at 8 mg selenium/m³ for 4 hours (Dudley and Miller 1941), or in guinea pigs and rabbits exposed to elemental selenium dust (average particle diameter, 1.2 µm) every other day at 33 mg selenium/m³ for eight 4-hour exposure periods (Hall et al. 1951).

Gastrointestinal Effects. Vomiting and nausea were reported in workers exposed to high concentrations of selenium dioxide for a maximum of 20 minutes during a fire (Wilson 1962). Stomach pain was frequently reported by workers exposed to elemental selenium and selenium dioxide at a selenium rectifier plant (Glover 1967), and by copper refinery workers exposed to an unspecified form of selenium (Holness et al. 1989). Exposure concentrations were not reported for the rectifier plant, but were >0.2 mg selenium/m³ at the copper refinery.

No studies were located regarding gastrointestinal effects in animals after inhalation of selenium or selenium compounds.

Hepatic Effects. No studies were located regarding hepatic effects in humans after inhalation of selenium or selenium compounds.

Hepatoxicity has been observed in experimental animals following inhalation exposure to elemental selenium dust and to hydrogen selenide. One month after an 8-hour exposure to elemental selenium dust at a level of 33 mg selenium/m³, most rats exhibited slight liver congestion and a few exhibited mild centrilobular atrophy (Hall et al. 1951). In contrast, 1 week after exposure to 25,958 mg selenium/m³ as dimethyl selenide for 1 hour, rats showed no observable changes in the liver (Al-Bayati et al. 1992). Three weeks following acute exposure to elemental selenium dust at a level of 33 mg selenium/m³ for 4 hours every other day for 8 days, 4/10 guinea pigs exhibited slight hepatic congestion with mild central atrophy and 2/10 showed some fatty hepatocellular degeneration (Dudley and Miller 1941). In contrast, exposure of guinea pigs to lower concentrations of selenium (8 mg/m³), as hydrogen selenide, for a single 4-hour period produced mild fatty hepatocellular metamorphosis (Dudley and Miller 1941).

Renal Effects. No studies were located regarding renal effects in humans after inhalation of selenium or selenium compounds.

The kidneys do not appear to be affected in guinea pigs (Dudley and Miller 1941; Hall et al. 1951) after acute inhalation exposure to 33 mg selenium/m³ as hydrogen selenide for 8 hours or to 8 mg selenium/m³ as elemental selenium dust for 4 hours. Likewise, the kidneys were not affected in rabbits following acute inhalation exposure to 33 mg selenium/m³ as hydrogen selenide for 8 hours (Hall et al. 1951) or in rats following acute inhalation exposure to 25,958 mg selenium/m³ as dimethyl selenide for 1 hour or to 33 mg selenium/m³ as hydrogen selenide for 8 hours (Al-Bayati et al. 1992; Hall et al. 1951).

Endocrine Effects. No studies were located regarding endocrine effects in humans after inhalation of selenium or selenium compounds.

No histopathological changes in the adrenal gland were observed in guinea pigs exposed to hydrogen selenide at 8 mg selenium/m³ for 4 hours (Dudley and Miller 1941) or in rats exposed to elemental selenium at 33 mg selenium/m³ for 8 hours (Hall et al. 1951).

Body Weight Effects. No studies were located regarding effects on body weight in humans following inhalation of selenium or selenium compounds.

No effects on body weight were observed in guinea pigs following a single 8-hour exposure to elemental selenium at 33 mg selenium/m³ or in guinea pigs and rabbits exposed to elemental selenium dust at 33 mg selenium/m³ every other day for 4 hours for a total of eight exposures (Hall et al. 1951).

3.2.1.3. Immunological and Lymphoreticular Effects

No studies were located regarding immunological or lymphoreticular effects in humans after inhalation exposure to selenium or selenium compounds.

Lymphoid hyperplasia was noted in the spleen of guinea pigs following a single 4-hour exposure at 8 mg selenium/m³ as hydrogen selenide (Dudley and Miller 1941). Histopathological changes in the spleen were not observed in guinea pigs exposed to elemental selenium dust (average particle diameter, 1.2 µm) at 33 mg selenium/m³ for 8 hours (Hall et al. 1951). Injury to the spleen was observed in guinea pigs following exposure for 4 hours, every other day, for 8 days to elemental selenium dust at a level of 33 mg selenium/m³ (Hall et al. 1951). Specific effects included congestion of the spleen, fissuring red pulp, and increased polymorphonuclear leukocytes (Hall et al. 1951).

3.2.1.4. Neurological Effects

Information concerning possible neurological effects caused by inhalation of selenium or selenium compounds is limited. Severe frontal headaches were reported by workers exposed during an accident to high concentrations of selenium fumes (compound not stated) for approximately 2 minutes (Clinton 1947). Workers at a selenium rectifier plant reported symptoms of malaise and irritability when working with selenium (exposure was probably to selenium dioxide and elemental selenium, but the form was not stated) (Glover 1967). The symptoms resolved whenever the workers were moved to other work. Urinary concentrations of selenium were about 0.08 mg/L, compared to 0.024–0.034 mg/L in unexposed workers.

No studies were located regarding neurological effects in animals after inhalation of selenium or selenium compounds.

No studies were located regarding the following health effects in humans or animals after inhalation exposure to selenium or selenium compounds:

3.2.1.5. Reproductive Effects

3.2.1.6. Developmental Effects

3.2.1.7. Cancer

There are no epidemiologic data that support a causal association between the inhalation of elemental selenium dusts or selenium compounds and the induction of cancer in humans (Gerhardsson et al. 1986; Wester et al. 1981). In one study, postmortem samples were collected from copper smelter workers who were exposed to several different airborne compounds, including selenium compounds. Samples from lung cancer cases had lower concentrations of selenium in lung tissue than samples from controls or from workers who had died from other causes (Gerhardsson et al. 1986). In another autopsy study of smelter workers, Wester et al. (1981) found that the selenium concentrations in kidney tissues from workers who had died of malignancies were lower than the selenium concentrations in kidney tissues from workers who died of other causes. Further discussions regarding the cancer protective effects of selenium can be found in Section 3.2.2.7.

No studies were located regarding carcinogenic effects in laboratory animals after inhalation exposure to selenium or selenium compounds.

3.2.2.1. Death

Accidental selenium poisonings in humans have occurred, but few fatalities have been reported. The selenium doses associated with the reported deaths are unknown (Carter 1966; Koppel et al. 1986). One 3-year-old boy died 1.5 hours after ingestion of an unknown quantity of selenious acid contained in a gun-bluing preparation (Carter 1966). Clinical signs included excessive salivation, garlic odor on the breath, and shallow breathing. A 15-year-old female survived ingestion of a solution of sodium selenate estimated to have provided 22 mg selenium/kg body weight, probably because she was forced to vomit soon after exposure (Civil and McDonald 1978). Clinical signs included garlic odor of the breath and diarrhea.

No cases of human death in the United States have been attributed to intermediate or chronic oral exposures to selenium or selenium compounds. In the Hubei Province of China, in an area of endemic selenosis, a woman who died was suffering from hemiplegia thought to have been caused by chronic selenosis induced by eating locally grown foods that contained high levels of organic selenium compounds (Yang et al. 1983). However, an autopsy was not performed and no clinical history of previous illness was available.

In nonhuman animals, the most acutely toxic selenium compounds by ingestion appear to be sodium selenite and sodium selenate (Olson 1986). Oral LD₅₀ values for sodium selenite, expressed as mg selenium/kg body weight, were reported as 4.8–7.0 in rats, 1.0 in rabbits, 3.2 in mice, and 2.3 in guinea pigs (Cummins and Kimura 1971; Pletnikova 1970). Minimum lethal doses of sodium selenite, expressed as mg selenium/kg body weight, reported for larger animals were 13–18 for pigs and 9.9–11.0 for cows (Miller and Williams 1940); however, these values were estimated on the basis of a small number of animals. Two of four 12-week-old lambs died within 16 hours of administration of 5 mg selenium/kg as sodium selenite (Smyth et al. 1990). Selenium dioxide is reported to have LD₅₀ values of 16 mg selenium/kg for mice and 48 mg selenium/kg body weight for rats, but these values are also based on a small number of animals (Singh and Junnarkar 1991). An oral LD₅₀ of 35.9 mg selenium/kg has been reported for L-selenocystine given to mice (Sayato et al. 1993). Elemental selenium is less toxic than most selenium compounds, because of its extremely low solubility; an LD₅₀ of 6,700 mg selenium/kg body weight has been reported for oral administration of elemental selenium as a suspension (particle size 1–30 µm) in 0.5% methylcellulose to rats (Cummins and Kumura 1971).

Lower doses of selenium can cause signs of toxicity if administered over extended periods of time. Eight weaned 5-week-old pigs receiving 1.3 mg selenium/kg/day as sodium selenite in gelatin capsules daily for 10 days died during one study; only one dose level was tested (Wilson et al. 1989). Two long-tailed macaques administered 0.60 mg selenium/kg/day as selenomethionine by nasogastric intubation died of either anorexia or aspirated vomitus secondary to emesis and gastritis after 10 or 15 days of treatment (Cukerski et al. 1989). Seven of 12 female rats receiving diets containing 0.418 mg selenium/kg/day as sodium selenate for 14 days died before the end of the experiment (NTP 1996). Exposure to selenium in drinking water at a level of 0.84 mg selenium/kg/day as selenite or selenate for 4–6 weeks resulted in the death of four of six or two of six male rats, respectively (Palmer and Olson 1974). Feeding male rats diets containing 0.48 mg selenium/kg/day as sodium selenite or 0.4 mg selenium/kg/day as seleniferous wheat for 6 weeks resulted in the death of one of eight rats in each group (Halverson et al. 1966). Administration of sodium selenite in drinking water at a level of 0.28 mg selenium/kg/day for 58 days resulted in the death of 25 of 50 male rats (Schroeder and Mitchener 1971a). Mortality was observed in rats, but not in mice, receiving either 1.67 mg selenium/kg/day as sodium selenite or 2.54 mg selenium/kg/day as sodium selenate in drinking water for 13 weeks (NTP 1994). Gavage treatment of male mice with selenocystine 6 days per week for 30 days at a dose of 14.2 mg selenium/kg killed all 15 treated animals, while no deaths were noted at 9.4 mg selenium/kg (Sayato et al. 1993). The longevity of hamsters was not affected by dietary administration of sodium selenite at a dose of 0.42 mg selenium/kg/day for 124–144 weeks (Birt et al. 1986).

Sodium selenate and sodium selenite exhibit similar toxicity in female rats, but male rats appear more susceptible to the toxicity of sodium selenite than selenate (Palmer and Olson 1974; Schroeder and Mitchener 1971a). Sodium selenate in drinking water at 0.28 selenium mg/kg/day for 1 year did not increase mortality of male or female rats compared with control rats (Schroeder and Mitchener 1971a). Ingestion of 0.28 mg selenium/kg/day of sodium selenite in drinking water for 1 year did not increase mortality in female rats, whereas 50% of the males died by day 58 of administration (Schroeder and Mitchener 1971a).

The relative acute toxicities of sodium selenite, potassium selenite, sodium selenate, and potassium selenate in aqueous solution have been examined in mice (Pletnikova 1970). No significant differences among the toxicities of the potassium and sodium salts of selenium were apparent in this study. In another study, rats tolerated a dose of 1.05 mg selenium/kg/day administered in drinking water as potassium selenate for over 8 months with no deaths, but three of five females and one of three males died by the end of 1 year (Rosenfeld and Beath 1954). Decreased survival was reported in rats fed sodium selenate or selenite at 0.5 mg selenium/kg/day in a 2-year cancer study (Harr et al. 1967; Tinsley et al. 1967). No mortality was observed in hamsters fed 0.42 mg selenium/kg/day as sodium selenite in the diet for 82–142 weeks (Birt et al. 1986).

Selenium sulfide (i.e., selenium monosulfide) and selenium disulfide are less water soluble and are of lower acute toxicity than sodium selenate or sodium selenite. There are no reported human deaths due to ingestion of selenium sulfide. The LD₅₀ value for the gavage administration of 1–20% selenium disulfide in aqueous 0.5% methylcellulose to rats was 138 mg selenium disulfide/kg (Cummins and Kimura 1971). When 1% selenium disulfide shampoo was administered by gavage, the LD₅₀ value was lower (78 mg selenium disulfide/kg) (Cummins and Kimura 1971). The compound administered may have been a mixture of selenium sulfide and selenium disulfide; analysis of the compound was not reported. Henschler and Kirschner (1969) reported an LD₅₀ of 3,700 mg selenium sulfide/kg for mice administered by gavage in aqueous 0.5% carboxymethylcellulose. Administration of single gavage doses of selenium monosulfide to rats produced death in 3/15 animals dosed with 50 mg/kg, 3/6 animals dosed with 75 mg/kg, 1/2 animals dosed with 100 mg/kg, and 2/2 animals dosed with 125 mg/kg (Moore et al. 1996b).

In the case of selenium sulfide, mice are more tolerant than rats, and males of both species appear to be more tolerant than females (NTP 1980c). The daily doses producing 50% mortality for a 17-day gavage administration of a mixture of selenium mono- and disulfides were 112 mg selenium sulfides/kg for male rats, 56 mg selenium sulfides/kg for female rats, and 805 mg selenium sulfides/kg for male mice (NTP 1980c). A 13-week gavage study using the same mixture of selenium mono- and disulfides reported survival as 10/10, 10/10, 10/10, 9/9, 8/9, and 6/10 in female mice and 10/10, 10/10, 10/10, 10/10, 10/10, and 9/10 in male mice receiving 0, 21.6, 46.4, 100, 216, and 464 mg selenium sulfides/kg/day, respectively (NTP 1980c). Although the researchers intended to administer selenium monosulfide to the animals, elemental analysis, melting point, and x-ray diffraction revealed that the compound administered included some selenium disulfide. No other chemical or physical analyses of the selenium compound administered were reported.

The LD₅₀ and lethal LOAEL values from each reliable study following oral exposure to elemental selenium dust, selenium dioxide dissolved in water (selenious acid), sodium selenate, sodium selenite, potassium selenate, and dietary selenium for each species and exposure duration are recorded in Table 3-2 and plotted in Figure 3-2. The LOAEL values for death in rats and mice following acute and intermediate oral exposures to selenium sulfide or selenium disulfide are recorded in Table 3-3 and plotted in Figure 3-3.

3.2.2.2. Systemic Effects

The highest NOAEL value and all LOAEL values for each reliable study for systemic effects in each species and duration category are recorded in Tables 3-2 and 3-3 and plotted in Figures 3-2 and 3-3.

Respiratory Effects. Pulmonary edema and lesions of the lung have been noted in case reports of humans (Carter 1966; Koppel et al. 1986) and animals (Glenn et al. 1964a; Rosenfeld and Beath 1947) after ingestion of lethal doses of selenium compounds. Rabbits orally administered sodium selenite (subroute not specified) at levels approximating the LD₅₀ (1–5 mg selenium/kg body weight) developed pulmonary congestion, hemorrhages, and edema; dyspnea; general muscular weakness; and asphyxial convulsions (Smith and Westfall 1937). Pulmonary edema and hemorrhages were observed in four sheep treated orally (subroute not specified) with a single dose of sodium selenite of 5 mg selenium/kg (Smyth et al. 1990). The lungs may be a target of acute exposure to excess selenium because the metabolite, dimethyl selenide, is exhaled.

The effects of intermediate or chronic exposures to selenium compounds are less clear. Although Harr et al. (1967) stated that absolute lung weights decreased with increasing doses of selenite or selenate chronically administered to rats in the diet in a 2-year cancer study, they did not report lung weights at specific dose levels. Selenium administration also might have contributed to pneumonic lesions, but again, the authors did not statistically analyze their results or relate the severity of the effect to the doses of selenium administered. Respiratory effects were not observed in rats treated with selenite in the diet for 8 weeks at a dose of 0.45 mg selenium/kg/day (Chen et al. 1993). Effects on the lungs were not observed in pigs fed 1.25 mg selenium/kg as organic selenium found in the plant Astragalus bisulcatus for up to 5 days, or D,L-selenomethionine or selenate in the diet for up to 6 weeks (Panter et al. 1996). Treatment of steers with selenomethionine or selenite in food at doses up to 0.808 mg selenium/kg/day for 120 days did not produce any signs of respiratory distress or changes in lung weight or histology (O’Toole and Raisbeck 1995). Ingestion of selenium in drinking water for 13 weeks at doses up to 1.67 and 7.17 mg selenium/kg as selenate in rats and mice, respectively, and 1.57 and 3.83 mg selenium/kg as selenite in rats and mice, respectively, did not cause any respiratory effects (NTP 1994). Nelson et al. (1943) reported that no effects on the lungs were apparent in rats administered 0.50 mg selenium/kg/day as seleniferous corn for 2 years.

An increased incidence of amyloidosis of the major organs, including the lungs, was observed in mice following lifetime exposure to sodium selenate or sodium selenite in drinking water at a level of 0.57 mg selenium/kg/day (Schroeder and Mitchener 1972). This effect was noted in 30% of control mice and 58% (p>0.001) of selenium-treated mice. Data for individual organs were not provided.

Administration of lethal doses of selenium sulfide particles in carboxymethylcellulose by gavage has been reported to cause irregular breathing in mice (Henschler and Kerschner 1969), but not in rats (Cummins and Kimura 1971). No respiratory effects were seen in mice administered 464 mg selenium sulfides/kg/day or in rats administered 31.6 mg selenium sulfides/kg/day by gavage once daily for 13 weeks (NTP 1980c).

Cardiovascular Effects. Tachycardia has occasionally been reported as a result of a lethal, acute oral exposure to selenium compounds in humans (Carter 1966); however, the dose was not reported in this lethal exposure to a gun-bluing solution containing selenious acid. Although myocardial disorders (cardiogenic shock, congestive heart failure, arrhythmia, multifocal necrosis of the myocardium) have been associated with selenium deficiencies (Yang et al. 1988), none has been reported to be associated with chronic dietary selenosis in humans observed at doses of ≥0.016 mg/kg/day (Yang et al. 1989a). A preliminary study completed in China suggests that selenium supplementation (100 µg/day, form not stated) during pregnancy may reduce the incidence of pregnancy-induced hypertension (Li and Shi-mei 1994).

In contrast, postmortem studies of sheep that died from acute oral exposure to sodium selenite or sodium selenate have revealed petechial hemorrhages of the endocardium (Glenn et al. 1964a; Smyth et al. 1990). The sheep were treated with a time-weighted average dose of 0.65 or 0.9 mg selenium/kg/day as selenate over a 171-day period (Glenn et al. 1964a, 1964b), or a single dose of selenite at 5 mg selenium/kg (Smyth et al. 1990). Vacuolation and pyknosis of nuclei were observed in the hearts of pigs fed an unspecified form of selenium at a dose of 0.46 mg selenium/kg/day for 34 days (Stowe et al. 1992). In a 2-year cancer study, Harr et al. (1967) reported the occurrence of myocardial hyperemia, hemorrhage, and degeneration, as well as pericardial edema, in young rats administered sodium selenite or sodium selenate in the feed at doses of 0.5 mg selenium/kg/day, although the authors did not specify the duration of exposure required to produce the effects.

Exposure of pigs to feed containing 54 mg/kg selenium for 1–7 days resulted in severe toxicity and death of several of the animals (Penrith and Robinson 1996). Histological examination of heart tissue from pigs that died revealed myocardial lesions consisting of widespread hypertrophy, atrophy, and disorganization of fibers, occasional fibrosis, and marked medial hypertrophy of the arterioles.

Wistar rats administered 0.324 mg selenium/kg/day as sodium selenite in food for 12–14 weeks showed severe diffuse degenerative changes, including edema in the sub-endocardial connective tissue and the interfibers of prevascular regions, and myofibril swelling with profuse intercellular edema (Turan et al. 1999a). Myocyte borders were irregular, and there was a loss of striations and a degeneration of the sarcolemma and myofibril structure and order. Examination of the mechanical function of the heart in vitro using either Langendorff perfusion or papillary muscle recordings showed increased coronary perfusion pressure, increased resting force, and increased heart rate with irregular beating. No difference in contractile force was observed. Chronic heart failure did not occur in any of the animals in the study.

Cardiac damage was also observed in mice exposed to 0.2 mg selenium/kg/day as sodium selenite in food for 12 weeks (Skowerski et al. 1997b). Ultrastructural examination revealed cardiomyocytes that had numerous damaged mitochondria, a large number of lipid droplets, and numerous lysosomes.

Hearts of New Zealand white rabbits administered 0.137 mg selenium/kg/day as sodium selenite in food for 3 months showed distinct, degenerative changes indicating disintegration of the internal structure of the myocytes (Turan et al. 1999b). Muscle fibers were fragmented and separated. Disruption and loss of myofibrils was observed, sarcomeres were irregular, and the I, Z, and H bands were disorganized and discontinuous. Mitochondria were fewer and more variable in size and shape, with disoriented cristae and a loss of matrix substance. Hearts of control animals (0.007 mg selenium/kg/day) had normal histology.

Treatment of steers with selenomethionine or selenite in food at doses up to 0.808 mg selenium/kg/day for 120 days did not produce any changes in heart weight or histology (O’Toole and Raisbeck 1995). Histopathological changes in the heart were not observed in pigs fed selenium at 1.25 mg selenium/kg as organic selenium found in the plant A. bisulcatus for up to 5 days, or D,L-selenomethionine or selenate for up to 6 weeks (Panter et al. 1996). Histopathological changes were not observed in the hearts of rats treated with selenite in the diet for 8 weeks at a dose of 0.45 mg selenium/kg/day (Chen et al. 1993). Selenium administered to rats and mice in drinking water for 13 weeks at doses up to 1.57 and 7.17 mg selenium/kg/day as selenate, respectively, and up to 1.67 and 3.83 mg selenium/kg/day as selenite, respectively, did not cause any histopathological changes in the heart tissue (NTP 1994). No histopathological changes were noted in mice administered 464 mg selenium sulfides/kg/day or in rats administered 31.6 mg selenium sulfides/kg/day by gavage once daily for 13 weeks (NTP 1980c). An increased incidence of amyloidosis of the major organs, including the heart, was observed in mice following lifetime exposure to sodium selenate or sodium selenite in drinking water at a level of 0.57 mg selenium/kg/day (Schroeder and Mitchener 1972). This effect was noted in 30% of control mice and 58% (p>0.001) of selenium-treated mice. Data for individual organs were not provided.

Although myocardial degeneration and necrosis have been experimentally induced in laboratory animals and livestock including cattle, sheep, and swine by acute and longer-term exposures to inorganic salts of selenium, it is unclear whether seleniferous grains or forages, or other natural sources of selenium, can also cause cardiomyopathy (Raisbeck 2000).

Gastrointestinal Effects. In humans, gastrointestinal distress, including nausea, vomiting, diarrhea, and abdominal pain, has been reported following ingestion of aqueous sodium selenate (Civil and McDonald 1978; Gasmi et al. 1997; Helzlsouer et al. 1985; Koppel et al. 1986; Sioris et al. 1980). Two studies provided an estimate of dose. In a case report by Civil and McDonald (1978), diarrhea was observed in a 15-year-old girl about 45 minutes after she swallowed sheep drench containing selenate at a dose of about 22 mg selenium/kg. This effect was observed despite the induction of vomiting shortly after the exposure. In a second case report of a suicide attempt, a 56-year-old man reported that vomiting, diarrhea, and abdominal pain occurred 1 hour after he ingested approximately 11 mg/kg selenium as sodium selenite (Gasmi et al. 1997). Postmortem examinations following two deaths from selenium ingestion revealed dilation of the stomach and small intestine (Carter 1966) and erosive changes of the gastrointestinal tract (Koppel et al. 1986). High (unspecified) levels of dietary selenium compounds have been implicated as causing gastrointestinal disturbances in chronically exposed humans (Smith et al. 1936), but such symptoms are not specific to selenium intoxication.

Exposure of pigs to feed containing 54 mg/kg selenium for 1–7 days resulted in severe toxicity and death of several animals. Clinical signs included anorexia and vomiting, and histological examination (70–79 days after exposure) of three of the exposed animals that died found lesions ranging from small erosions (1–2 mm diameter) to extensive mucosal necrosis (up to 100 mm diameter) near the cardia of the stomach (Penrith and Robinson 1996).

Gross necropsy of steers that died after ingestion of sodium selenite revealed severe gastrointestinal irritation (Baker et al. 1989; Maag et al. 1960). In addition, cattle and other livestock exhibiting alkali disease, perhaps as a result of long-term consumption of range plants high in selenium, ate and drank less and suffered from ulcers in the upper intestinal tract (Shamberger 1986). A single oral dose of 5 mg selenium/kg as selenite caused edema and congestion of abdominal viscera in lambs (Smyth et al. 1990). However, treatment of steers with selenomethionine or selenite in food at doses up to 0.808 mg selenium/kg/day for 120 days did not produce any changes in histology of the gastrointestinal tissues (O’Toole and Raisbeck 1995).

Gastrointestinal effects were not observed in rats treated with selenite in the diet for 8 weeks at a dose of 0.45 mg selenium/kg/day (Chen et al. 1993). Selenium treatment in drinking water for 13 weeks at doses up to 1.57 and 7.17 mg selenium/kg/day as selenate in rats and mice, respectively, and 1.67 and 3.83 mg selenium/kg/day as selenite in rats and mice, respectively, did not cause any gastrointestinal effects (NTP 1994). Gastrointestinal effects were not observed in rats fed organic selenium (seleniferous corn or wheat) at 0.5 mg selenium/kg/day for 24 months (Nelson et al. 1943). Vomiting and anorexia were reported in monkeys receiving 0.15 mg/kg/day selenium as L-selenomethionine by oral intubation during gestation days 20–50 (Tarantal et al. 1991).

Selenium sulfide administration by gavage at lethal levels has been reported to cause diarrhea and anorexia in rats (Cummins and Kimura 1971). No gastrointestinal effects were seen in mice administered 464 mg selenium sulfide/kg/day or in rats administered 31.6 mg selenium sulfide/kg/day by gavage once daily for 13 weeks (NTP 1980c).

Hematological Effects. Hematological changes were evaluated in a 120-day double blind study of healthy men who consumed a controlled diet of foods naturally low or high in selenium (Hawkes et al. 2001). Eleven subjects were fed 0.0006 mg selenium/kg/day in the diet for 21 days (baseline period), followed by 0.0002 mg/kg/day (6 subjects) or 0.004 mg/kg/day (5 subjects) for the following 99 days. Complete blood counts (white blood cells, lymphocytes, granulocytes, platelets, erythrocytes, hematocrit) and hemoglobin concentration measurements showed no adverse effects of selenium supplemenation. Mean within-subject changes from baseline in white blood cell counts were significantly different in the low- and high-selenium groups at last two time points in the study (days 70 and 99). At the end of the study, the white blood cell counts were decreased by 5% in the high-selenium group and increased by 10% in the low-selenium group, due mainly to changes in numbers of granulocytes. Lymphocyte counts were significantly increased in the high-selenium group on day 45, but not at the end of the study. There were no clear effects of selenium on numbers of activated or cytotoxic T-cells, lymphocyte phenotypes, serum immunoglobulins, or complement fractions, as summarized in Section 3.2.2.3 (Immunological Effects).

Increased prothrombin time was reported for individuals chronically exposed to estimated dietary doses of 0.016 mg selenium/kg/day in a high-selenium region of China (Yang et al. 1989a). However, no increase in prothrombin time was found in another study of individuals consuming diets that supplied up to 0.0098 mg/kg/day selenium (Longnecker et al. 1991). A study that compared children from seleniferous and nonseleniferous areas of Venezuela found slightly reduced (no statistical analysis was performed) hemoglobin levels and hematocrit values for the children from the seleniferous area (Jaffe et al. 1972). However, the children from the seleniferous zone had a poorer diet, consumed less milk and meat, and had a greater incidence of intestinal parasites, which may account for the differences observed. Red blood cell counts were significantly increased in mice that received drinking water containing 9 ppm (0.82 mg selenium/kg/day) selenium as sodium selenite for 14 days (Johnson et al. 2000). However, these mice also had a severe reduction in water consumption (43%) and this may have led to a decrease in blood volume. No significant increase in red blood cell count (or decrease in water consumption) was observed for mice receiving 3 ppm (0.38 mg selenium/kg/day) selenium as sodium selenite, or up to 9 ppm (1.36 mg selenium/kg/day) selenium as selenomethionine for 14 days (Johnson et al. 2000).

No hematological changes (hemoglobin concentration, hematocrit, erythrocyte count, and cell volume) were reported for male Sprague-Dawley rats fed diets providing up to 0.27 mg selenium/kg/day as sodium selenite for 40 days (Eder et al. 1995). Increased hematocrit was observed in rats treated with selenate (1.56 mg selenium/kg/day) or selenite (1.67 mg selenium/kg/day) in the drinking water for 13 weeks, but only at concentrations that decreased water intake (NTP 1994). No effects on hematology end points were observed in mice treated with selenate or selenite in drinking water for 13 weeks at 7.17 mg selenium/kg for selenate and 3.83 mg selenium/kg/day for selenite (NTP 1994).

No differences in blood cell counts or hematological parameters were found in rabbits administered 0.137 mg selenium/kg/day as sodium selenite in the diet for 3 months, compared with control animals receiving a normal laboratory diet (Turan et al. 1999b).

A dose-related decrease in hematocrit was observed in rats fed seleniferous wheat (Halverson et al. 1966). Compared to controls, hemoglobin was decreased 23 and 79% at 0.32 and 0.56 mg selenium/kg/day, respectively. Hemoglobin reductions were most evident in the animals that had died during the experiments. In a 2-year cancer study, Harr et al. (1967) reported that the hemoglobin concentration decreased by 0.5 g/100 mL with each 2-fold increase of sodium selenate in the diet, but did not specify the lowest dose at which hemoglobin concentrations were significantly reduced compared to the controls (the range of selenium doses used was 0.025–0.40 mg selenium/kg/day). Hematocrit was increased in rats given selenite and selenate in drinking water for 13 weeks at concentrations that also resulted in decreased water intake (NTP 1994). No hematological effects were noted in rats or mice treated with selenate at 0.92 and 7.17 mg selenium/kg/day, respectively, or selenite at 0.86 and 3.83 mg selenium/kg/day, respectively (NTP 1994).

No studies were located regarding hematological effects in humans or other animals after oral exposure to selenium sulfide or selenium disulfide.

Musculoskeletal Effects. No adverse musculoskeletal effects were reported following chronic oral exposure of humans to dietary levels of selenium of up to 0.0098 mg selenium/kg/day (Longnecker et al. 1991).

A single oral (subroute not specified) dose of sodium selenite (5 mg selenium/kg/day) caused edema in skeletal muscles of the diaphragm in sheep (Smyth et al. 1990). Exposure of pigs to feed containing 54 mg/kg selenium for 1–7 days resulted in severe toxicity and death of several animals (Penrith and Robinson 1996). Histological examination of skeletal muscle from animals that died found damage with interstitial oedema and diffuse swelling of fibers. Livestock suffering from chronic alkali disease, a disease once common in the southwestern United States where selenium levels are high, showed lameness due to joint erosion and hoof deformation (Shamberger 1986). However, treatment of steers with selenomethionine or selenite in food at doses up to 0.808 mg selenium/kg/day for 120 days did not produce any changes in muscle or bone histology (O’Toole and Raisbeck 1995). Hyperplasia of the sarcolemma nuclei and disintegration of myofibrils were observed in the skeletal muscles of pigs fed an unspecified form of selenium for 34 days (Stowe et al. 1992). In a 2-year cancer study, Harr et al. (1967) fed graded doses of selenium in the form of sodium selenate or selenite to rats and reported frank osteotoxicity at doses as low as 0.2 mg selenium/kg/day given for several months (duration specified as less than 100 days). Selenium administered in drinking water for 13 weeks at doses up to 1.57 and 7.17 mg selenium/kg/day as selenate in rats and mice, respectively, and 1.67 and 3.83 mg selenium/kg as selenite in rats and mice, respectively, failed to cause adverse musculoskeletal effects (NTP 1994). Musculoskeletal effects were not observed in rats fed seleniferous corn or wheat at 0.5 mg selenium/kg/day for 24 months (Nelson et al. 1943). No musculoskeletal effects were seen in mice administered 464 mg selenium sulfide/kg/day or in rats administered 31.6 mg selenium sulfide/kg/day by gavage once daily for 13 weeks (NTP 1980c).

Hepatic Effects. Limited data suggest that hepatotoxicity can occur in humans following acute oral exposure to sodium selenate, but no definitive studies were located regarding hepatic effects in humans after intermediate or chronic oral exposure to selenium compounds. Tests following an acute poisoning of a 15-year-old girl with sodium selenate revealed abnormally elevated serum bilirubin and alkaline phosphatase (Civil and McDonald 1978). Hepatic effects, such as changes in serum liver enzymes or liver morphology (identified by ultrasonography), have not been observed in humans at chronic dietary intakes of 0.0098 mg selenium/kg/day (Longnecker et al. 1991) or 0.025 mg selenium/kg/day (Yang et al. 1989a). Selenium-induced hepatoxicity is documented in animals as summarized below. The lack of evidence of liver damage in humans due to selenosis, despite the animal data to the contrary, suggests a problem with the animal models of the disease.

Congestion and/or edema and hemorrhage in the liver have been reported in sheep following the acute oral (subroute not specified) administration of lethal levels of sodium selenate (Hopper et al. 1985) or sodium selenite (Smyth et al. 1990) and in mules and pigs following administration of lethal levels of sodium selenite (Miller and Williams 1940). A significant decrease in relative liver weight was reported for mice exposed to 9 ppm (0.82 mg selenium/kg/day) selenium as sodium selenite in drinking water for 14 days, but not to 3 ppm (0.38 mg selenium/kg/day) (Johnson et al. 2000). No effect on liver weight was observed for mice receiving up to 9 ppm (1.36 mg selenium/kg/day) selenium as selenomethionine in drinking water for 14 days (Johnson et al. 2000).

Administration of single gavage doses of selenium monosulfide to rats produced death and widespread hepatic necrosis in 3/6 animals dosed with 75 mg/kg, 1/2 animals dosed with 100 mg/kg, and 2/2 animals dosed with 125 mg/kg (Moore et al. 1996b).

Hepatic effects have also been reported following intermediate-duration exposure in pigs, but not in cattle. Pigs exposed for 7 weeks to either dietary organic selenium in dried plants (either A. pruelongus or A. bisulcatus) or sodium selenate (at 1.1 or 1.3 mg selenium/kg/day) exhibited diffuse swelling and vacuolar degeneration of hepatocytes (Baker et al. 1989). The doses used in this study reduced mean survival to only 44 days. Pigs exposed to sodium selenite in feed for 35 days at doses less than half as high as those tested by Baker et al. (1989) (0.47 versus 1.1 or 1.3 mg selenium/kg/day) exhibited no liver damage (Mahan and Magee 1991). A study of pigs treated with 0.08, 0.33, 0.59, or 1.07 mg selenium/kg/day as sodium selenite in the feed for 8 weeks found hepatic nodules/granules in two pigs treated with 0.59 or 1.07 mg selenium/kg/day (Mihailovic et al. (1992). The lesions were diagnosed as postdystrophic atrophic cirrhosis. However, only these two severely affected pigs (one from each of the highest dose groups of 40 animals each) were selected for histopathological examination. Hepatic effects were not observed in pigs fed selenium at 1.25 mg selenium/kg as organic selenium of the type(s) found in the plant A. bisulcatus for up to 5 days, or D,L-selenomethionine or selenate for up to 6 weeks (Panter et al. 1996). Treatment of steers with selenomethionine or selenite in food at doses up to 0.808 mg selenium/kg/day for 120 days did not produce any changes in liver weight or histology (O’Toole and Raisbeck 1995).

Alterations or cirrhosis of the liver in experimental animals following intermediate or chronic oral exposure to selenium compounds have been reported by Bioulac-Sage et al. (1992), Fitzhugh et al. (1944), Halverson et al. (1970), Harr et al. (1967), Kolodziejczyk et al. (2000), Nelson et al. (1943), and Schroeder and Mitchener (1972). Halverson et al. (1966) reported reduced liver-to-body-weight ratios and increased bilirubin in rats administered 0.44 mg selenium/kg/day for 6 weeks as naturally occurring selenium in wheat. At this level, five of eight rats died. At a dose of 0.84 mg selenium/kg/day administered as sodium selenate in drinking water for 4–6 weeks, rats developed cirrhosis of the liver (Palmer and Olson 1974). At this level, two of six rats died.

Hepatic damage was observed in mice exposed to 0.2 mg selenium/kg/day as sodium selenite in food for 12 weeks (Skowerski et al. 1997a), and ultrastructural examination showed that the cytoplasm of the hepatocytes contained extremely large and irregularly-shaped vacuoles. Wistar rats administered 0.324 mg selenium/kg/day as sodium selenite in food for 12–14 weeks showed degenerative changes to the liver (not fully described in text) (Turan et al. 1999a). Livers of rats fed 0.002 or 0.005 mg selenium/kg/day as sodium selenite for 3 months showed damage that increased with dose (Kolodziejczyk et al. 2000). Rats from the 0.002 mg selenium/kg/day group had a distinct swelling of Küpffer cells in dilated sinusoidal vessels, mainly in the proximity of portal fields, and occasional necrotic areas comprising groups of hepatocytes, while livers from rats receiving 0.005 mg selenium/kg/day showed activation and swelling of the Küpffer cells in widened sinusoidal vessels, relatively abundant infiltrations of mononuclear cells into portal canals, and sporadic areas of necrosis within individual lobules.

Young rats treated with sodium selenite in the feed for 2 months had nodular hyperplasia at a dose of 0.2 mg selenium/kg/day. However, clinical tests of liver function (bilirubin, alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, and gamma-glutamyltransferase activities) showed no significant changes (Bioulac-Sage et al. 1992). Diffuse panlobular vacuolar changes were reported in rats fed sodium selenite in the diet for 8 weeks at 0.45 mg/kg/day (Chen et al. 1993).

In a 2-year cancer study, acute toxic hepatitis was common among rats fed sodium selenite or sodium selenate at 0.25 mg selenium/kg/day or higher (Harr et al. 1967; Tinsley et al. 1967). Liver surfaces were mottled, and parenchymatous degeneration was present. Hepatic lesions occurred at a dose as low as 0.10 mg selenium/kg/day. Absolute liver weights decreased with increasing levels of sodium selenate or sodium selenite in the diet. The average liver weight of animals administered selenate (14.5 g) was twice the average liver weight of animals administered selenite (7.2 g); however, the average liver weight of control animals was not reported, and possible dose-related hepatic effects were not discussed by these authors.

Increased serum bile acids, suggesting cholestasis, were observed in rats treated with 1.57 mg selenium/kg/day as sodium selenate in drinking water for 13 weeks, but no effects were noted at 0.92 mg/kg/day (NTP 1994). In a 13-week drinking water study, hepatic effects were not observed in mice treated with sodium selenate at 7.17 mg selenium/kg/day, in mice treated with sodium selenite at doses up to 3.83 mg selenium/kg/day, or in rats treated with sodium selenite at doses up to 1.67 mg selenium/kg/day (NTP 1994). Increased serum aspartate aminotransferase and alanine aminotransferase activities were observed in mice treated by gavage with selenocystine at doses of 9.4 mg selenium/kg/day for 30 days (Sayato et al. 1993) or 4.7 mg selenium/kg/day for 90 days (Hasegawa et al. 1994). No effects on liver enzymes were observed in mice treated with selenocystine at 4.7 mg selenium/kg/day for 30 days (Sayato et al. 1993) or at 2.5 mg selenium/kg/day for 90 days (Hasegawa et al. 1994). Chronic dietary administration of selenium as seleniferous corn or wheat at doses ranging from 0.25 to 0.50 mg/kg/day for 24 months produced cirrhosis of the liver in rats (Nelson et al. 1943).

An increased incidence of amyloidosis of the major organs, including the liver, was observed in mice following lifetime exposure to sodium selenate or sodium selenite in drinking water at a level of 0.57 mg selenium/kg/day (Schroeder and Mitchener 1972). This effect was noted in 30% of control mice and 58% (p>0.001) of selenium-treated mice. Data for individual organs were not provided.

Selenium sulfide administered to rats daily by gavage for 13 weeks produced focal coagulation necrosis in the liver with infiltration by inflammatory cells. These changes developed at a dose of 31.6 mg selenium sulfide/kg/day, but not at a dose of 17.8 mg selenium sulfide/kg/day (NTP 1980c). In mice, on the other hand, oral intubation of selenium sulfide at 464 mg selenium sulfide/kg/day did not produce hepatic effects (NTP 1980c).

Renal Effects. No studies were located regarding renal effects in humans after oral exposure to selenium or selenium compounds.

In domestic and experimental animals, renal effects have been observed following both acute and chronic oral exposures to selenium compounds. Administration of a single oral (subroute not specified) dose of sodium selenite at 5 mg selenium/kg/day produced hydropic degeneration of the kidney in sheep (Smyth et al. 1990). In a study of the toxicity of L-selenomethionine to long-tailed macaques by nasogastric intubation, two animals administered 0.24 mg selenium/kg/day aspirated vomitus secondary to emesis, developed obvious gastritis, and died of anorexia, one after 10 days and the other after 15 days of administration (Cukierski et al. 1989). Histopathologic examination of the kidneys of these animals revealed glomerulonephritis and proximal convoluted tubule nephropathy. The study authors indicated that these changes were consistent with macaque fatal fasting syndrome and may not have resulted from the direct effects of L-selenomethionine. Following long-term ingestion of plants high in selenium, livestock suffering from alkali disease exhibited nephritis (Shamberger 1986). However, treatment of steers with selenomethionine or selenite in food at doses up to 0.808 mg selenium/kg/day for 120 days did not produce any changes in kidney weight or histology (O’Toole and Raisbeck 1995).

A dose-related increase in degeneration of the renal papilla (described as mild to minimal) was observed in rats treated with selenate or selenite in the drinking water at about 0.5 mg selenium/kg/day for 13 weeks (NTP 1994). No evidence of renal toxicity was observed in rats given 0.3 mg selenium/kg/day in this study. In contrast to rats, the only kidney effect noted in mice treated with sodium selenate or selenite in the drinking water was increased relative kidney weight (NTP 1994). This effect, which occurred at 1.87 mg selenium/kg/day as selenate and 1.61 mg selenium/kg/day as selenite, was only noted at doses at which drinking water intake was decreased, leading the investigators to suggest that the effect may have been a result of dehydration. A similar increase in relative kidney weight associated with decreased water consumption was observed in mice consuming approximately 0.38 mg selenium/kg/day as selenite in drinking water, but no effect on kidney weight or water consumption was observed in mice consuming up to 1.36 mg selenium/kg/day as selenomethionine (Johnson et al. 2000). No renal effects were observed in pigs fed selenium at 1.25 mg selenium/kg as organic selenium found in the plant A. bisulcatus for up to 5 days, or D,L-selenomethionine or selenate for up to 6 weeks (Panter et al. 1996). No effects on the kidneys were observed in rats treated with selenite in the diet for 8 weeks at a dose of 0.45 mg selenium/kg/day (Chen et al. 1993). Gavage treatment of mice with selenocystine for 30 days at a dose of 9.4 mg selenium/kg/day had no adverse effect on the kidneys (Sayato et al. 1993).

Rats chronically fed selenite in the diet were reported to exhibit more frequent and more severe nephritis than those given equivalent amounts of selenate (Harr et al. 1967); however, the study authors did not quantify these observations or statistically compare data from the two groups. An increased incidence of amyloidosis of the major organs, including the kidneys, was observed in mice following lifetime exposure to sodium selenate or sodium selenite in drinking water at a level of 0.57 mg selenium/kg/day (Schroeder and Mitchener 1972). This effect was noted in 30% of control mice and 58% (p>0.001) of selenium-treated mice. Data for individual organs were not provided.

A mixture of selenium sulfide and selenium disulfide administered to mice daily by gavage for 13 weeks at a dose of 464 mg selenium sulfides/kg/day produced an increase in the incidence and severity of interstitial nephritis compared with the controls, whereas a daily dose of 216 mg selenium sulfides/kg did not elicit renal toxicity (NTP 1980c). In rats, selenium sulfide by oral intubation at 31.6 mg selenium sulfides/kg/day for 13 weeks did not produce renal effects (NTP 1980c).

Endocrine Effects. A balance in selenium and iodine levels is needed for normal thyroid hormone metabolism. Selenium is an essential component of the iodothyronine 5’-deiodinase enzymes, which convert the prohormone thyroxine (T₄) to the active form, triiodothyronine (T₃) (Delange 2000; Köhrle 1994; St Germain and Galton 1997). Selenium is also a component of glutathione peroxidase (GPX), the main enzyme responsible for protecting thyroid cells against oxidative damage. Selenium deficiency causes decreases in metabolic clearance of iodothyronines, extrathyroidal production of T₃, and thyroid iodine concentrations in experimental animals (Arthur and Beckett 1989, 1994; Behne and Kyriakopolous 1993). Deficiency in both selenium and iodine has been associated with goiter and cretinism in humans and causes thyroid gland necrosis and fibrosis in rats (Delange 2000; Goyens et al. 1987; Vanderpas et al. 1990). Additional information on thyroid effects of selenium and iodine deficiency is discussed in Section 3.9. Thyroid hormone levels in humans and animals can also be affected by selenium supplementation; these effects include decreases in serum T₃ and T₄ levels and increases in serum TSH levels, suggesting suppression of thyroid hormone production, as discussed below.

A limited amount of information is available regarding endocrine effects in humans following oral exposure to selenium. Serum levels of thyroid and reproductive hormones were evaluated in a double blind 120-day study of healthy men (20–45 years old) who consumed a controlled diet of foods naturally high or low in selenium (Hawkes and Turek 2001). Eleven subjects were fed 0.0006 mg/kg/day of selenium in the diet for the first 21 days of the study, followed by 0.0002 µg selenium/kg/day (6 subjects) or 0.004 mg selenium/kg/day (5 subjects) for 99 days. Blood samples were analyzed for serum levels of selenium, thyroid hormones (T₃ and TSH), and reproductive hormones (testosterone, follicle-stimulating hormone, luteinizing hormone, prolactin, estradiol, and progesterone) during week 3 (baseline), week 17 (ending value), and at several interim time points.

Selenium levels in blood plasma began to change within 3 days of starting the low- and high-selenium diets and progressively continued throughout the study (Hawkes and Turek 2001). By week 17, mean plasma selenium concentrations had increased by 109% in the high-selenium group and decreased by 38.5% in the low-selenium group. Group mean serum T₃ concentrations (averages of within-subject changes from baseline) were significantly different in the low-selenium subjects and high-selenium subjects at all time points, but the changes are insufficient to be considered adverse as discussed below. In the low-selenium group, serum T₃ levels increased an average of 14 and 8% from baseline at weeks 8 and 17, respectively. In the high-selenium group, serum T₃ levels decreased an average of 23 and 11% from baseline at weeks 8 and 17, respectively. Analysis of variance (ANOVA) showed a significant main effect of dietary selenium on serum T₃ concentrations, as well as a significant selenium x time interaction, indicating that the changes in T₃ levels decreased over time. Although the decreases in serum T₃ in the high-selenium group and increases in serum T₃ in the low-selenium group lessened in magnitude during the study, all group mean values appear to have remained within the normal range. The baseline and week 17 serum total T₃ values (mean±SD) were 1.82±0.36 and 1.57±0.07 nmol/L in the high-selenium group, and 1.57±0.25 and 1.64±0.16 nmol/L in the low-selenium group, compared to the normal human range of 1.1–2.7 nM/L (Stockigt 2000), indicating that the changes in serum T₃ were subclinical and not toxicologically significant. Serum TSH concentrations increased significantly by 32% over its baseline concentration in the high-selenium group, but did not change significantly in the low-selenium group. Baseline and ending mean TSH values in the high-selenium group were 2.25±0.81 and 2.96±1.05 mU/L, respectively, both of which are in the normal range of 0.3–4.0 mU/L (Stockigt 2000). The lack of clinically significant changes in serum T₃ and TSH values is not surprising because the study was designed as a nutritional study and not as a toxicological study; the selenium intakes bracketed the current recommended dietary allowance (RDA) (~0.8 µg Se/kg/day) and were well below the tolerable upper limit level (~5.7 µg Se/kg/day) recommended by the Food and Nutrition Board (NAS 2000). There were no significant changes in serum levels of free or total testosterone, follicle-stimulating hormone, luteinizing hormone, prolactin, estradiol, or progesterone. This study also found no adverse immunologic or male reproductive changes as summarized in Sections 3.2.2.3 and 3.2.2.5.

An examination of thyroid hormone levels in lactating women residing in areas of Venezuela with high levels of selenium in the soil (selenium intake ranged from 250 to 980 µg per day as estimated from selenium content of breast milk) revealed a significant decrease in serum T₃ levels, as compared with women having normal selenium intakes (90–350 µg/day), but these hormone levels remained within the normal range (Brätter and Negretti De Brätter 1996). Additionally, a significant inverse correlation for selenium and serum T₃ concentration was found using the Spearman Rank test. The study authors noted that the effect of selenium on T₃ levels became significant at dietary intake levels of 350–450 µg/day. No significant alterations in serum T₄ or TSH levels or correlations with selenium intake were found.

Twenty weeks of selenium supplementation (10, 20, 30, or 40 µg/day) of New Zealanders who normally consume a diet low in selenium (unsupplemented intake of 28–29 µg/day), but show no signs of deficiency, produced a reduction in T₄ concentration in all groups (Duffield et al. 1999). However, only the differences between the 10 µg-group and controls and the combined supplemented individuals and controls were significant. T₃ and TSH levels were not measured. Thyroglobulin concentration did not change significantly with supplementation.

In a study of 68 male Latvian fish consumers (Hagmar et al. 1998), a significant inverse correlation was found between serum levels of selenium and TSH. No correlation was found between serum selenium concentration and the serum concentrations of T₃ or T₄. No measurements were made of dietary selenium intake.

Selenium supplementation has been shown to affect type-I-deiodinase activity in male rats (Behne et al. 1992; Eder et al. 1995; Hotz et al. 1997). Exposure to 0.055 or 0.27 mg selenium/kg/day as sodium selenite in food for 40 days produced a significant decrease (approximately 50%) in serum levels of T₃ and a nonsignificant reduction in type-I-deiodinase activity compared with rats receiving 0.009 or 0.026 mg selenium/kg/day (Eder et al. 1995). Exposure to 0.27 mg selenium/kg/day did not produce any other adverse signs, such as weight loss or decreased food consumption, and serum T₄ levels were similar in all groups.

Exposure of weanling male Sprague-Dawley rats to 0.09 mg selenium/kg/day as sodium selenate in food for 6 weeks produced a significant (~30%) increase in TSH, compared with controls receiving 0.009 mg selenium/kg/day (Hotz et al. 1997). Serum T₃ and T₄ levels and thyroid glutathione peroxidase levels were unaffected by dietary selenium. Kidney type-I-deiodinase levels were decreased (~10%) in high selenium animals compared with controls, but the differences were not significant, and liver type-I-deiodinase levels were unaffected by dietary selenium. Iodine-deficient diets produced greater thyroid glutathione peroxidase activity at each dietary level of selenium, and the greatest activity was in rats with high selenium.

No significant changes in thyroid levels of T₃ or T₄ were found in male Wistar rats fed diets containing high selenium (0.105 mg selenium/kg/day as sodium selenite or 0.118 mg selenium/kg/day as L-selenomethionine) for 3 months, compared with controls receiving adequate selenium (0.0015 mg selenium/kg/day as sodium selenite) (Behne et al. 1992). However, rats eating the high selenium diet showed a significant reduction in hepatic type I deiodinase activity, compared with controls, with a 29% reduction in the production rate of T₃ from T₄ and a 45% reduction in the production rate of 3,3’-diiodothyronine from T₄.

Many studies have documented reduced body weight gain in young animals treated with selenium compounds, and abnormal weight loss in older animals (Grønbaek et al. 1995; Halverson et al. 1966; Harr et al. 1967; Jacobs and Forst 1981a; Johnson et al. 2000; Nelson et al. 1943; NTP 1994; Palmer and Olson 1974; Panter et al. 1996; Schroeder 1967; Tarantal et al. 1991; Tsunoda et al. 2000). There is evidence to suggest that these effects may be due in part to the interactions of selenium or selenium compounds with hormones that regulate normal growth and body weight. In a 14-day study suggesting that selenium may inhibit pituitary function, Thorlacius-Ussing (1990) treated nursing rats with sodium selenite in drinking water (0.64 or 0.96 mg/kg/day). The resulting decrease in the body weight gain of the pups observed at both doses may be associated with a reduction in somatomedin C levels (no other hormone levels were tested), and the weight deficiency could be reversed by administration of a growth hormone. Postweanling female Wistar rats treated with sodium selenite (0.64 mg selenium/kg/day) in drinking water for 3 or 6 weeks exhibited decreased weight gain and decreased somatomedin C serum concentrations. When the selenium supplement was removed after 3 weeks, body weight gain returned to normal, but the serum somatomedin C concentrations did not return to control levels. Growth hormone secretion in response to growth hormone releasing factor was also reduced in the selenium-exposed group (Thorlacius-Ussing et al. 1988). Serum somatomedin C levels were not significantly different among three exposure categories (>200, 201–240, and >240 ng selenium/mL) in 44 long-term residents of seleniferous areas in South Dakota, despite >50% differences in serum, whole blood, and toenail selenium levels among the groups (Salbe et al. 1993). A 10% reduction in body weight and a reduction in tibia lengths, compared to pair-fed controls, were found in rats provided with sodium selenite in the drinking water at 0.46 mg selenium/kg/day for 35 days (Grønbaek et al. 1995). A significant reduction in insulin-like growth factor-binding protein-3 was also noted. The investigators concluded that the reduction in growth caused by excess selenium is not due to reduced caloric intake.

Selenium administered in drinking water for 13 weeks at doses up to 1.57 and 7.17 mg selenium/kg/day as selenate in rats and mice, respectively, and 1.67 and 3.83 mg selenium/kg as selenite in rats and mice, respectively, failed to cause changes in the weights or histology of the thyroid, adrenal glands, parathyroid, or pancreas (NTP 1994).

Lambs given a single oral (subroute not specified) dose of 5 mg selenium/kg as sodium selenite exhibited cytoplasmic flocculation of the pancreas (Smyth et al. 1990). Increased pancreas weights were observed in rats fed organic selenium (seleniferous wheat) at a dose of 0.4 mg selenium/kg/day for 6 weeks (Halverson et al. 1966). Chronic exposure of rats fed sodium selenite or sodium selenate in their diet for a lifetime was associated with pancreatic damage. Although Harr et al. (1967) reported a dose-related increase in the incidence and severity of pancreatic lesions in treated rats, they did not specify the lowest dose at which pancreatic lesions were observed.

Treatment of steers with selenomethionine or selenite in food at doses up to 0.808 mg selenium/kg/day for 120 days did not produce any changes in weight or histology of the pancreas, adrenal glands, thyroid, or pituitary gland (O’Toole and Raisbeck 1995).

An increased incidence of amyloidosis of the major organs, including the adrenal gland, was observed in mice following lifetime exposure to sodium selenate or sodium selenite in drinking water at a level of 0.57 mg selenium/kg/day (Schroeder and Mitchener 1972). This effect was noted in 30% of control mice and 58% (p>0.001) of selenium-treated mice. Data for individual organs were not provided.

Dermal Effects. Jensen et al. (1984) described both marked alopecia and the deformity and loss of fingernails in a woman who had consumed a selenium supplement containing 31 mg total selenium (in the form of sodium selenite and elemental selenium) per tablet for 77 days. The woman consumed one tablet each day in addition to vitamin supplements (vitamins C, A, D, E, B complex) and a mineral supplement “labeled as containing all 72 trace elements in undefined quantities.” In epidemiological studies of populations chronically exposed to high levels of selenium in food and water, investigators have reported discoloration of skin, pathological deformity and loss of nails, loss of hair, and excessive tooth decay and discoloration (Smith et al. 1936; Yang et al. 1983, 1989a, 1989b). The 1989 studies by Yang et al. follow up their original 1983 study of Chinese populations living in areas classified as having low-, medium-, and high-selenium exposure based on local soils and food supplies. The average and standard error of selenium intakes in the low-, medium-, and high-intake regions were 0.0012±0.00009, 0.0037±0.0004, and 0.025±0.001 mg/kg/day, respectively. The whole blood (average ± standard error) concentrations of selenium in the low-, medium-, and high-intake regions were 0.16±0.00, 0.35±0.02, and 1.51±0.05 mg/L, respectively. The estimated daily dietary selenium intake required to produce these symptoms in an area of China characterized by endemic selenosis was at least 0.016 mg selenium/kg/day (Yang et al. 1989a). This corresponds to a blood concentration of 1.054 mg/L and an estimated daily intake of 0.91 mg/day, assuming a 55-kg Chinese man or woman and using the regression analysis provided by Yang et al. (1989b). The NOAEL from the highest intake population not affected by nail disease is 0.015 mg selenium/kg/day, which corresponds to a blood concentration of 0.97 mg/L. Foods that contributed the greatest levels of selenium were smoked pork, coal-dried corn, chestnuts, pumpkin seeds, dried fruits, and garlic. It has been noted that the selenosis problem in China began when coal with high levels of selenium was burned as the main source of fuel (Whanger 1989). Food was cooked and dried over the open flame, adding selenium to the food. In addition, the people breathed large amounts of smoke, but the contribution of volatilized selenium to the total dose of selenium has not been adequately characterized (Whanger 1989). Coal was also burned on the fields as a fertilizer source. Environmental selenium concentrations in the low-, medium-, and high-intake regions were 0.37–0.48, 0.73–5.66, and 7.06–12.08 mg/kg in soil, and 370, 1,720, and 12,270 µg/L in water, respectively (Yang et al. 1989b).

No evidence of nail disease was observed in a population living on selenium-rich ranches in the western United States (Longnecker et al. 1991). Doses of selenium were calculated to be between 0.001 and 0.01 mg/kg/day, corresponding to a maximum intake of 0.724 mg/day. Whole blood selenium concentrations were 0.18–0.67 mg/kg. Although these values for the United States are consistent with studies of the Chinese population, only one or a few individuals ingested the highest doses.

The highest selenium intake for villagers in a high-selenium area of China in which endemic selenosis did not occur was estimated at 1.51 mg selenium/person/day (0.027 mg selenium/kg/day), with the average dietary selenium intake in this area of selenosis occurrence estimated to be 3.2 mg selenium/person/day (0.058 mg selenium/kg/day) (Yang et al. 1983). The lowest daily dietary selenium intake associated with dermal effects, 0.91 mg selenium/day, was converted to equivalent daily doses from food (0.016 mg/kg/day) for presentation in Table 3-2.

Five individuals from the high selenium region of China described by Yang et al. (1989a) who had been diagnosed with overt signs of selenosis (hair loss and nail sloughing) in 1986 were reexamined in 1992 (Yang and Zhou 1994). The results of this examination showed that these individuals had recovered from selenosis (overt symptoms of nail sloughing were absent) and that the average selenium concentrations in their blood had fallen from 1,346 to 968 µg/L. The corresponding dietary intakes of selenium were 1,270 and 819 µg/day. This study has been used to establish a LOAEL of 0.023 mg selenium/kg/day and a NOAEL of 0.015 mg selenium/kg/day. Based on the occurrence of these dermal effects, a chronic oral MRL of 0.005 mg selenium/kg/day has been derived from the NOAEL, as described in the footnote in Table 3-2 and detailed in Appendix A. This MRL is approximately 6 times greater than the NAS (2000) RDA for selenium of 55 µg/day (~0.0008 mg/kg/day).

In a 30-day study of oral administration of L-selenomethionine to long-tailed macaques, skin lesions appeared on the forearm of one of two macaques given 0.01 mg selenium/kg/day. However, the limited number of animals precludes identifying the dose as a LOAEL for dermal effects (Cukierski et al. 1989). Pigs receiving dietary administration of the same doses of selenium for 35 days exhibited hoof cracking (Mahan and Magee 1991). Symmetrical hair loss, dry scaling skin, and cracked overgrown hooves were observed in one of five pigs and three of five pigs fed sodium selenate or D,L-selenomethionine at a dose of 1.25 mg selenium/kg/day for up to 6 weeks, respectively (Panter et al. 1996). In an experiment limited to a duration of 5 days because of severe paralysis, similar dermal effects were not observed in pigs fed 1.25 mg selenium/kg/day as selenium contained in the plant A. bisulcatus. The form of selenium in A. bisulcatus is unknown, although Panter et al. (1996) indicate that it is nonprotein.

Exposure of pigs to feed containing 54 mg/kg selenium for 1–7 days resulted in severe toxicity and death of several animals; however, none of the pigs developed coronitis or hoof separation (Penrith and Robinson 1996). Skin from four pigs with alopecia was examined about a month after exposure and was found to have epidermal thickening due to acanthosis and hyperkeratosis, vacuolar degeneration of the basal cells and acanthocytes, necrosis of individual keratinocytes, and serocellular crusts.

In the late 19^th and the early 20^th century, livestock grazing on plants growing on seleniferous soils in areas of the Great Plains of the United States suffered from alkali disease attributed to the high selenium content of some plants. Alkali disease in horses, cattle, and swine is characterized by alopecia, inflammation at the coronary band, followed by cracked or malformed hooves and rough hair coat (Draize and Beath 1935). Daily selenium intakes associated with these effects were not quantified. However, treatment of steers with selenomethionine in food at doses of 0.288 mg selenium/kg body weight/day or selenite at doses of 0.808 mg selenium/kg/day for 120 days produced hoof lesions (O’Toole and Raisbeck 1995). In intermediate-duration studies, cracked hoof walls have been observed in pigs fed selenate, selenite, or an unspecified form of selenium at doses of 0.25 mg selenium/kg/day and greater (Baker et al. 1989; Mahan and Magee 1991; Mihailovic et al. 1992; Wahlstrom and Olson 1959b). Poor quality of the hair coat has also been reported in mice administered sodium selenite or selenate in the diet at 0.57 mg selenium/kg/day (Schroeder and Mitchener 1972). Exposure of female BALB/c mice to 0.21 mg selenium/kg/day for 6 months from diets containing selenium as sodium selenite resulted in alopecia around the nose (Boylan et al. 1990).

No studies were located regarding dermal effects in humans or other animals after oral exposure to selenium sulfide or selenium disulfide.

Ocular Effects. A case-control study using a hospital discharge register indicated that there was no correlation between low serum selenium concentrations and cataract occurrence in humans (Knekt et al. 1992). Since this is a case-control study, it does not provide information on the potential dietary factors, exposure to specific selenium compounds, or duration of exposure.

Treatment of steers with selenomethionine or selenite in food at doses up to 0.808 mg selenium/kg/day for 120 days did not produce any changes in the histology of the eyes (O’Toole and Raisbeck 1995). Selenium given to rats and mice in drinking water for 13 weeks at up to 1.6 and 7.2 mg selenium/kg as selenate, respectively, or 1.7 and 3.8 mg selenium/kg as selenite, respectively, did not cause any ocular effects (NTP 1994).

Body Weight Effects. Two studies reported body weight effects in humans after oral exposure to selenium. Selenium intake was found to affect body weight in a study of 11 men (20–45 years old) who were fed 0.0006 mg/kg/day of selenium in the diet for the first 21 days of the study, followed by diets naturally low (0.0002 µg selenium/kg/day, 6 subjects) or high (0.004 mg selenium/kg/day, 5 subjects) for 99 days at 2,800 kcal/day (Hawkes and Keim 1995). Despite minor adjustments of intake to maintain body weight, by the 6^th week, the high selenium group started to gain weight relative to the low selenium group, and the difference between the two groups became significant after the 10th week. A similar increase in lean body mass was observed in both groups. The study was designed as a nutritional study and not as a toxicological study, as the selenium intake levels were well below the tolerable upper limit level (~5.7 µg Se/kg/day) recommended by the Food and Nutrition Board (NAS 2000). The weight gain observed in this study therefore has nothing to do with weight loss due to selenosis.

A study that compared children from seleniferous and nonseleniferous areas of Venezuela found slightly reduced height and weight (no statistical analysis was performed) for the children from the seleniferous area (Jaffe et al. 1972). However, the children from the seleniferous zone had a poorer diet, consumed less milk and meat, and had a greater incidence of intestinal parasites, which may account for the differences observed.

In contrast, reduced growth rates of young animals and reduced body weight in older animals are common observations associated with oral administration of excess sodium selenate, sodium selenite, or organic selenium compounds to experimental animals (Boylan et al. 1990; Cukierski et al. 1989; Donaldson and McGowan 1989; Grønbaek et al. 1995; Halverson et al. 1966; Harr et al. 1967; Hasegawa et al. 1994; Johnson et al. 2000; Nelson et al. 1943; NTP 1994, 1996; Palmer and Olson 1974; Panter et al. 1996; Penrith and Robinson 1996; Raisbeck et al. 1996; Sayato et al. 1993; Schroeder 1967; Tarantal et al. 1991; Thorlacius-Ussing 1990, Tsunoda et al. 2000; Turan et al. 1999a). This reduction in growth is often accompanied by reduced food and water consumption, and in dietary or drinking water studies, may be an effect of poor palatability of selenium compounds. However, reduced growth has also been observed in gavage studies (Cukierski et al. 1989; Hasegawa et al. 1994; Sayato et al. 1993) and, as discussed under endocrine and neurological effects, the growth retardation may have an endocrine or neurotransmitter component. Selenium effects on the levels of thyroid hormones (Behne and Kyriakopoulos 1993; Behne et al. 1992; Eder et al. 1995; Hotz et al. 1997), dopamine metabolites (Tsunoda et al. 2000), insulin-like growth factor–binding protein-3 (Grønbaek et al. 1995), and somatomedin C (Thorlacius-Ussing 1990) have been observed in selenium-treated animals, although somatomedin C was not a sensitive end point in humans from a high selenium area of South Dakota (Salbe et al. 1993).

Other Systemic Effects. Urinary excretion of selenium was about twice as great in children with a high incidence of dental caries than in children with a low incidence of caries (Hadjimarkos 1969b). Possible confounding factors (e.g., fluoride status and socioeconomic status) were not considered, however. In Yang et al. (1989a), the incidence of mottled teeth in the medium- and high-selenium groups was increased, but the effect was attributed to interactions between selenium and fluoride.

3.2.2.3. Immunological and Lymphoreticular Effects

No studies were located regarding adverse immunologic or lymphoreticular effects in humans after oral exposure to selenium or selenium compounds. Immune system effects were evaluated in a 120-day double blind study of healthy men who ingested a controlled diet of foods naturally low or high in selenium (Hawkes et al. 2001). Eleven subjects were fed 0.0006 mg selenium/kg/day in the diet for 21 days (baseline period), followed by 0.0002 mg/kg/day (6 subjects) or 0.004 mg/kg/day (5 subjects) for the following 99 days. The results show that the high-selenium diet was not immunotoxic and had some mild and transient immune-enhancing properties. There is an indication that selenium supplementation increased the secondary immune response to diphtheria vaccine when rechallenged at the end of the study. The mean within-subject ratio of diphtheria antibody titers 14 days after reinoculation (day 116) to titers 14 days after the initial challenge at baseline (day 19) was significantly greater in the high-selenium group than in the low-selenium group (2.7±1.8-fold vs. 0.9±0.6-fold, p=0.03). Lymphocyte counts were significantly increased in the high-selenium group on day 45, but not at the end of the study, and there were no clear effects of selenium on numbers of activated or cytotoxic T-cells. The proliferative response of peripheral lymphocytes to stimulation with pokeweed mitogen (a B-cell mitogen) was significantly higher in the high-selenium group than in the low-selenium group on days 45 and 72, although not at the end of the study. There was no selenium-induced lymphocyte proliferation in response to T-cell mitogens (phytohemagglutinin or concanavalin A), or changes in lymphocyte phenotypes, serum immunoglobulins (IgA, IgG, IgM), complement fractions, natural-killer cell activity, delayed-type hypersensitivity skin responses to seven recall antigens (tuberculin purified-protein derivative, mumps, tetanus toxoid, candida, trichophyton, streptokinase strepase, and coccidioidin), or antibody responses to diptheria-tetanus and influenza vaccines. This study was designed as a nutritional study and not as a toxicological study, as the selenium intake levels were well below the tolerable upper limit level (~5.7 µg Se/kg/day) recommended by the Food and Nutrition Board (NAS 2000).

Other human studies also indicate that selenium contributes to enhancing immune function (Baum et al. 1997; Kiremidjian-Schumacher et al. 1994; Peretz et al. 1991). Lymphocyte response was enhanced by dietary selenium, as measured by the T-lymphocyte proliferative response to pokeweed mitogen in elderly people taking a selenium-enriched yeast supplement (0.0014 mg/kg/day for 6 months) (Peretz et al. 1991). This finding is similar to results of the Hawkes et al. (2001) study summarized above, although it was noted that the elderly as a group generally tend to have both lower blood selenium concentrations and lower lymphocyte proliferation than the general population. Dietary supplementation with approximately 0.001 mg selenium/kg/day (as sodium selenate) for 8 weeks caused increased proliferation of active T cells in a group of 11 volunteer subjects (Kiremidjian-Schumacher et al. 1994). The lymphocytes in the exposed subjects had an increased response to stimulation with alloantigen and developed into cytotoxic lymphocytes capable of destroying tumor cells. There was a 118% increase in cytotoxic lymphocyte-mediated tumor cytotoxicity, as well as an 82.3% increase in natural killer cell activity, compared to baseline values. The selenium supplementation regimen used in this study did not cause significant increases in selenium levels in the plasma or red blood cells.

Immune function was evaluated in 40 volunteers from a Finnish population with low blood selenium concentrations that were supplemented with selenium or placebo for 11 weeks (Arvilommi et al. 1983). At the end of the supplementation period, plasma selenium levels were 74 µg/L in the placebo group and 169 µg/L in the supplemental group. Intracellular killing of Staphylococcus aureus by granulocytes was slightly lower in the placebo group than in the selenium group (77.2% compared to 85.2%, p>0.05). No significant changes were observed in phagocytosis, chemotactic factor generation, antibody or leukocyte migration inhibitory factor production by lymphocytes, or proliferative responses to the T-cell mitogens phytohemagglutinin or concanavalin A.

There is evidence that selenium has a role in protecting patients with HIV virus. Immune parameters and nutrients known to affect immune function were evaluated at 6-month intervals in 125 HIV-1-seropositive drug-using men and women (Baum et al. 1997). When all factors that could affect survival were considered jointly, only reduced number of CD4 helper T cells over time and selenium deficiency were significantly associated with mortality. Low plasma selenium (>85 µg/L) represented a significantly greater risk factor for mortality than low helper T cell counts, and conferred a more significant risk than any other nutrient studied, indicating that selenium-deficient HIV patients were more likely to die from HIV infection than those with adequate levels of selenium.

Studies of mice, rats, and cattle suggest that exposure to high doses of sodium selenite, but not selenomethionine, may reduce immunological responses (Johnson et al. 2000; Koller et al. 1986; Raisbeck et al. 1998; Yaeger et al. 1998). BALB/c mice (five males/group) were exposed to drinking water containing 0, 1, 3, and 9 ppm selenium as sodium selenite (0.024, 0.17, 0.38, and 0.82 mg selenium/kg/day) or seleno-L-methionine (0.024, 0.17, 0.47, and 1.36 mg selenium/kg/day) for 14 days (Johnson et al. 2000). The mice exposed to sodium selenite showed significant decreases in the relative spleen weight at 9 ppm and the relative thymus weight at 3 and 9 ppm. The number of splenocytes in the spleens of the 9 ppm group was reduced by 62%. Single-cell splenocyte cultures were made from the spleens of treated animals and used to determine the effects of selenium treatment on mitogen-induced lymphocyte blastogenesis and cytokine production. Cultured splenic lymphocytes from mice exposed to 9 ppm selenium as sodium selenite showed a significant (260%) increase in the basal rate of proliferation and a nonsignificant increase in mitogen-induced proliferation. Exposure to 9 ppm selenium as sodium selenite also produced a significant increase in the amount of tumor necrosis factor α (TNFα) and interleukin-1β (IL-1β) produced by lipopolysaccharide (LPS)-stimulated splenic macrophages. However, the results of this experiment must be interpreted with caution as treatment with 9 ppm selenium as sodium selenite also produced a large and statistically significant decrease in food (21%) and water (43%) consumption, so that some of the effects observed (e.g., changes in organ weights) may reflect effects of dehydration rather than selenium toxicity. In contrast, the similar groups of mice treated with up to 9 ppm selenium as seleno-L-methionine (up to 1.36 mg selenium/kg/day)showed no significant changes in body weight gain, organ weights, water consumption, or food consumption compared with controls. There were no changes in the basal or mitogen-stimulated lymphocyte proliferation following treatment with seleno-L-methionine, and no alteration in the production of TNFα or IL-1β from splenic macrophages was observed.

Another study in BALB/c mice examined the effects of consumption for 47 days of drinking water containing 7 ppm selenium as selenocystine, selenomethionine, or sodium selenite on immune function (Raisbeck et al. 1998). On the 14^th day of the experiment, the mice received a subcutaneous injection of ovalbumin (OVA). Examination of mitogen-stimulated blastogenesis, B-cell function, and IgG concentrations at the end of the experimental period showed a significant decrease in B-cell function for mice treated with the two organic forms of selenium and a significant reduction in the concentration of OVA-specific antibodies for animals treated with any of the three forms of selenium. Total IgG concentration and OVA-stimulated blastogenesis did not vary between groups.

Rats given sodium selenite in drinking water at 0.7 mg selenium/kg/day for 10 weeks exhibited reduced humoral antibody (IgG) production in response to an administered antigen, and reduced prostaglandin synthesis, but there was no effect on natural killer cell (NKC) cytotoxicity (Koller et al. 1986). At lower doses (0.07 or 0.28 mg selenium/kg/day), NKC cytotoxicity was significantly increased, enhancing the immune response to antigenic stimulation, although the delayed-type hypersensitivity (DTH) and prostaglandin E₂ synthesis were significantly reduced. Selenium administration did not affect the ability of resident peritoneal cells to produce interleukin IL-1. Given the enhanced NKC activity at 0.07 and 0.28 mg selenium/kg/day, but not at 0.7 mg selenium/kg/day, and given the reduced antibody and prostaglandin synthesis at 0.7 mg selenium/kg/day, the dose of 0.7 mg selenium/kg/day is identified as the lowest LOAEL. A NOAEL cannot be identified because of the conflict between enhanced NKC activity and reduced DTH and prostaglandin E₂ synthesis occurring at the same dose levels in this study. Antibody responses to ovalbumin were significantly lower in five male antelope (Antilocapra americana) fed a diet containing 15 ppm selenium (a mixture of alfalfa and hay naturally high in selenium) for 164 days than in controls fed a similar diet containing only 0.3 ppm selenium, but there was no difference in total globulin concentration between groups (Raisbeck et al. 1996). No clinical signs of selenosis or treatment-associated lesions were observed in these animals.

Treatment of steers with selenomethionine or selenite in food at doses up to 0.808 mg selenium/kg/day for 120 days produced symptoms of selenosis (hoof lesions), but did not produce any changes in the weight or histology of the spleen or thymus or in the histology of the lymph nodes (O’Toole and Raisbeck 1995).

Leukocyte function was significantly reduced in pregnant cows supplemented with 0.135 mg/kg/day selenium for 3 months from diets that contained 0.25 (control), 6, or 12 ppm selenium as sodium selenite. Treated animals showed a significant decrease in forced antibody production and a depression in mitogenic response compared with controls (0.005 mg selenium/kg/day) (Yaeger et al. 1998). No clinical signs of selenium toxicosis were observed in any of the cows during the experiment.

As selenium can enhance some immune system functions, selenium may have a normal physiological function in the immune system. This is supported by an 8-week study in which treatment of mice with selenium as sodium selenite (0.33 mg selenium/kg/day, dietary) resulted in enhanced ability of cytotoxic T-lymphocytes to destroy tumor cells (Kiremidjian-Schumacher et al. 1992).

Selenium appeared to play a protective role against viral infection in rats (Beck et al. 1995). When selenium-deficient rats were inoculated with a benign strain of Coxsackie’s virus (CVB3/0), six separate point mutations were identified with the progression of virulence, causing myocarditis. Cocksackie’s virus appears to act as a cofactor in the development of the myocardits; this was shown when the virus was isolated from blood and tissue of people with Keshan disease (a cardiomyopathy particularly prevalent in selenium-deficient growing children and women of child-bearing age).

No studies were located concerning immunological or lymphoreticular effects in humans or experimental animals following oral exposure to selenium sulfide or selenium disulfide.

The highest NOAEL values and all LOAEL values from each reliable study for immunological effects following oral exposure to selenium or selenium compounds for each species and duration category are recorded in Table 3-2 and plotted in Figure 3-2.

3.2.2.4. Neurological Effects

Following acute oral exposure to selenium compounds in humans, aches and pains and irritability (Civil and McDonald 1978), as well as chills and tremors (Sioris et al. 1980) have been reported. The dizziness associated with selenium inhalation exposure has not been documented after selenium ingestion.

In a 1964 study, Rosenfeld and Beath reported listlessness, a general lack of mental alertness, and other symptoms of selenosis in a family exposed for approximately 3 months to well water containing 9 mg selenium/L (0.26 mg selenium/kg/day from drinking water). All of the symptoms resolved after use of the seleniferous water was discontinued. Because Rosenfeld and Beath (1964) did not estimate the family’s exposure to dietary selenium, it is not possible to identify the total daily selenium dose associated with the symptoms of selenosis in this family.

In a dietary study of 11 men in a metabolic unit, selenium intake (80 µg/day for the first 21 days, then either 13 [n=6] or 356 [n=5] µg/day for 14 weeks) was reported to have no significant effect on mood, as measured using the Bi-Polar form of the profile of mood states (POMS) (Hawkes and Hornbostel 1996). However, subjects with initially low selenium levels did show significantly greater decreases in mood scores during selenium depletion.

In areas of the People’s Republic of China where populations suffer from chronic selenosis, peripheral anesthesia and pain in the limbs were reported (Yang et al. 1983). In extreme cases, exaggerated tendon reflexes, convulsions, and some paralysis and hemiplegia occurred (Yang et al. 1983). These latter cases were associated with an estimated daily dietary intake of selenium of at least 3.22 mg selenium/person/day, averaging 4.99 mg selenium/person/day (Yang et al. 1983). Assuming a weight of 55 kg for Chinese men (Yang 1989b), these dietary levels represent 0.027 mg selenium/kg/day and 0.09 mg selenium/kg/day, respectively. In another high selenium area, no neurological effects were observed in individuals who consumed up to 1.51 mg selenium/day (0.027 mg/kg/day) (Yang et al. 1983). Danish geriatric patients with a mean age of 75.3 years were given daily either a placebo or an antioxidant cocktail containing 0.004 mg/kg/day of selenium as L-selenomethionine along with zinc, vitamins C, A, B6, and E, and gamma-linolenic acid. After 1 year, whole blood selenium concentrations increased in the treated group, and slight but significant improvements in psychological scores were observed (Clausen et al. 1989). Because a mixture of nutrients was administered, the improvement in the patients cannot be attributed to selenium. People living on ranches with high selenium soils where selenium toxicity in livestock had historically been observed were compared to randomly selected residents in Wyoming and South Dakota. Daily selenium intake was measured by analysis of duplicate food portions. Subjects received a complete physical exam with a symptom questionnaire and laboratory tests. There were no biologically significant changes in clinical signs or blood chemistry. Calculated doses ranged from 0.001 to 0.01 mg selenium/kg/day in the diet (Longnecker et al. 1991).

An increased incidence (4 observed cases, 0.97 expected, standardized incidence ratio=4.14, 95% confidence interval [CI]=1.13–10.60) of amyotrophic lateral sclerosis, a human motor neuron disease of unknown origin, was reported for a cohort of 5,182 residents of Reggio Emilia, Italy who had been exposed to drinking water containing increased selenium (7–9 µg/L)from 1972 to 1988, compared with the incidence among residents of the surrounding area who had received municipal water containing >1 µg/L selenium (Vinceti et al. 1996). A subcohort of 2,065 of these individuals who had been exposed from 1974 (the earliest date for which a chemical analysis of the municipal tap water was available) was also examined and found to have an increased incidence ratio (4 observed cases, 0.47 expected, standardized incidence ratio=8.59, 95% CI=2.34–21.98). However, the study is limited by a water level of selenium that is not generally considered to be high, a lack of individual measurements of selenium exposure, and insufficient information on confounding variables. The lack of data on selenium status indicates that the study found a correlation but not causation.

In a 30-day study of the administration of L-selenomethionine to long-tailed macaques, severe hypothermia was observed in two of five animals administered 0.12 mg selenium/kg/day, but not in any of the eight animals receiving 0.08 mg selenium/kg/day (Cukierski et al. 1989). However, the increased incidence of hypothermia was not statistically significant. Following 1 week of treatment, all animals administered L-selenomethionine, including the two macaques treated with 0.01 mg selenium/kg/day, exhibited increased drowsiness and lethargy (Cukierski et al. 1989).

Symmetrical focal poliomyelomalacia and other forms of paralysis were seen in swine exposed to 0.58–2.1 mg selenium/kg/day after both acute and intermediate exposures (Baker et al. 1989; Goehring et al. 1984; Harrison et al. 1983; Mihailovic et al. 1992; Panter et al. 1996; Penrith and Robinson 1996; Stowe et al. 1992; Wilson et al. 1983, 1988, 1989). This lesion was noted in animals that showed ataxia, inability to stand, and paralysis of the hind limbs. Additionally, bilateral lesions were noted in the ventral horns of the cervical and lumbar/sacral intumescences of the spinal cord. Necrosis and cavitation were evident in the larger lesions (Harrison et al. 1983). Bilateral lesions were also observed in several nuclei of the brain stem and in the reticular formation (Wilson et al. 1983). Wilson et al. (1983) reproduced the syndrome in growing pigs by feeding them sodium selenite at 50 mg selenium/kg in the diet for 20–40 days. The study authors did not provide sufficient information to calculate doses on a mg selenium/kg body weight basis, but assuming that young swine consume approximately 4% of their body weight each day, this dose was approximately 2.1 mg/kg/day.

In a study of weaned 5-week-old pigs, a dose of 1.3 mg selenium/kg/day given as sodium selenite in capsules killed all eight pigs within 10 days. Histopathological lesions were found in the brain and spinal cord (Wilson et al. 1989). In a study in which pigs were fed 1.25 mg selenium/kg/day in the form of A. bisulcatus, D,L-selenomethionine, or selenate, the selenium in A. bisulcatus was the most potent neurotoxin, resulting in complete paralysis in four of five pigs after 5 days of treatment, and in the last pig after 3 weeks of treatment (Panter et al. 1996). In pigs fed selenate, three of five developed complete paralysis, and one pig developed posterior paralysis after 4–21 days of treatment. Although D,L-selenomethionine resulted in the greatest incidence of selenosis, it was the least potent neurotoxicant, resulting in posterior paralysis in two of five pigs after 9 and 24 days of treatment; the pigs that did not develop paralysis were fed D,L-selenomethionine for approximately 31 days. The form of selenium in A. bisulcatus is unknown, although Panter et al. (1996) indicate that it is nonprotein.

It has long been believed that the blind staggers syndrome in livestock results from consumption of plants high in selenium (100–10,000 mg selenium/kg plant) (Rosenfeld and Beath 1964). These plants, which include A. bisulcatus, are known as selenium-indicator plants. “Blind staggers” is characterized by impaired vision, aimless wandering behavior, reduced consumption of food and water, and finally paralysis and death (Rosenfeld and Beath 1964; Shamberger 1986). Trembling of the skeletal muscles was observed in steers fed sodium selenite mixed in the feed at doses between 0.6 and 1.1 mg selenium/kg/day (Maag et al. 1960). At necropsy, two of six steers exhibited neuronal degeneration of the cerebral and cerebellar cortices. However, more recent studies in which cattle were treated with known amounts of selenium have not replicated these effects, and it is likely that “blind staggers” is not solely the result of selenium toxicity, but may also be attributable to other unidentified causes. For example, treatment of 20 steers with selenomethionine or selenite in food at doses up to 0.808 mg selenium/kg/day for 120 days produced symptoms of selenosis (hoof lesions), but did not produce any neurological signs associated with “blind staggers” or any treatment-related changes in the histology of the central nervous system (O’Toole and Raisbeck 1995).

Neurological effects have also been reported for mice after acute or intermediate exposures to selenium. A single oral dose of selenium dioxide dissolved in water given to mice at 1/10^th the LD₅₀ (1.7 mg/kg) caused moderate reductions in alertness, spontaneous activity, touch response, muscle tone, and respiration. Pentobarbital sleeping time was also significantly increased, and there was moderate hypothermia (Singh and Junnarkar 1991). Brain tissue from male BALB/c mice administered sodium selenite or seleno-L-methionine in drinking water at 0, 1, 3, or 9 ppm selenium (sodium selenite: 0.03, 0.24, 0.58, or 1.34 mg selenium/kg/day; seleno-L-methionine: 0.03, 0.26, 0.63, or 1.96 mg selenium/kg/day) for 14 days was examined for changes in the concentrations of norepinephrine (NE), dopamine (DA), dihydroxyphenylacetic acid (DOPAC), homovanillic acid (HVA), serotonin (5-HT), and 5-hydroxyindolacetic acid (5-HIAA) (Tsunoda et al. 2000). Treatment with seleno-L-methionine produced no significant changes in the concentrations of any of the neurotransmitters or their metabolites. DOPAC, DA, and HVA were increased in the striatum of mice receiving 3 or 9 ppm selenium as selenite. The increase was significant at both concentrations for DOPAC and at 3 ppm (but not 9 ppm) for HVA, but was not significant at either concentration for DA. No changes were observed for levels of NE, 5-HT, or 5-HIAA levels in any brain region of mice treated with sodium selenite.

Exposure of female BALB/c mice to 0.21 mg selenium/kg/day for 6 months from diets containing selenium as sodium selenite resulted in significant changes in behavior during open field testing (Boylan et al. 1990). Open field testing measures the arousal level of small rodents and can differentiate between fear-related behavior and general arousal. Mice receiving excessive selenium had reduced sniffing behavior and exhibited greater activity entering more squares, and more interior squares than mice receiving normal selenium diets. These behaviors are indicative of a general state of arousal rather than fear-motivated activity.

No studies were located concerning neurological effects in humans or experimental animals following oral exposure to selenium sulfide or disulfide.

The highest NOAEL values and all LOAEL values from each reliable study for neurological effects following oral exposure to selenium or selenium compounds for each species and duration category are recorded in Table 3-2 and plotted in Figure 3-2.

3.2.2.5. Reproductive Effects

No studies were located regarding adverse effects on human reproduction following oral exposure to elemental selenium or to selenium compounds. Associations between high seminal plasma selenium and impaired sperm count or motility have been inconsistently observed in humans (Bleau et al. 1984; Hansen and Deguchi 1996; Roy et al. 1990). A 120-day double blind experimental study found no adverse changes in sperm indices or reproductive hormone in men (20–45 years old) who consumed a controlled diet of foods naturally high or low in selenium (Hawkes and Turek 2001). Eleven subjects were fed a diet that provided 0.0006 mg selenium/kg/day for the first 21 days of the study, followed by diets providing 0.0002 mg selenium/kg/day (6 subjects) or 0.004 mg selenium/kg/day (5 subjects) for 99 days. Semen quality (sperm concentration, semen volume, sperm total number, fraction motile sperm, percent progressive sperm, mean forward velocity, and various sperm morphology parameters), reproductive hormone levels (serum testosterone, follicle-stimulating hormone, luteinizing hormone, prolactin, estradiol, and progesterone), and thyroid hormone levels (serum T₃ and TSH) were evaluated during weeks 3 (baseline values), 8, and 17 (ending values).

Selenium levels in blood plasma began to change within 3 days of starting the low- and high-selenium diets and progressively continued throughout the study (Hawkes and Turek 2001). By week 17, mean plasma selenium concentrations had increased by 109% in the high-selenium group and decreased by 38.5% in the low-selenium group. A similar pattern of changes occurred in seminal plasma selenium, although selenium levels in sperm did not change significantly in either group. Mean sperm motility was significantly different in the low-selenium subjects and high-selenium groups at week 13, but not at weeks 8 or 17. The fraction of motile sperm increased an average of 10% in the low-selenium group by week 13, and was essentially the same as the baseline value at week 17. Sperm motility decreased an average of 32% in the high-selenium group at week 13, and ended 17% lower than baseline value at week 17. ANOVA showed a significant main effect of dietary selenium on sperm motility, as well as a significant selenium x time interaction, indicating that the group responses diverged over time. Baseline and ending motile sperm fractions in the high-selenium group were 0.588±0.161 and 0.488±0.193, respectively; ≥50% motility is considered normal (FDA 1993). The decrease in sperm motility in the high-selenium group cannot be clearly attributed to selenium because the effect was not consistent over the duration of exposure, is unlikely to be adverse because it is at the low end of the normal range, and is not accompanied by any significant changes in other indices of sperm movement (progression or forward velocity), or sperm numbers or morphology. Additionally, there were no effects of selenium on serum levels of reproductive hormones, and changes in thyroid hormones, which could affect sperm function, were not outside normal ranges (see Endocrine Effects in Section 3.2.2.2).

A nonsignificant increase in spontaneous abortions (relative risk [RR]=1.73; 95% CI=0.62–4.80) was reported among births in the municipality of Reggio Emilia, Italy, where women had been exposed to drinking water containing 7–9 ug/L levels of selenium (as selenate, reported estimated intake 10–20 ug/day) between 1972 and 1988 (Vinceti et al. 2000a). This study is limited by a level of selenium in water that is not considered high, lack of data on selenium status, and insufficient information on confounding variables. Selenium deficiency has been implicated as a risk factor for recurrent miscarriage in humans (Al-Kunani et al. 2001; Barrington et al. 1996, 1997; Güvenc et al. 2002; Kumar et al. 2002).

Data from animal studies suggest that exposure to excessive selenium has adverse effects on testosterone levels and sperm production and increases the percentage of abnormal sperm (El-Zarkouny et al. 1999; Kaur and Parshad 1994; NTP 1994). A significant reduction (49%) in serum testosterone levels was reported for New Zealand white rabbits gavaged with 0.001 mg selenium/kg/day as sodium selenite once a week for 6 weeks (El-Zarkouny et al. 1999). The percentage of spermatozoa without an acrosome was also increased in treated rabbits compared with controls, but the difference was not significant. Sperm motility, ejaculate volume, sperm concentration, and total sperm output were all reduced by selenium treatment, but statistical analysis of these data was not presented.

Exposure of Wistar rats to 0.234 mg selenium/kg/day as sodium selenite in water produced testicular hypertrophy (Turan et al. 1999a). A dose-related increase in abnormal sperm and a decrease in live sperm were observed in wild-caught rats exposed to selenite in the diet at 0.1 and 0.2 mg selenium/kg/day (Kaur and Parshad 1994). The percentage of abnormal sperm was 3.9% at 0.1 mg/kg/day and 24.6% at 0.2 mg/kg/day. The abnormalities observed were principally in the midpiece region of the sperm, the region that contains a selenoprotein (Sunde 1990). Decreased sperm counts were observed in rats provided with selenate or selenite in drinking water for 13 weeks at a dose of 0.29 mg selenium/kg/day for selenate and a dose of 0.17 mg selenium/kg/day for selenite (NTP 1994). Effects on sperm were not observed in mice treated with selenate or selenite in the drinking water at doses up to 5.45 mg selenium/kg/day for selenate or up to 3.31 mg selenium/kg/day for selenite (NTP 1994). The administration of 1.05 mg selenium/kg/day as potassium selenate to rats in drinking water for 1 year did not affect male fertility (Rosenfeld and Beath 1954), and the administration of 0.57 mg selenium/kg/day as sodium selenate for three generations did not reduce male fertility in mice (Schroeder and Mitchener 1971b). A short-term reproductive study of the effects of sodium selenate in drinking water on rats at doses (0.418 mg selenium/kg/day) that produced signs of systemic toxicity did not cause any increase in sperm abnormalities or lesions of the testis or epididymis (NTP 1996). Selenium administered in the diet or in drinking water over short exposure periods (e.g., 1 month) does not appear to affect the fertility of female animals unless the intake is sufficiently high to cause general toxicity (Nobunaga et al. 1979). Despite a small increase in the number of abnormal length estrous cycles, Nobunaga et al. (1979) found no adverse effect on the fertility of female mice from administration of sodium selenite at doses of 0.34 mg selenium/kg/day in drinking water for 30 days before mating and for 18 days during pregnancy. On the other hand, chronic exposure of mice and rats to otherwise nontoxic doses has been shown to reduce fertility and to reduce markedly the viability of the offspring of pairs that are able to conceive (Schroeder and Mitchener 1971b; Wahlstrom and Olson 1959b).

A study of supplementation of female pigs with 0.1 or 0.3 ppm selenium (doses not available) administered as a selenium-enriched yeast or sodium selenite in the diet, from 60 days before breeding until weaning found no adverse effects on reproductive performance (measured by number of offspring) or growth (Mahan and Kim 1996). In another study of the effect of selenium on fertility in pigs, females fed sodium selenite at 0.4 mg selenium/kg/day from 8 weeks of age exhibited reduced rates of conception and also produced offspring with significantly reduced birth weight and weaning weights in the first and second litters (Wahlstrom and Olson 1959b). An altered menstrual cycle was reported in monkeys administered 0.08 mg selenium/kg/day as L-selenomethionine for 30 days (Cukierski et al. 1989).

Vaginal cytology of female rats provided with drinking water containing selenate or selenite indicated that the rats spent more time in diestrus and less time in proestrus and estrus than the controls (NTP 1994). This effect occurred following treatment with 0.31 mg selenium/kg/day as selenate or 0.86 mg selenium/kg/day as selenite. The animals in these studies were not mated, so it is not known if the effects on the estrous cycle had any effect on fertility. Effects on the estrous cycle were not observed in mice treated with selenate or selenite in the drinking water at doses up to 7.17 mg selenium/kg/day for selenate, or at doses up to 3.83 selenium/kg/day for selenite (NTP 1994).

In a three-generation reproduction study, selenium administered as sodium selenate (0.57 mg selenium/kg/day) in the drinking water of breeding mice produced adverse effects on reproduction (Schroeder and Mitchener 1971b). The most notable observed effects included the failure of about half of the F3 generation pairs to breed successfully. In a two-generation study using rats, selenium administered as potassium selenate had no effect on reproduction at a dose of 0.21 mg selenium/kg/day for 1 year; however, decreased fertility and pup survival were noted at 1.05 mg selenium/kg/day (Rosenfeld and Beath 1954). At 0.35 mg selenium/kg/day for 1 year, the number of young successfully reared by the females was reduced by 50%, and the body weight of the females was approximately 20% less than that of the control females (Rosenfeld and Beath 1954).

A short-term reproductive study of the effects of sodium selenate in drinking water on rats reported some female reproductive toxicity (reduced corpora lutea, reduced implants per litter, shorter estrous cycle), but only at doses (0.418 mg selenium/kg/day) that produced signs of severe maternal toxicity, including a large reduction in water consumption (NTP 1996).

In a review of selenium poisoning in domestic animals, Harr and Muth (1972) noted a decreased conception rate and an increased fetal resorption rate in cattle, sheep, and horses fed diets naturally containing organic selenium compounds at 25–50 mg selenium/kg diet. Assuming that large animals consume an amount of food equal to about 2–3% of their body weight daily, the doses would have been approximately 0.5–1.5 mg selenium/kg/day. These levels of selenium also produced other signs of toxicity, including hair loss, lameness, and degeneration and fibrosis of the heart, liver, and kidneys. In a case control study of 136 Holstein cows from four herds, an association of cystic ovaries with blood selenium concentrations >108 ng/mL was found (Mohammed et al. 1991). The concentration of progesterone in the milk was significantly higher in the controls than in the cows receiving selenium supplementation, but no information on the selenium dose was presented. No change in estrus cycle length, estrus behavior, progesterone, or estrogen profiles or pregnancy rate was observed in a study of the reproductive response of ewes fed alfalfa pellets containing sodium selenate (24 ppm selenium) or A. bisculcatus (29 ppm selenium) as a selenium source for 88 days, from >52 days before pregnancy up to day 28 of gestation (Panter et al. 1995). Doses could not be calculated as food consumption was not listed, and the paper states that the food supply was limited to match that of the group with the lowest intake.

The highest NOAEL value for reproductive effects following intermediate oral exposure to sodium selenite and all reliable LOAEL values for reproductive effects following intermediate or chronic oral exposure to selenium compounds other than selenium sulfide are recorded in Table 3-2 and plotted in Figure 3-2.

3.2.2.6. Developmental Effects

No studies have demonstrated that selenium or its compounds are teratogenic in humans. Robertson (1970) reported on the outcome of pregnancies in a laboratory in which workers handled sodium selenite. Of the five pregnancies, four ended in spontaneous abortion and one resulted in an infant with bilateral clubfoot. The urinary selenium levels in all subjects were similar to those in other individuals living in the same area. The limited number of cases, possible exposure to other toxic agents, and other confounding factors leave the relationship between sodium selenite and developmental effects inconclusive.

No significant increase in spontaneous abortions (RR=1.73; 95% CI=0.62–4.80) was reported among births in the municipality of Reggio Emilia, Italy, where women had been exposed to drinking water containing 7–9 µg selenium/L (as selenate) between 1972 and 1988 (Vinceti et al. 2000a). Body weight and length at birth were similar in infants of exposed and unexposed women, and no significant increase in the prevalence of congenital abnormalities was found for 353 infants of exposed mothers compared with the 14,481 births among unexposed women. This study is limited by a level of selenium in water that is not considered high, lack of data on selenium status, and insufficient information on confounding variables.

Zierler et al. (1988) performed a case control study of 270 children born in Massachusetts with severe congenital heart disease and 665 controls randomly selected from birth certificates. The study compared the selenium concentrations in the public drinking water supply used by the mothers close to the time of conception to the selenium concentrations in the water consumed by the controls. The results indicated that selenium exposure via drinking water was associated with beneficial effects, particularly a reduction in the risk of congenital heart defects (cono-truncal defects, venticular septal defects, coarctation of the aorta, and patent ductus arteriosis), but many variables are unknown, including other possible confounders (no adjustment for age, parity, tobacco, alcohol, drug use, or socioeconomic status), other sources of selenium in the mothers’ diet and environment, the amount of drinking water consumed, and the selenium concentrations in the water during the first trimester.

Excess selenium is a demonstrated teratogen in birds. However, there is no clear evidence linking selenium exposures to teratogenic effects in mammals. Several studies have documented the sensitivity of chick embryos to selenium poisoning. Hatchability of eggs is reduced by dietary levels of organic selenium in grain that are too low to cause toxicity in other farm animals. The eggs are fertile but often produce grossly deformed embryos lacking eyes and beaks and having deformed wings and feet (Franke and Tully 1935; Franke et al. 1936; Gruenwald 1958; Palmer et al. 1973). Deformed embryos have also been produced by injection of aqueous sodium selenite or sodium selenate into the air cell of the normal, fertile eggs of chickens (Franke et al. 1936; Khan and Gilani 1980). The incidence of malformation among coot, duck, stilt, and grebe embryos from eggs of birds ingesting plant and other food from irrigation drainwater ponds in the San Joaquin Valley of California was much higher than expected (10–42%, depending on the species, versus >1% based on data from other areas) (Ohlendorf et al. 1986a, 1988). Selenium concentrations in these ponds were >0.3 mg/L.

The consumption of naturally high seleniferous diets by sheep (Rosenfeld and Beath 1964) and cattle (Dinkel et al. 1963) may interfere with normal fetal development and produce malformations. Malformations were associated with alkali disease and occurred at dietary levels that produced other toxic manifestations, but it is not clear if these reports took into account consumption of other toxic range plants. The specific selenium compound or compounds possibly associated with livestock developmental toxicity have not been identified. No change in the outcome of pregnancy was observed in a study of the reproductive response of ewes fed alfalfa pellets containing sodium selenate (24 ppm selenium) or A. bisculcatus (29 ppm selenium) as a selenium source for 88 days, from >52 days before pregnancy up to day 28 of gestation (Panter et al. 1995). All lambs appeared normal, and there was no significant difference in the number or weight of lambs born to treated and control ewes. Doses could not be calculated, as food consumption was not listed, and the paper states that the food supply was limited to match that of the group with the lowest intake.

In an intermediate-duration study, an increased number of deaths between birth and weaning, reduced birth weight, and reduced body weight at weaning were observed in offspring of pigs fed selenite at 0.4 mg/kg/day for an unstated period of time (Wahlstrom and Olson 1959b). Treatment of 15 pregnant cows with diets containing 0.25 (control), 6, or 12 ppm selenium (0.005, 0.135, or 0.265 mg Se/kg/day) as sodium selenite beginning at 80–110 days gestation and continuing for 3 months resulted in no abnormalities among the offspring apart from one calf in the 12 ppm group that was born weak and subsequently died (Yaeger et al. 1998). This calf had myocardial lesions similar to those described for selenium toxicosis and had markedly elevated hepatic selenium levels, although selenium levels in blood and hair of this calf and its dam were lower than average for the 12 ppm group.

In studies of laboratory mammals, the administration of inorganic selenium compounds at levels that are not maternally toxic has not produced terata (Bergman et al. 1990; Chiachun et al. 1991; Ferm et al. 1990; NTP 1996; Poulsen et al. 1989; Rosenfeld and Beath 1954; Schroeder and Mitchener 1971b; Thorlacius-Ussing 1990). Ferm et al. (1990) administered a single dose of sodium selenate, sodium selenite, or L-selenomethionine (0, 1.8, 2.2, 2.7, 4.0, 5.0, or 5.9 mg selenium/kg/day) to pregnant Syrian hamsters on gestation day 8. Pathological examination of the fetuses on day 13 showed that the percentage of abnormal litters was significantly increased at doses of ≥2.7 mg/kg. Encephalocele was the major malformation noted, and incidences were as follows: 0/71 controls; 4/55 (7.3%) at 1.8 mg/kg; 1/49 (2%) at 2.2 mg/kg; 7/66 (10.6%) at 2.7 mg/kg; 15/70 (21.4%) at 4 mg/kg; 9/38 (23.7%) at 5 mg/kg; and 6/16 (37.5%) at 5.9 mg/kg. Nobunaga et al. (1979) found that administration of sodium selenite in drinking water at 0.34 mg selenium/kg/day for 30 days before mating and for 18 days during pregnancy slightly, but significantly, reduced fetal growth in mice. However, there was no effect on fetal growth in the same study at a dose of 0.17 mg selenium/kg/day. A short-term developmental study (from gestation day 6 until birth) of the effects of sodium selenate in drinking water on rats produced some developmental toxicity (decreased number of live births, reduced pup weight, increased gestation period), but only at doses (0.418 mg selenium/kg/day) that produced signs of severe maternal toxicity including a large reduction in water consumption (NTP 1996). Selenium administered as potassium selenate in drinking water to male and female rats at a dose of 1.05 mg selenium/kg/day for 1–8 months for two successive generations did not cause congenital malformations (Rosenfeld and Beath 1954). Similarly, administration of 0.57 mg selenium/kg/day as sodium selenate in the drinking water of breeding mice for three generations did not have teratogenic effects, although there was an increased incidence in fetal deaths, and a high proportion of the surviving offspring were runts (Schroeder and Mitchener 1971b).

Poulsen et al. (1989) demonstrated that pigs exposed to 42.4 mg/day of selenium as sodium selenite in feed throughout pregnancy produced normal litters, with no adverse effect on piglet survival, litter size, or body weight at birth. Body weights of the pigs during pregnancy were not provided, and therefore, mg/kg/day doses could not be calculated. Body weight gains of pigs fed selenium as selenite at a dose of 0.4 mg selenium/kg/day after weaning (duration not specified) were reduced (Wahlstrom and Olson 1959a). The reduction in body weight gain was greater among pigs from dams not fed selenium during gestation and lactation compared to pigs fed selenium (0.4 mg/kg/day) during gestation and lactation. Without providing data, the study authors indicated that there was a greater loss of pigs at birth and during lactation from sows fed selenium, which may have eliminated susceptible pigs.

In a teratology study of long-tailed macaques, no gross abnormalities or growth retardations were observed in fetuses from mothers administered L-selenomethionine at levels of 0.003, 0.025, 0.15, or 0.30 mg selenium/kg/day on gestational days 20–50 (10 animals per group); the mid and high doses were maternally toxic (Tarantal et al. 1991).

The highest NOAEL value and all reliable LOAEL values for developmental effects following intermediate or chronic oral exposure to selenium compounds are recorded in Table 3-2 and plotted in Figure 3-2.

3.2.2.7. Cancer

Early studies reporting that selenium was carcinogenic in mammals after being provided as seleniferous corn or wheat in the diet (Nelson et al. 1943), as sodium selenite or sodium selenate in drinking water (Schroeder and Mitchener 1971a), or as sodium selenate in the diet (Volgarev and Tscherkes 1967) were flawed. The majority of subsequent studies of humans and animals have revealed no association between selenium intake and the incidence of cancer (Azin et al. 1998; Beems 1986; Coates et al. 1988; Harr et al. 1967; Ma et al. 1995; Menkes et al. 1986; Ratnasinghe et al. 2000; Thompson and Becci 1979; Vinceti et al. 1995; Virtamo et al. 1987) or a clear chemopreventive association (Birt et al. 1982; Clark et al. 1996a, 1999; Finley et al. 2000; Ip 1981, 1983; Ganther and Lawrence 1997; Ip and Lisk 1995; Ip et al. 1996, 1997, 1998, 2000a, 2000b; Jiang et al. 1999; Ma et al. 1995; Medina and Shepherd 1981; Moyad 2002; Overvad et al. 1985; Schrauzer et al. 1976, 1977; Shamberger et al. 1976; Soullier et al. 1981; Thompson and Becci 1980; Woutersen et al. 1999; Yoshizawa et al. 1998). Some epidemiological and experimental evidence suggests that selenium exposure, under certain conditions, may contribute to a reduction in cancer risk (Clark et al. 1996a, 1999; El-Bayoumy 2001; Ganther 1999; Moyad 2002; Spallholz 2001; Yoshizawa et al. 1998), and the chemopreventive potential of supplemental selenium is currently under research (Clark et al. 1999; Duffield-Lillico et al. 2002; Reid et al. 2002).

The only selenium compound that has been shown to be carcinogenic in animals is selenium sulfide (NTP 1980c), although there is also some evidence for carcinogenicity due to ethyl selenac (selenium diethyldithiocarbamate) (Innes et al. 1969; NCI 1968). These compounds are very different chemically from the organic and inorganic forms found in foods and the environment. Human dietary studies generally do not identify the selenium form specifically; both organic (from grains and other plant and animal products) and inorganic (from drinking water) forms are ingested. Animal bioassays in which selenium was administered as sodium selenate, sodium selenite, or organic forms of selenium have all shown similar negative results.

Excess incidence of melanoma was reported for a cohort of 2,065 individuals that were exposed to 7–9 µg/L levels of selenium as selenate in the municipal water supply in Reggio Emilia, Italy from 1972 until 1988 (Vinceti et al. 1998). Eight individuals among the exposed cohort developed melanoma compared with 128 in the remainder of the municipal population (total number of individuals not given). The standardized mortality ratios (SMRs) were 5.0 (95% CI=1.6–12.0) for males and 3.2 (95% CI=1.0–7.7) for females. The authors estimate the general dietary intake of selenium in the area to be 45–50 µg/day and the excess selenium supplied in the contaminated tap water to be 10–20 µg/day. However, the study is limited by the fact that no individual measurements of selenium exposure were made, and individuals were classed as exposed or unexposed depending on their place of residence. The lack of data on selenium status indicates that the study found a correlation but not causation. Other limitations include a water level of selenium that is not generally considered to be high and insufficient information on confounding variables.

A study of the effects of nutritional supplementation with selenium found a significant reduction in overall cancer mortality and in the incidence of lung (RR=0.54, 95% CI=0.3–0.98, p=0.04), colorectal (RR=0.42, 95% CI=0.18–0.95, p=0.03), and prostate (RR=0.37, 95% CI=0.18–0.71, p=0.003) cancer (Clark et al. 1996a, 1999). The original intent of the study was to assess the effects of selenium supplementation on nonmelanoma skin cancer. Patients with a history of skin carcinoma (1,312 individuals) were randomized into two groups; one group received a selenium supplement of 200 µg/day, and the other received a placebo. Groups were treated for an average of 4.5 years and followed for an average of 6.5 years. Supplementation produced no difference in skin cancer incidence; however, secondary end point analyses of the data found a protective effect for selenium for the cancers described above. A reanalysis of the lung cancer data added eight cases to the selenium-treated group, four cases to the placebo group, and increased follow-up to 7 years (Duffield-Lillico et al. 2002; Reid et al. 2002). Selenium supplementation did not reduce lung cancer incidence in the full population (RR=0.70, 95% CI=0.40–1.21, p=0.18; hazard ratio [HR]=0.74, 95% CI=0.44–1.24, p=0.26), although a nominally significant decrease was observed among subjects with baseline plasma selenium concentrations in the lowest tertile (HR=0.42, 95% CI=0.18–0.96, p=0.04). The analysis for the middle and highest tertiles of baseline selenium level showed HRs of 0.91 and 1.25, suggesting that there was a trend toward a reduction in risk of lung cancer with selenium supplementation.

Supporting evidence for an antiprostate cancer effect of selenium was obtained for a nested case-control design within the Health Professional Follow-up study (Yoshizawa et al. 1998), which found that higher prediagnostic selenium levels were associated with reduced prostate cancer incidence. This study included 33,737 male health professionals aged 40–75 years who provided toenail clippings in 1987. The cohort was assessed by questionnaire for incidence of new cases of prostate cancer from 1989 to 1994. Higher levels of selenium in toenail clippings were significantly associated with a reduced risk of prostate cancer. After controlling for factors such as a family history of prostate cancer, body mass index, calcium intake, lycopene intake, saturated fat intake, vasectomy, and geographical region, the odds ratio (OR) was 0.35 (95% CI=0.16–0.78, p for trend=0.03). Studies of selenium supplementation have generally shown a reduction in prostate cancer risk only in individuals who had lower levels of baseline plasma selenium, whereas subjects with normal or higher levels did not benefit and may have an increased risk for prostate cancer (Moyad 2002).

Epidemiological studies that focused on the selenium concentration of forage crops as an indicator of available dietary selenium indicated an inverse association between selenium levels and cancer occurrence, with few exceptions. In the United States, male mortality due to cancer of the tongue, esophagus, stomach, intestine, rectum, liver, pancreas, larynx, lungs, kidneys, and bladder was significantly lower in states with high selenium levels in forage crops (concentrations in excess of 0.10 mg selenium/kg) (Shamberger et al. 1976). For females in states with high selenium levels, significantly lower cancer death rates were found for cancer of the esophagus, stomach, intestine, rectum, liver, pancreas, lungs, bladder, thyroid, breast, and uterus (Shamberger et al. 1976). Only male and female mortality due to cancer of the skin and eye, male mortality due to cancer of the lip and aleukemic leukemia (a deficiency or absence of leukocytes in the blood due to leukemia), and female mortality due to dermal melanoma were associated with high selenium levels in forage crops. Many of the high selenium areas are in the southwestern portion of the United States, and therefore, exposures to ultraviolet light may have contributed to the skin cancers observed in these areas (Shamberger et al. 1976). In a comparison of selenium intake and cancer mortality rates in different countries, Schrauzer et al. (1977) detected a cancer chemopreventive association between the selenium content of the diet and age-corrected cancer mortality from leukemia and cancers of the intestine, rectum, breast, ovary, prostate, lung, pancreas, skin, and bladder.

In a case control study of lung cancer patients, Menkes et al. (1986) found that the risk of lung cancer was not associated with serum selenium levels (0.113 and 0.110 mg selenium/L in cases and controls, respectively), but was significantly increased with decreasing serum levels of vitamins A and E. The study authors suggested that high serum selenium levels were significantly associated with an increased incidence of squamous cell carcinoma as compared to other cellular tumor types, but the statistical analysis used was flawed. In a region of China with high rates of stomach cancer and low intake of several micronutrients (selenium not specifically stated), an intervention trial in 29,584 adults for 5.25 years demonstrated a 21% decrease in stomach cancer in the randomly selected group receiving a nutritional supplement of beta carotene (15 mg/day), vitamin E (30 mg/day), and selenium (50 µg selenium/day as selenium yeast) (Blot et al. 1993). However, because the three nutrients were given in combination to a nutritionally deficient population, it is not possible to determine what part of this effect (if any) was due to selenium.

Low serum selenium levels have been associated with an increased incidence of cancer in some prospective epidemiological studies (Salonen et al. 1984, 1985; Willet et al. 1983). In the United States, Willet et al. (1983) found that the risk of cancer for subjects in the lowest quintile (fifth) of serum selenium concentrations (>0.115 mg selenium/L) was twice that of subjects in the highest quintile (>0.154 mg selenium/L). In Finland, Salonen et al. (1985) found the risk of fatal cancer for subjects in the lowest tertile (third) of serum selenium concentrations (>0.047 mg selenium/L) was 5.8 times higher than that of the remaining subjects. Mean serum selenium levels in Americans (0.129 mg selenium/L cases; 0.136 mg selenium/L controls) (Willett et al. 1983) are more than twice the mean serum selenium levels in the Finns (0.0505 mg selenium/L for cases; 0.0543 mg selenium/L for controls) (Salonen et al. 1984). Although the age-specific risk of fatal cancers in the two populations cannot be calculated from the data reported, the overall incidence of cancer in the 4-year Finnish study was less than half that in the 5-year U.S. study. In addition, other prospective studies of Americans have found no correlation between fatal cancer and blood selenium concentrations (Coates et al. 1988). Thus, one may not be able to predict relative cancer risks with serum selenium levels in one population based on data from another population.

A 9-year prospective follow-up study was conducted by Virtamo et al. (1987) on a group of men in Finland. At the beginning of the study, blood samples were taken as part of a study of coronary heart disease and other atherosclerotic diseases. Cancer data were collected from central registries for the years 1976 through 1983. The results indicated no association between low serum selenium levels (>0.045 mg selenium/L) and an increased risk of cancer. Evidence suggests that combined dietary deficiencies of vitamin E and selenium may be associated with increased cancer risk (Salonen et al. 1985).

Epidemiological studies of breast cancer have found inverse correlations, positive correlations, and no correlations between tissue selenium concentrations and cancer incidence (recently reviewed by Garland et al. 1993). In a case control study of plasma selenium and breast cancer in which the controls had benign breast disease, a preventive effect of selenium was found only among individuals who had higher plasma selenium and were not taking selenium supplementation (Hardell et al. 1993). This effect was significant (odds ratio 0.38) at a serum selenium concentration range of 0.08–0.09 mg/L in women 50 years old or more. GPX activity in erythrocytes was not found to be a marker for the risk of breast cancer. A case control study of 162 cases of breast cancer in Dutch women did not find a significant difference in dietary, plasma, erythrocyte, or toenail selenium between cases and 529 controls when multivariate-adjusted odds ratios were calculated. Dutch women have lower selenium intake than women in the United States and one of the highest incidences of breast cancer in Western Europe. The authors of this study surmised that other studies reporting an inverse relationship between selenium levels and breast cancer may be seeing an effect of the cancer (e.g., decreased uptake of selenium or anorexia), rather than lower selenium level contributing to the development of cancer (van’t Veer et al. 1990). Similarly, a large prospective study of 434 cases in the United States found no correlation between selenium content in nails, established as a long-term marker of selenium (Hunter et al. 1990a), and breast cancer (Hunter et al. 1990b). It is interesting to note that a more recent investigation of the same cancer cases found an inverse correlation between vitamin A (retinoids) in the diet and breast cancer (Hunter et al. 1993). Retinoids are believed to have chemopreventive activity (Clausen et al. 1989; Hunter et al. 1993). Although the data as a whole for breast cancer and tissue selenium concentrations do not support a clear chemopreventive effect for selenium, it is possible that very high selenium concentrations or very low selenium concentrations outside the ranges observed in these studies could play a role in human cancer risk (Garland et al. 1993).

There were several inadequacies in the early studies that reported carcinogenic effects in animals following oral administration of selenium-containing compounds. Nelson et al. (1943) (also reported as Fitzhugh et al. 1944) administered naturally seleniferous corn or wheat diets containing 5, 7, or 10 mg selenium/kg diet (0.25, 0.35, or 0.50 mg selenium/kg/day) to female rats for 2 years. Selenium administration produced high mortality (69%) in all treatment groups by the end of the first 12 months, and the first tumors appeared after 18 months of treatment. Tumors developed only in animals with cirrhotic livers, and the tumors were reported to be nonmalignant. The possible contribution of overt hepatotoxicity to the development of liver tumors is not known. The incidences of tumors in the surviving animals in the three dose groups were 6/25, 3/21, and 2/7, respectively. The investigators had difficulty discerning malignant from nonmalignant tumors, and most animals had died of cirrhosis of the liver before the appearance of liver tumors. These difficulties cast doubt on the conclusion of the investigators that selenium induced tumor formation in these rats.

A statistically significant increase was reported in the incidence of all tumors and malignant tumors in rats administered 0.28–0.42 mg selenium/kg/day as sodium selenite or sodium selenate in drinking water for a lifetime (Schroeder and Mitchener 1971a). Not all autopsied animals were examined histologically, however, and high mortality in all groups occurred as a result of a virulent pneumonia epidemic that occurred during the study. In addition, the statistical analysis failed to account for the fact that the selenium-treated rats lived longer than did the control rats. Analysis of the incidence of tumors among animals with equal longevities indicates that the incidence of tumors in the selenate-treated rats was not significantly different from that in the controls.

A series of dietary studies assessed the effects of various dietary supplements on selenium tumor induction in male rats (Volgarev and Tscherkes 1967), but the conclusions that can be drawn from these experiments are limited since they did not include controls. Tumors (primarily liver) were found in 10/23 male rats administered sodium selenate in the diet at a dose of 0.34 mg selenium/kg/day for more than 18 months (Volgarev and Tscherkes 1967). The first tumors appeared after 18 months of selenium administration, by which time, 43% of the animals had already died (group started with 40 animals). Tumors were also found in 3/16 male rats administered sodium selenate in the diet at an initial dose of 0.34 mg selenium/kg/day for 6 months, followed by 0.68 mg/kg/day until the animals’ death (Volgarev and Tscherkes 1967). In a third group of experiments, no tumors were found in 200 male rats administered sodium selenate in the diet (0.34 mg selenium/kg/day) for 26 months. However, there was very high mortality among these rats, and survival time was 10 months shorter than among the similarly fed animals in the first experiment. The authors noted that an additional 200 male rats were maintained in their laboratory during these experiments and fed stock rations. The life spans of these animals exceeded those used in the experiments and no tumors were found at autopsy.

More recent animal bioassays have failed to demonstrate any association between excessive selenium exposure and carcinogenesis. Chen et al. (2000) reported a significant increase in rat esophageal adenocarcinogenesis in response to supplementation with 0.06 mg selenium/kg/day as sodium selenite for 40 weeks. However, selenium supplementation has generally been shown to significantly inhibit tumors induced by chemicals, viruses, or ultraviolet light (Birt et al. 1982; Finley et al. 2000; Ip 1981, 1983; Ip and Lisk 1995; Ip et al. 1996, 1997, 1998, 2000a, 2000b; Jabobs 1983; Jacobs et al. 1977a, 1977b, 1979, 1981; Jiang et al. 1999; Medina and Shepherd 1981; Overvad et al. 1985; Schrauzer et al. 1976; Soullier et al. 1981; Thompson and Becci 1980; Woutersen et al. 1999). Results following administration of selenium as sodium selenate, sodium selenite, and organic forms of selenium are similar. Additional research reviewed in El-Bayoumy (1991, 1995, 1997) indicates that synthetic organoselenium compounds may be more potent cancer preventive agents than selenate, selenite, or the selenoamino acids.

Two sources reported the results of a study of rats administered sodium selenate or sodium selenite in the diet for a lifetime (Harr et al. 1967; Tinsley et al. 1967). A vehicle control and two positive control groups (administered a known hepatocarcinogen, N-2-fluorenyl-acetamide [FAA]) were included. Mortality was high in the highest dose group (0.8 mg selenium/kg/day), and therefore, selenium administration was discontinued. Longevity was reduced in animals fed 0.4 mg selenium/kg/day, but not in the animals administered lower doses. Of the original 1,437 experimental animals, 1,126 were necropsied. Half of the 88 FAA-fed rats developed neoplasms, half of which were hepatic carcinomas, indicating that the strain of rat and dietary conditions were compatible with the development of hepatic carcinogenesis. The incidence of cancer of all types in the necropsied control rats (11 out of 482, or 2.3%) was somewhat higher than the incidence of cancer in the selenium-treated animals that were necropsied (9 out of 553, or 1.6%). A statistical analysis of the data from this study was not reported. Although the reduced longevity of animals administered 0.4 mg selenium/kg/day might have prevented the observation of some late-developing cancers, the large number of rats necropsied, the end points examined, and the doses administered provide credible evidence of the lack of carcinogenic potential of sodium selenate or selenite.

Mice were fed tortula yeast diets containing up to 1.0 mg selenium/kg diet (equivalent to 0.13 mg selenium/kg body weight/day) as sodium selenite for 2 weeks prior to a single application of 0.125 mg 7,12-dimethylbenz[a]anthracene (DMBA) to the skin or repeated daily applications of 0.25 mL of a 0.03% solution of benzo[a]pyrene in acetone for 27 weeks (Shamberger 1970). The highest dose of selenite used, 0.13 mg selenium/kg/day, significantly decreased the number of tumors induced by both aromatic compounds. No significant increase was found in the incidence of spontaneous tumors in mice following administration of 3 mg selenium/L in drinking water as either sodium selenite or sodium selenate for a lifetime (Schroeder and Mitchener 1972). This level corresponds to doses of 0.31–0.34 mg selenium/kg/day for the males and 0.42 mg selenium/kg/day for the females. The single dose administered, however, might not have been the maximal dose that could be tolerated. There were 7% more malignant tumors in the selenium-treated animals (13 out of 88 sectioned, or 15%) than in the controls (10 out of 119 sectioned, or 8%), but the difference was not statistically significant. The forms of selenium administered did not influence the incidence of tumors. In this study, only 88 out of 211 selenium-treated animals and 109 out of 209 control animals were examined histologically.

The only selenium compound that has been shown to be carcinogenic in animals is selenium sulfide (NTP 1980c), although there is some inconclusive evidence that ethyl selenac may also be carcinogenic (Innes et al. 1969; NCI 1968). A statistically significant increase in hepatomas (0/16 controls; 12/16 treated) was observed in male mice of one strain (C57BL/6 x C3H/Anf)F₁) receiving 2 mg selenium/kg as ethyl selenac, but not in male or female mice of another strain (C57BL/6 x AKR)F₁) receiving the same dose (Innes et al. 1969; NCI 1968).

Statistically significant increases in hepatocellular carcinomas and adenomas in rats and hepatic carcinomas and adenomas, as well as alveolar/bronchiolar carcinomas and adenomas, in female mice have been observed following chronic oral exposure to selenium sulfide (NTP 1980c). The incidence of hepatocellular carcinomas in rats was 1/50, 0/50, and 15/49 in males and 0/50, 0/50, and 21/50 in females at 0, 3, and 15 mg selenium sulfide/kg/day, respectively. In mice, the incidences of hepatocellular carcinomas and adenomas were 15/50, 14/50, and 23/50 in males, and 0/49, 2/50, and 25/49 in females at 0, 20, and 100 mg selenium sulfide/kg/day, respectively. Selenium sulfide is a pharmaceutical compound used in some antidandruff shampoos and is not administered orally. Because selenium sulfide is not absorbed through the skin, use of shampoos containing this compound should be safe, unless one intentionally consumes the product or has open cuts or sores on the scalp or hands. Chemically, selenium sulfide and ethyl selenac are very different from the organic and inorganic selenium compounds found in foods and in the environment.

In 1975, the International Agency for Research on Cancer (IARC) evaluated the literature relating selenium to carcinogenesis in both humans and animals. The Agency stated that the available data provided no suggestion that selenium is carcinogenic in humans (IARC 1975a), and IARC subsequently assigned selenium to Group 3: not classifiable as to its carcinogenicity to humans (IARC 1987). The forms of selenium considered included sodium selenate, sodium selenite, and the organic forms of selenium contained in plant materials. Separate evaluations of ethyl selenac and methyl selenac assigned them to Group 3, also (IARC 1975a, 1987). According to EPA, selenium is not classifiable as to its carcinogenicity in humans and is rated as Group D (IRIS 2003). The evidence for selenium sulfide, however, is sufficient to classify it as Group B2 (probable human carcinogen) (IRIS 2003).

3.2.3.1. Death

No studies were located concerning death in humans after dermal exposure to selenium or selenium compounds. In a range-finding study using mice dermally exposed to selenium sulfide for a maximum of 17 applications, 8 out of 20 animals died at 714 mg selenium sulfide/kg (NTP 1980a). However, the effects noted in this study were equivocal since there was no indication that the application sites were covered to prevent ingestion. Further, severe skin damage developed, and this may have led to direct systemic absorption of the compound.

3.2.3.2. Systemic Effects

No studies were located concerning respiratory, cardiovascular, gastrointestinal, hematological, musculoskeletal, hepatic, renal, endocrine, or body weight effects in humans or other animals following dermal exposure to selenium or selenium compounds.

Dermal Effects. Skin toxicity in humans, notably skin rashes, burns, and contact dermatitis, has been reported for both acute and chronic exposure to selenium fumes and acute exposures to selenium dioxide (Middleton 1947). No effects were detected in a study of eight women exposed daily for 2 weeks to an experimental sunscreen lotion containing up to 0.003 mg/kg/day selenium as L-selenomethionine (Burke et al. 1992a). A case report of a severe allergic skin response following intermediate exposure to sodium selenite (Senff et al. 1988) is discussed under immunological effects. Single topical exposures to selenious acid resulted in purpura, inflammation around hair follicles, and a pustular rash with some ulceration in exposed workers (Pringle 1942). However, these effects may have been due to the caustic effects of the acid. A single case report of hyperpigmentation and hair loss after use of a shampoo containing 1% selenium sulfide was located (Gillum 1996), but a study of the efficacy of an antidandruff shampoo containing 1% selenium sulfide found no adverse effects after 6 weeks of use by 150 individuals (Neumann et al. 1996).

Application of 100 µL of a lotion (oil-in-water emulsion) containing 0.02% selenium as selenomethionine 3 times a week to the shaved backs of mice for 39 weeks did not result in significant dermal effects (Burk et al. 1992b). Dermal effects were also not observed in hairless mice treated in the same manner for 49 weeks.

In mice, topical application of selenium sulfide resulted in erythema and skin irritation at 29 mg/kg, acanthosis at 143 mg/kg, and severe skin damage at 714 mg/kg (NTP 1980a).

Ocular Effects. No studies were located regarding ocular effects in humans after dermal exposure to selenium or selenium compounds. However, older reports on eye contact with selenium or selenium compounds indicate that acute exposure to selenium dioxide caused ocular and conjunctival irritation, and caused severe pain, lacrimation, blurred vision, and dulled corneas upon contact (Middleton 1947). Brief exposure to clouds of selenium fumes resulted in lacrimation, irritation, and redness of the eyes (Clinton 1947).

No studies were located regarding ocular effects in laboratory animals after dermal exposure to selenium or selenium compounds.

3.2.3.3. Immunological and Lymphoreticular Effects

A 1988 case report describes a female laboratory technician who developed severely pruritic vesicles between the fingers after 6 months of exposure to a medium containing selenium. After 2 years, the severity of the symptoms increased to include eczema on the face and neck, watering eyes, and two asthma attacks within a 2-month period. Sodium selenite or the medium containing selenium were the only positive patch tests (Senff et al. 1988).

No studies were located regarding immunological and lymphoreticular effects in laboratory animals after dermal exposure to selenium or selenium compounds.

No studies were located regarding the following health effects in humans or laboratory animals after dermal exposure to selenium or to selenium compounds:

3.2.3.4. Neurological Effects

3.2.3.5. Reproductive Effects

3.2.3.6. Developmental Effects

3.2.3.7. Cancer

No studies were located regarding carcinogenic effects in humans after dermal exposure to selenium or selenium compounds.

The results of most animal studies have not indicated that elemental selenium or selenium compounds are carcinogenic when topically applied to the skin of experimental animals (NTP 1980a, 1980b; Shamberger 1970). Several studies indicate that selenium compounds may protect against effects of known dermal carcinogens (polynuclear aromatic hydrocarbons [PAHs] and ultraviolet light). Shamberger (1970) reported that topical application of a solution containing 0.0005% sodium selenide significantly reduced the number of dermal papillomas induced by painting DMBA on the shaved backs of mice. More recently, Burke et al. (1992b) reported orally and topically administered l-selenomethionine decreased ultraviolet burns and skin cancer in albino (BALB:c) and hairless pigmented (Skh:2) female mice.

Only one study was found in which tumor development was reported after topical administration of selenium ointment (Tsuzuki et al. 1960). An unspecified number of mice were exposed to an unspecified amount of ointment containing 2.5, 5.0, 7.5, or 10% elemental selenium 6 days/week for an unspecified period of time by topical administration to hip skin. Tumors developed on the base of the necks of two female mice. Ingestion of the compound was possible because the animals may have licked the ointment. No other details were reported. The study is inconclusive because of the lack of controls, short duration, and inadequate description of the study protocol and results.

The National Toxicology Program (NTP 1980a) conducted a dermal application study of selenium monosulfide. The compound was applied to the skin of groups of 50 male and 50 female Swiss mice at 0, 0.5, or 1.0 mg selenium sulfide/mouse, 3 days/week for 86 weeks. The application sites were not covered; therefore, ingestion of the test compounds was possible. The incidence of tumors in the treated groups did not differ significantly from that in the control group.

NTP (1980b) also tested Selsun, a prescription dandruff shampoo containing 2.5% selenium sulfide (also a mixture of the mono- and disulfides), for carcinogenic properties. Groups of 50 male and 50 female Swiss mice were dermally exposed to a 0, 25, or 50% solution of Selsun in distilled water 3 days/week for 86 weeks. These doses were equivalent to 0, 0.31, or 0.625 mg selenium sulfide/mouse/day. The incidences of alveolar or bronchiolar adenomas or carcinomas in male mice were significantly increased over vehicle control values, but not over untreated control values. There was no significant effect in female mice. Some ingestion of the compound was possible, because the application sites were not covered. Also, the male mice may have been susceptible to another ingredient in the shampoo (the chemical composition of the shampoo was not reported), or the bioassay may have been too short due to decreased survival to detect a carcinogenic effect in females. Male mice that received dermal application of slightly larger doses of selenium sulfide (NTP 1980a) did not develop significantly more cancers than the controls.

3.4.1.1. Inhalation Exposure

Studies regarding the absorption of selenium in humans following inhalation exposure are limited to occupational studies. Glover (1970) examined urinary selenium levels of workers employed in a selenium rectifier plant. Workers exposed to higher levels of unspecified inorganic selenium compounds in the air excreted higher levels of selenium in their urine than workers in other areas of the plant with lower concentrations of selenium in the air. Although the study indicates that selenium was absorbed from the lungs of the workers, the nonspecific exposure levels and lack of compound identification precluded an estimate of the extent and rate of absorption from the lungs. Significantly increased serum selenium levels were reported for workers at a rubber tire repair shop in Toluca City, Mexico compared with a group of unexposed individuals from the same city (Sánchez-Ocampo et al. 1996). The workers in this study were exposed to selenium (no levels reported) from vulcanized rubber, both as dust in the air and from handling the tires; thus, it is not possible to attribute absorption to a single route.

Table 3-6

Selenium Concentrations in Human Tissues^,.

Table 3-7

Biomarkers: Selenium Concentrations in Human Tissues and Fluids.

Studies using dogs and rats indicate that absorption of selenium following inhalation exposure is extensive, although the rate of absorption depends on the chemical form of selenium. In rats (Medinsky et al. 1981a) and dogs (Weissman et al. 1983), the absorption of selenium following inhalation exposure to selenious acid aerosol is approximately twice as rapid as the absorption of selenium following inhalation exposure to elemental selenium aerosol. However, Medinsky et al. (1981a) found that with either form after 4 days most of the selenium was absorbed following inhalation exposure and that the distribution of selenium in the body tissues was identical, suggesting that selenium entered the same body pool following pulmonary uptake (Medinsky et al. 1981a).

3.4.1.2. Oral Exposure

Selenium compounds are generally readily absorbed from the human gastrointestinal tract. The bioavailability of ingested selenium can be affected by the physical state of the compound (e.g., solid or solution), the chemical form of selenium (e.g., organic, inorganic), and the dosing regimen. However, in general, it appears that the degree of selenium absorption (i.e., percent of administered dose absorbed) in humans is independent of the exposure level, but that in some cases, absorption is greater when selenium deficiency exists.

In humans, absorption of sodium selenite or selenomethionine can exceed 80% for both small and relatively large doses (Griffiths et al. 1976; Thomson 1974; Thomson and Stewart 1974; Thomson et al. 1977). A total of 90–95% of a small amount of sodium selenite (0.010 mg selenium/person) administered in aqueous solution was absorbed (Thomson 1974). Absorption of a large dose (1.0 mg/person) of either sodium selenite or selenomethionine was 90–95 and 97% of the administered dose, respectively (Thomson et al. 1977). These data indicate a lack of homeostatic control over the dose range tested. Martin et al. (1989a) found no clear evidence of increased gastrointestinal absorption of selenium as sodium selenite in aqueous solution by healthy male volunteers kept on a selenium-deficient diet. Griffiths et al. (1976) reported 96–97% absorption of a single dose of 0.002 mg selenium administered as selenomethionine in solution. Similarly, Thomson et al. (1977) reported 97% absorption of a single large dose of 1.0 mg selenium administered as selenomethionine in solution to one subject. The subjects in these studies were New Zealand women.

Other studies have indicated that humans might absorb selenomethionine more efficiently than sodium selenite (Moser-Veillon et al. 1992; Swanson et al. 1991). Young et al. (1982) studied human absorption of dietary selenium in young men in the United States. The men ate either ⁷⁵Se-labeled chicken alone (0.013 mg selenium/person) or the chicken plus supplemental labeled sodium selenite (0.071 mg selenium/person in a solution mixed with the meal). Eighty percent of the selenium in the chicken meat was absorbed, but less than 30% of the selenium administered as sodium selenite was absorbed. Similarly, Robinson et al. (1978) found that 75% of selenomethionine, but only 46% of selenite, was absorbed during a 10–11-week administration of solutions providing 0.0013–0.0023 mg selenium/kg/day to New Zealand women. It is not clear why the estimated absorption of sodium selenite varied between 46 and 30% in these trials.

Experimental animals also efficiently absorb selenium compounds from the gut independent of the level of selenium exposure. Several studies have reported absorption of 80–100% in rats given dietary selenium administered as sodium selenite, sodium selenate, selenomethionine, or selenocystine (Furchner et al. 1975; Thomson and Stewart 1973). Other animal species also readily absorb orally administered selenium compounds. Furchner et al. (1975) estimated that over 90% of an oral dose of selenious acid was absorbed in mice and dogs, although monkeys absorbed less of the administered dose (amount unspecified). Using an in vivo perfusion method in which selenite was added directly to the duodenal end of the small intestine, the absorption of selenite was linearly related to concentration (slope=0.0386) in the range of 1–200 µM (Chen et al. 1993).

In one study of rats, absorption of selenite or selenomethionine into the blood stream following oral exposure occurred primarily in the duodenum and, to a lesser extent, in the jejunum and ileum (Whanger et al. 1976). Compared to the small intestine, little selenium was absorbed from the stomach (Whanger et al. 1976), and it was not determined whether absorption occurred in the large intestine. In an in vitro study using everted intestinal sacs from hamsters, Spencer and Blau (1962) found that selenomethionine was transported against a concentration gradient with the same characteristics as methionine. Selenomethionine was not found to be degraded during transport. This study suggests that in the intestines, methionine and selenomethionine share the same transport mechanism.

A comparison of absorption of selenium by selenium-depleted rats after oral administration of sodium selenate, selenomethionine, or methyl selenocysteine (from high-selenium broccoli) found that gross absorption of selenium from methyl selenocysteine was significantly lower (85%) than from sodium selenate or selenomethionine (91%); further, true selenium absorption adjusted for urinary excretion was significantly different for methyl selenocysteine, sodium selenate, and selenomethionine, with the lowest absorption for methyl selenocysteine and the highest for selenomethionine (Finley 1998). Absorption of selenium from selenomethionine was not significantly lower than from sodium selenate.

In vivo experiments with ligated rat intestines have shown that there is significantly higher absorption and transfer to the body of selenium as selenocystine or selenodiglutathione than selenium as selenite from ligated loops of ileum, but that absorption of the three forms of selenium in the jejunum was approximately similar (Vendeland et al. 1992). In vitro experiments with brush border membrane vesicles derived from rat intestines have shown dramatic differences in the uptake and binding of selenium depending on the form in which it is presented, with absorption of organic forms being much more efficient than absorption from selenite or selenate (Vendeland et al. 1992, 1994). Selenium from selenocystine or selenodiglutathione was absorbed 10 times more quickly than selenium from sodium selenite (Vendeland et al. 1992). Similarly, selenium was much more efficiently absorbed from selenomethionine than from selenite or selenate (Vendeland et al. 1994). Binding also varied between selenomethionine, selenite, and selenate, with selenite binding exceeding that of selenate by 37-fold and selenomethionine exceeding selenite by 14-fold (Vendeland et al. 1994). These studies indicate that absorption of selenium from the gastrointestinal tract of animals is pH-dependent and influenced by the presence of sulfhydral-containing compounds, and that the increased absorption of selenium with sulfhydral compounds is likely due to complex formation with these compounds.

3.4.1.3. Dermal Exposure

Dermal absorption was tested in eight women at a maximum dose of 0.0029 mg selenium/kg as selenomethionine (0.05% L-selenomethionine in a lotion). No detectable increase in serum selenium concentrations was observed, but because the concentrations tested were so low, absorption cannot be ruled out (Burke et al. 1992a). Absorption of selenium disulfide was examined using a monthly 24-hour urine specimen in 16 persons who washed their hair weekly with a 1% selenium disulfide shampoo. No differences were found from control urinary selenium levels over the 1-year exposure period (Cummins and Kimura 1971). No absorption of selenium from selenium sulfide was seen in 15 persons who applied a 2.5% selenium sulfide suspension to their torsos and allowed it to remain on the body overnight (Kalivas 1993).

Mice were treated with a maximum of 0.02% selenium as selenomethionine by topical application of a lotion 3 times per week for 39 weeks to the shaved back and ears (size of area not specified). The applied dose was 0.29 mg/kg/day. Controls received the lotion without selenium. Dermal effects were not observed in the selenomethionine-treated mice. However, treated animals had significantly higher concentrations of selenium than the controls in the liver and ventral skin away from the application site (Burke et al. 1992b). These data suggest that mice can absorb topically applied selenomethionine, but since the areas were not occluded, some oral absorption during grooming is also possible.

3.4.2.1. Inhalation Exposure

No studies were located regarding the distribution of selenium in humans after inhalation of elemental selenium or selenium compounds.

Weissman et al. (1983) reported that selenium concentrated in the liver, kidney, spleen, and lungs of dogs following inhalation exposure to selenious acid or elemental selenium aerosols.

3.4.2.2. Oral Exposure

A study of 100 paired samples of maternal and neonate hair found that the concentration in neonatal hair (0.63±0.52 µg/g) was lower than in maternal hair (1.02±1.04 µg/g), but the results were not analyzed statistically (Razagui and Haswell 1997). Levels of selenium in 30 paired samples of the hair of a mother and her child found no correlation between the selenium concentration of the hair of the mother and her child (Bermejo Barrera et al. 2000). The average level of selenium in the children’s hair (0.77±0.24 µg/g) was higher than that of their mothers (0.54±0.34 µg/g). The higher concentration of selenium in the children’s hair could represent increased absorption or retention, but no information was provided in the study as to the age of the children or to possible differences in dietary intake of selenium between mother and child.

A study in rats found that young (weanling) animals accumulated more selenium in their tissues than adults (Salbe and Levander 1989). Selenium-deficient rats were fed diets supplemented with the same amounts of selenium, as sodium selenate or L-selenomethionine, for 4 weeks. Hair and nail selenium levels in adults were 10–20% and ~50% lower, respectively, than the amounts found in weanlings. Skeletal muscle and red blood cell selenium levels were ~50 and ~35% lower, respectively, in adults than weanlings, whereas levels in the liver were generally similar between the two growth phases.

Selenomethionine caused greater deposition of selenium in the tissues than sodium selenate in both adults and weanlings, although the percent increase was similar for the two compounds in both growth phases. In rats and dogs, the selenium arising from sodium selenite administered in drinking water or in the diet is widely distributed in the body, although concentrated primarily in the liver and kidney (Furchner et al. 1975; Sohn et al. 1991; Thomson and Stewart 1973).

In most studies, selenium from selenomethionine accumulates in tissues to a greater extent than equal administered doses of selenium from selenite or selenate. Behne et al. (1991) reported higher liver and muscle selenium concentrations in rats receiving selenium orally as selenomethionine for 3 or 6 weeks than as selenite for the same length of time. Ip and Hayes (1989) reported similar results for blood, liver, kidney, and skeletal muscle. Salbe and Levander (1990b) compared distribution of dietary selenomethionine and selenate in rats and found higher selenium concentrations in plasma, erythrocytes, liver, muscle, hair, and nails in animals receiving selenomethionine. (Hair and nails have been used to gauge long-term human selenium exposure and were, therefore, included in this study.) Monkeys receiving selenomethionine in drinking water for 11 months had selenium concentrations in plasma, erythrocytes, liver, muscle, and hair that were 3–10-fold greater than monkeys receiving selenite (Butler et al. 1990). The higher levels of selenium found after selenomethionine compared to selenite treatment are likely a result of a greater retention of selenium from selenomethionine, rather than a difference in absorption. Butler et al. (1990) indicate that dietary ascorbic acid can reduce selenite absorption, but not selenomethionine absorption. Therefore, the differential effect of ascorbic acid on selenium absorption may have contributed to the difference in selenium content of tissues observed in monkeys treated with selenite, compared to monkeys treated with selenomethionine. Studies of rats indicate that the central nervous system also concentrates more selenium when administered as selenomethionine than when administered as inorganic selenium compounds (Grønbaek and Thorlacius-Ussing 1989, 1992; Zi-Jian Jie 1992).

A comparison of distribution of selenium in selenium-depleted rats after oral administration of sodium selenite, sodium selenate, selenomethionine, or methyl selenocysteine (from high-selenium broccoli) revealed that the rate of restoration of selenium in the liver and muscle was significantly slower for methyl selenocysteine than other forms of selenium (Finley 1998). The rate of repletion in muscle was significantly faster for selenomethionine than other groups, but kidney and plasma showed no significant difference in the rate of repletion for any form of selenium. The rate of repletion of glutathione peroxidase activity in the tissues was similar to the rate of repletion of the tissue itself and was slowest when methyl selenocysteine was the administered form.

Another study of distribution of selenium in selenium-deficient rats fed either sodium selenite or selenomethionine found that the concentration of selenium in blood and hair increased with administered dose, but was higher for selenium administered as selenomethionine (Shiobara et al. 1998).

A study of dietary supplementation of female pigs with 0.1 or 0.3 ppm selenium from a selenium-enriched yeast or from sodium selenite (doses not given) from 60 days before breeding until weaning found that the concentration of selenium in milk, dam, and offspring tissues increased with the dose of selenium administered and was higher when the source of selenium was the selenium-enriched yeast (Mahan and Kim 1996).

A study using pigs indicates that tissue levels of selenium do not correlate with effects. Tissue concentrations of selenium were higher in pigs fed 1.25 mg selenium/kg/day as D,L-selenomethionine than in pigs fed the same dose of selenium as A. bisulcatus or selenate, although neurological effects were more severe and occurred after fewer days of treatment with A. bisulcatus (Panter et al. 1996). The form of selenium in A. bisulcatus is unknown, although Panter et al. (1996) indicate that it is nonprotein.

In poultry, selenium is concentrated in the pancreas to a greater extent following oral administration of selenomethionine than following oral administration of sodium selenite (Cantor et al. 1975). The differential ability of the two compounds to concentrate in the pancreas of birds may explain why selenium administered as selenomethionine is more effective than the same dose of selenium administered as sodium selenite in preventing pancreatic fibrosis in chicks, a condition indicative of selenium deficiency (Cantor et al. 1975).

The distribution profiles of single oral or intravenous doses of selenium (2 mg selenium/kg as sodium selenite) administered to Wistar rats were dependent on the route of administration (Kaneko et al. 1999). Selenium concentration was highest in the kidney or liver, followed by the heart, lung, or spleen; then plasma and the brain. Oral administration produced lower doses of selenium than injection in all organs except the kidney where levels produced by the two routes were comparable (this may reflect the importance of urine as a route of excretion).

Following oral exposure, selenium is found in human milk (Brätter and Negretti De Brätter 1996; Brätter et al. 1991b; Li et al. 1999; Michalke and Schramel 1998; Moser-Veillon et al. 1992; Rodríguez Rodríguez et al. 1999; Viitak et al. 1995; Yang 1989b). Selenium is also found in the milk of mice, rats, dogs, pigs, cows, and monkeys (Abdelrahman and Kincaid 1995; Archimbaud et al. 1992; Baňuelos and Mayland 2000; Chhabra and Rao 1994; Hawkes et al. 1994; Mahan and Kim 1996; Parizek et al. 1971a). This supplies offspring with selenium during the time period in which they are fed exclusively on milk (about 6 months for humans). Transplacental transfer of selenium has been demonstrated in humans, rats, hamsters, dogs, pigs, and monkeys (Archimbaud et al. 1992; Choy et al. 1993; Hawkes et al. 1994; Jandial et al. 1976; Mahan and Kim 1996; Parizek et al. 1971a; Willhite et al. 1990).

3.4.2.3. Dermal Exposure

Although unable to detect increased selenium in human females exposed to selenomethionine dermally, Burke and coworkers found elevated liver and skin selenium concentrations in mice treated with a topical lotion containing selenomethionine applied to the shaved back and ears (size of area not specified), although since the areas were not occluded, some oral absorption during grooming is also possible (Burke et al. 1992a, 1992b). In rats, between 9 and 27% of dermally applied selenious acid was absorbed, as measured in ⁷⁵Se radioisotope studies (Medinsky et al. 1981b).

3.4.2.4. Other Routes of Exposure

In humans, selenium has been found to be widely distributed to organs and tissues following injection of sodium selenite, sodium selenate, and selenomethionine, with the highest concentrations generally found in the liver and kidneys (Ben-Porath and Kaplan 1969; Cavalieri et al. 1966; Jereb et al. 1975; Lathrop et al. 1972). In studies involving injection of radiolabelled selenium, the pancreas accumulated high concentrations of radiolabelled selenium immediately following injection, but within hours, the selenium rapidly disappeared from this organ (Lathrop et al. 1972). Using an in vitro, dually perfused, human term placenta, selenite has also been shown to cross the human placenta (Eisenmann and Miller 1994). Further, following intravenous injection, ⁷⁵Se from selenomethionine was found to cross the near-term human placenta (Jandial et al. 1976).

There is a rapid decline in serum selenium levels 1 hour after intravenous administration of sodium selenite or sodium selenate to humans (Burk 1974; Nelp and Blumberg 1965). Burk (1974) found that 50% of the plasma selenium was protein-bound within the first 2 hours after administration; 85% was bound within 4–6 hours after administration; and 95% was bound after 24 hours. Circulating alpha-2 globulins have been reported to have the greatest affinity for selenium (Hirooka and Galambos 1966a).

Burk (1974) found that lipoproteins, primarily the very low density lipoprotein (VLDL) and the low-density lipoprotein (LDL) fractions, were also involved in selenium binding.

In vitro studies of human plasma and whole blood incubated with sodium selenite have indicated that selenite is accumulated in erythrocytes by an active transport mechanism (Lee et al. 1969). Several studies indicate that the selenite is chemically altered in the erythrocyte and then transported back into the plasma, where the selenium metabolite binds to plasma proteins (Burk 1974; Hirooka and Galambos 1966a; Lee et al. 1969).

A high degree of protein binding of selenium in the plasma has also been demonstrated in experimental animals. Sandholm (1973) found that selenite administered intravenously to mice can be metabolically altered by erythrocytes to a form that binds to plasma proteins. In mice, rats, and dogs, selenite initially binds to albumin. Later, selenite can be found bound to alpha and gamma globulins in rats and to alpha-2 and beta-1 globulins in dogs (Imbach and Sternberg 1967; Sternberg and Imbach 1967).

3.4.4.1. Inhalation Exposure

Following acute inhalation exposures to selenium compounds, humans excrete some of the absorbed dose in the expired air (Glover 1970), but no studies were located that actually quantified the rate of excretion or identified the selenium compounds in the expired air of humans.

3.4.4.2. Oral Exposure

Several human studies have indicated that the rate of urinary excretion is most rapid in the first 24 hours following oral administration or intravenous injection of sodium selenite (Kuikka and Nordman 1978; Thomson and Stewart 1974). Thomson and Stewart (1974) found that >6% of a trace dose (0.01 mg selenium) of orally administered sodium selenite was excreted in the urine within 24 hours of administration, whereas 64–73% of a 1-mg dose of selenium was excreted in the first 24 hours (Thomson 1974). Thomson et al. (1977) also found that a lower proportion of the selenium from an oral dose of 0.1 mg selenium administered as selenomethionine was excreted in the 24-hour urine than from a larger dose (1.0 mg selenium). Similarly, low selenium New Zealand residents excreted proportionally less selenium in their urine than North Americans of higher selenium status (Robinson et al. 1985), and there is limited evidence of such adaptation to selenium intake in some animal studies (Jaffe and Mondragon 1969, 1975; WHO 1987). Thus, when higher amounts of selenium are administered, a higher proportion of the selenium is excreted in the urine during the first 24 hours following exposure.

Figure 3-6

Activation and Reduction of Selenate to Selenite in Yeast Saccharomyces cerevisiae.

Decreasing urinary or fecal excretion appears to be the homeostatic mechanism by which the body retains greater amounts of selenium. Martin et al. (1989a) observed greater retention of selenium by individuals maintained on a selenium-deficient diet. This increase in retention was correlated with a decrease in fecal elimination. Similarly, the increased retention of selenium from selenomethionine compared to selenite was correlated with decreased elimination (Swanson et al. 1991). Lactating women have a greater retention of selenium from selenomethionine compared to selenite and a decreased urinary elimination (Moser-Veillon et al. 1992). Muscle activity seems to influence urinary excretion of selenium as demonstrated by the doubling of selenium concentration in the urine of women following vigorous exercise (Oster and Prellwitz 1990).

Less information is available regarding the elimination of selenium in the feces of humans than in the urine of humans. However, levels of fecal excretion of selenium have been reported to be similar to levels of urinary selenium excretion when dietary levels of selenium are not excessive (Patterson et al. 1989). Over a 14-day period, Stewart et al. (1978) found urinary elimination of selenium to average 0.013 mg selenium/day and fecal elimination of selenium to average 0.011 mg selenium/day in four New Zealand women exposed to 0.024 mg selenium/day in their normal diets. Balance data on 27 healthy U.S. adults (12 men and 15 women) similarly indicated an approximately even split between urine and fecal selenium excretion (Levander and Morris 1985). Determination of selenium balance at four time points (spring, summer, fall, and winter) showed respective average levels of selenium in the urine and feces of 48±2 and 34±1 µg/day in the men, and 39±1 and 23±1 µg/day in the women. Plasma selenium levels remained essentially constant during the year and were similar in the men and women, averaging 136±4 and 133±4 ng/L, respectively. Although the U.S. men consumed more selenium in the diet than the women, their selenium balance (8±4) was less positive than the women (12±3) because they tended to excrete more in the feces (Levander and Morris 1985). It has been suggested that some of the selenium content in feces can be attributed to biliary excretion (Levander and Baumann 1966a, 1966b).

In humans, whole body retention studies following oral administration of sodium selenite have indicated that selenium elimination is triphasic (Thomson and Stewart 1974). During the initial phase, which lasted about 1 week, elimination of selenium was rapid, with a half-life of approximately 1 day (Thomson and Stewart 1974). In the second phase, which also lasted approximately 1 week, selenium elimination was slower, with a half-life of 8–9 days. In the third phase, selenium elimination was much slower, with a half-life estimated to be 115–116 days. The first two elimination phases correspond to the fecal elimination of nonabsorbed selenium and the urinary excretion of absorbed but unutilized selenium (Thomson and Stewart 1974). Selenomethionine elimination is also triphasic; however, its terminal half-life is longer than that of sodium selenite. The average half-lives of selenomethionine for the three phases were measured to be approximately 0.4–2, 5–19, and 207–290 days, respectively (Griffiths et al. 1976). An examination of elimination data from 44 pigs exposed to excess selenium as sodium selenite in feed was found to fit a one-compartment model of selenium elimination (Davidson-York et al. 1999). Serum selenium levels were monitored over a period of 46 days beginning 1–14 days after termination of exposure to the feed containing excess selenium. Data were not adequate to depict the initial distribution phase, but a geometric mean elimination half-life of 12 days was calculated. It is likely that the period of elimination included in this study corresponds to the second phase described by Thomson and Stewart (1974).

The chemical form of selenium may play a role in determining how rapidly selenium is excreted in the urine. In rats, the rate of urinary excretion of selenium has been found to be greater following oral administration of sodium selenite than of selenomethionine (Thomson and Stewart 1973). A comparison of excretion of selenium by selenium-depleted rats after oral administration of sodium selenate, selenomethionine, or methyl selenocysteine (from high-selenium broccoli) found that excretion of selenium from methyl selenocysteine or selenomethionine was significantly lower than from sodium selenate; further, that there was no significant difference between secretion of selenium from methyl selenocysteine and selenomethionine (Finley 1998). This may contribute to the greater retention of selenium from selenomethionine, than from inorganic selenium (Martin et al. 1989a). However, another study of excretion of selenium from rats fed selenium as either sodium selenite or selenomethionine found that excretion of selenium increased with administered dose, but was similar for both forms of selenium (Shiobara et al. 1998).

As exposure to oral L-selenomethionine increased in macaques, the amount of selenium eliminated in the urine/day increased, as did the maximum rate of urinary excretion. However, the percentage of administered dose appearing in the urine decreased with an increase in dose (Hawkes et al. 1994).

3.4.4.3. Dermal Exposure

No studies were located regarding the excretion of selenium by humans or other animals after dermal exposure to elemental selenium or selenium compounds.

3.4.4.4. Other Routes of Exposure

Whole body retention studies in sheep following injection of selenium have indicated that selenium excretion in animals follows a triexponential profile (Blodgett and Bevill 1987b; Ewan et al. 1967). In a 2-week study, Blincoe (1960) estimated the half-life for ⁷⁵Se in rats following intraperitoneal injection of ⁷⁵Se-labeled sodium selenite (0.93 mg selenium/kg). Initially, the excretion of selenium was rapid, with a half-life of approximately 0.8 day; the second phase of excretion was slower, with a half-life of 13 days. These results parallel the initial phases of selenium excretion seen in humans. The abbreviated duration of the Blincoe (1960) study did not permit the determination of a terminal elimination phase half-life. In rats, Ewan et al. (1967) found the final phase of elimination of selenium following a single subcutaneous injection of sodium selenite to be dose independent (from 0.008 mg selenium/kg to 2 mg selenium/kg), with a half-life of 65–78 days. Blodgett and Bevill (1987b) found the elimination rate of selenium in sheep during the second phase following a single intramuscular injection of sodium selenite to be dose dependent, with larger doses resulting in longer half-lives (i.e., doses of 0.4, 0.6, 0.7, or 0.8 mg selenium/kg resulting in half-lives for selenium elimination of 6.3, 8.8, 15.1, and 20.4 hours, respectively). The reasons for the decreasing elimination rate with increasing dose during the second phase are not clear.

Dietary levels of selenium and the individual’s selenium nutritional status are the most important factors that influence the route and rate of selenium excretion. Selenium excretion in expired air is only significant when exposures to selenium are high. Rats injected subcutaneously with sodium selenite at doses of 2.2–5.4 mg selenium/kg excreted 41–62% of the administered selenium in exhaled air, whereas rats injected with sodium selenite at doses of 0.005–0.9 mg selenium/kg excreted only 0.2–11% of the administered selenium in expired air (McConnell and Roth 1966; Olson et al. 1963). As the amount of administered sodium selenite increased, the percent of the administered selenium excreted in the urine decreased (from approximately 22–33% of the administered selenium at doses of 0.005–0.9 mg selenium/kg to 3–14% of the administered selenium at doses of 3.1–5.4 mg selenium/kg) (McConnell and Roth 1966). Selenium in the feces was not measured in this study. Burk et al. (1972) found that as the dietary level of sodium selenite was increased, a larger proportion of an injected tracer dose of selenium (as sodium selenite) was excreted. At a dietary level of 0.005 mg selenium/kg, approximately 60% of the injected selenium had been excreted in the first 35 days following administration. At a dietary level of 0.05 mg selenium/kg, over 94% of the injected selenium had been excreted over the same period of time.

In experimental animals, other factors that can cause an increase in selenium levels in expired air are higher dietary levels of selenium, protein, or methionine (Ganther et al. 1966). Phenobarbital induction of microsomal enzymes has also led to increased exhalation of selenium following intravenous administration of sodium selenite (Sternberg et al. 1968).