Response of Selenium Status Indicators to Supplementation of Healthy North American Men with High-Selenium Yeast

  • Wayne Chris Hawkes
  • B. Diane Richter
  • Zeynep Alkan
  • Elaine C. Souza
  • Monique Derricote
  • Bruce E. Mackey
  • Ellen L. Bonnel


The essential nutrient selenium is required in microgram amounts [recommended dietary allowance (RDA) = 55 μg/day, 699 nmol/day] and has a narrow margin of safety (upper tolerable intake limit = 400 μg/day, 5 μmol/day). We conducted a randomized placebo-controlled study of high-selenium yeast, the form used in most supplements (300 μg/day, 3.8 μmol/day), administered to 42 free-living healthy men for 48 weeks. Dietary intakes of selenium, macronutrients, and micronutrients were not different between groups and did not change during the study. Supplementation more than doubled urinary selenium excretion from 69 to 160 μg/day (876 to 2,032 nmol/day). Urinary excretion was correlated with recent selenium intake estimated from 3-day diet records: urinary selenium excretion = 42 μg/day (533 nmol/day) + 0.132 × dietary selenium intake, p < 0.001. Dietary selenium intake was not significantly correlated with the other indicators of selenium status, presumably because urinary selenium excretion reflected recent intake, and tissue selenium was homeostatically controlled. After 48 weeks of supplementation, plasma selenium was increased 60% from 142 to 228 μg/l (1.8 to 2.9 μmol/l), and erythrocyte selenium was approximately doubled from 261 to 524 μg/l (3.3 to 6.6 μmol/l). Selenium concentrations increased more modestly in hair (56%) and platelets (42%). Platelets were the only blood component in which glutathione peroxidase activity was significantly related to selenium content. Selenium levels decreased rapidly after the end of supplementation, and there were no significant differences in selenium status indicators between groups by week 96. The absorption, distribution, and excretion of selenium from high-Se yeast were similar to selenium in foods.


Selenium supplementation Yeast Nutritional status Diet assessment 





recommended dietary allowance


upper tolerable intake limit


National Health and Nutrition Examination Survey


Selenium (Se) is an essential nutrient required in microgram quantities by humans, with a recommended dietary allowance (RDA) of 55 μg/day recently set for adults [1]. The main biological function of Se in mammals is as a source of selenocysteine for the synthesis of approximately 25 highly conserved selenoproteins [2], the most well-known of which are the four glutathione peroxidases that reduce hydrogen peroxide and/or organic hydroperoxides to water or alcohols at the expense of glutathione. The current RDA was set at a level sufficient to maintain maximal glutathione peroxidase activity in blood plasma [1]. The biological functions and activities of most mammalian selenoproteins remain to be discovered. Among the selenoproteins with known activities are methionine sulfoxide reductase B, the three thioredoxin reductases, and the three iodothyronine deiodinases.

Se deficiency primarily affects growing animals and is associated with liver necrosis in rats, pancreatic atrophy in chicks, gizzard myopathy in turkey poults, and “white muscle disease” in sheep. Two Se-responsive conditions have been documented in humans: Keshan disease is a fatal cardiomyopathy associated with Coxsackie virus B infections that afflicts children and women of child-bearing age in the Se-deficient areas of China and can be completely prevented by Se supplementation. Peripheral skeletal myopathy is observed in total parenteral nutrition patients provided feeding solutions without Se and is completely reversed by addition of Se, and is thus the only specific Se-deficiency syndrome identified in humans.

Assessment of Se status can be problematic. The Se intake of individuals is difficult to estimate accurately from food frequency questionnaires or food records, unless the foods are analyzed for their Se contents [3, 4]. This is due primarily to the variability in the Se content of individual foods, which is determined mostly by the soil Se concentration where the foods are produced [5]. Hair Se reflects dietary intake and provides an indicator of long-term intake, but its application is limited in industrialized countries, where the use of Se-containing anti-dandruff shampoos is common. Urinary Se excretion follows dietary intake closely but only reflects recent intakes and requires collection of complete 24-h specimens to be accurate. Se levels in blood and blood components have been used extensively for status assessment in individuals and populations with widely different intakes, and the activity of glutathione peroxidase in plasma and platelets can be useful in populations with low Se intakes (<60 μg/day, 762 nmol/day).

Because of reports that Se supplements may decrease the risk of some kinds of cancer [6, 7, 8, 9] sales of Se supplements, mostly as high-Se yeast tablets, grew to an estimated $60 million in 2003 [10]. News of a report on Se’s potential to prevent mutation of the H5N1 avian influenza virus [11] led to a reported 60% increase in Se supplement sales in the UK in 2 weeks [12]. The upper tolerable intake level (UL) for Se of 400 μg/day (5 μmol/day) was derived from a “no observed adverse effects level” of 800 μg/day (10 μmol/day) and an uncertainty factor of 2 to be protective against Se toxicity symptoms such as hair loss and nail sloughing in sensitive individuals [1]. Because the UL for Se is only seven times the RDA, it is important to understand how indicators of Se status respond to intakes near the UL. We administered high-dose Se supplements (300 μg/day, 3.8 μmol/day) in the form of high-Se yeast to healthy men for 48 weeks and measured the responses of Se status indicators in blood and urine as the subjects equilibrated to the high Se intake. We then followed the subjects for an additional 48 weeks during which they consumed their usual Se intake.

Subjects and Methods


Fifty-four healthy, non-smoking men, aged 18–45 years, with body mass index between 17.5 and 29.5 kg/m2, were randomized to treatment in this study. Potential volunteers were given a physical examination by a nurse practitioner and determined to be in good health. Inclusion criteria were: self-reported absence of disease (hypertension, diabetes, sexually-transmitted disease, cancer, etc.); and clinically normal blood count, blood chemistries, thyrotropin, and semen analysis. Exclusion criteria were: tobacco smoking; positive blood test for HIV, hepatitis B, syphilis, or positive urine tests for drugs of abuse (barbituates, benzodiazapines, cocaine metabolites, opiates, amphetamines, and cannabinoids); intention to father children within the next year; use of Se shampoos, Se supplements providing more than 50 μg/day, thyroid medications, weight-loss drugs, or anabolic steroids; more than a 10 lb weight change within the last 6 months; regular use of a hot tub; and exercise or physical training in excess of three 1-h sessions per week. Subjects were paid for their participation. The study protocol was reviewed and approved by the Institutional Review Board of the University of California at Davis School of Medicine, and informed consent was obtained in writing from all subjects.

Experimental protocol

Potential subjects meeting the recruitment criteria were enrolled into a run-in period lasting 3–6 weeks, during which baseline measurements were obtained, and compliance was assessed. Non-compliant subjects were dismissed before randomization and are not reported further. Two subjects at a time were randomized to treatment from July 2000 to November 2002, with one subject from each pair randomly assigned to each treatment group by coin flip. Fifty-four men satisfactorily completed the run-in period and were randomized to receive placebo yeast tablets or high-Se yeast tablets for 48 weeks. Neither the subjects nor the study staff were aware of subjects’ treatment assignments. Subjects took their first tablet the same day that all baseline measurements were completed. Subjects visited the Center at least once every 6 weeks during the 48 week supplementation period and then returned again 24 and 48 weeks after the end of supplementation. Visits were scheduled relative to each subject’s first day of supplementation. When necessary, visits were re-scheduled within 1 week of the original planned date, with the exception of the 48 weeks visits. Because this visit represented the end of treatment, subjects were continued on supplements up to three additional weeks in order to obtain these critical endpoint measurements. When visits were missed, the next visit was scheduled based on the first day of supplementation to restore the original schedule. Subjects were instructed to take two tablets if they forgot to take a tablet the previous day and never to take more than two tablets in 1 day or try to make up for lapses of more than a single day. Unused tablets were counted at each visit and were collected at the end of the treatment period to measure compliance.

At every visit during the 48 week treatment period, subjects were interviewed to determine the number of days tablets were not taken, the nature and duration of any illnesses, and if there had been any changes in diet, appetite, activity or lifestyle since the last visit. At every visit during the treatment period, subjects were counseled not to make any changes to their diets, normal activities, or lifestyles. During the post-treatment follow-up period, subjects were instructed to ignore these precautions and to do as they wished. Subjects kept a diary in which they recorded all illnesses, colds, coughs, sore throats, sick days missed from work, and any other health problems or symptoms they experienced. These records were reviewed by a medical assistant and discussed with the subjects at each visit. Subjects completed the Profile of Mood States-Bipolar Form questionnaire every 6 weeks and the Medical Outcomes Study 36-Item Short Form questionnaire (“SF-36”, Quality Metric, Lincoln, RI) every 12 weeks during the treatment period to monitor their overall health and quality of life. During the run-in period, at the end of the 48 weeks treatment period, and again at 96 weeks, subjects completed the Holmes-Rahe Social Readjustment Scale questionnaire to assess their social stress levels. Samples of blood, scalp hair, and 24 h urine collections were obtained twice during the run-in period and then every 6 weeks during the 48 week treatment period, and again at 72 and 96 weeks.

Se supplements

Supplements were provided as high-Se Baker’s yeast (Saccharomyces cerevisiae, strain PN0056) grown aerobically using a Pharmacopoeia-controlled growth medium containing sodium selenite (SelenoPrecise™, 300 μg Se per tablet, 3.81 μmol Se, Pharma Nord, Denmark). The Se in SelenoPrecise™ yeast is present primarily as selenoamino acid derivatives, with selenomethionine (the Se homolog of methionine) accounting for ca. 81% of the total Se and inorganic Se accounting for less than 1% [13]. Placebo tablets were compounded identically, except using the same yeast grown without added Se (≤1.3 μg Se per tablet, 16.5 nmol Se). Tablets contained 0.5 g of spray-dried yeast in an inert binder and were coated with titanium dioxide for an identical appearance, smell and taste. Tablets were provided in 28-tablet bubble packs.

Dietary intake assessment

The intake of Se from foods was estimated from 3-day diet records. Subjects were trained by a registered dietitian how to estimate the amount and type of foods eaten. During the run-in period and before the first set of 3-day diet records was collected, subjects maintained 2-day diet records that were reviewed by the dietitian for accuracy and discussed with each subject as part of their training. These records were not used in the final analysis. Twice during the run-in period and then at 24, 48, 72, and 96 w, subjects kept a written record of all foods eaten for a 3-day period, always including at least one weekday (Monday–Friday) and at least one weekend day (Saturday and Sunday). Every diet record was reviewed by a registered dietitian in an interview with the subject for completeness and to resolve any uncertainties regarding food identifications or quantities. Records were analyzed for nutrient contents with the Minnesota Nutrition Data System 5.0 (Nutrition Coordinating Center, Division of Epidemiology and Community Health, University of Minnesota, Minneapolis, MN, USA), using food composition data derived primarily from the USDA National Nutrient Database for Standard Reference [14] to calculate dietary intakes of Se, macronutrients, vitamins, and minerals.

Laboratory measurements

Blood samples were collected in the mornings after an overnight fast. Serum was prepared from blood drawn into serum separator tubes (Vacutainer® SST, BD Diagnostics, Franklin Lakes, NJ, USA), and the serum was separated by centrifugation and refrigerated until analyzed for sodium, potassium, chloride, carbon dioxide, urea nitrogen, creatinine, glucose, calcium, protein, albumin, alkaline phosphatase, aspartate aminotransferase, bilirubin, total cholesterol, low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol, and triglycerides each night at a Clinical Laboratory Improvement Amendments (CLIA)-certified reference laboratory (University of California at Davis School of Medicine Pathology Laboratory, Sacramento, CA, USA). Blood drawn into EDTA-coated tubes was centrifuged at 190×g for 15 min at 4°C, and the platelet-rich plasma was withdrawn with a plastic pipette. Two 40 μL aliquots were counted in a hematology analyzer (Serono-Baker Diagnostics, Allentown, PA, USA) to determine platelet numbers, and the platelets were collected by centrifugation at 950×g for 10 min at 4°C, and then washed once with ice-cold Hank’s Buffered Salt Solution (HBSS) and re-suspended in HBSS. Aliquots of serum, plasma, washed erythrocytes, and platelets were stored at −70°C until analyzed. Hair was clipped as close to the skin as practicable from the same approximately 1 in.2 area at the back of the neck each time, and the samples were stored in paper bags at room temperature and analyzed without further treatment. Aliquots of urine were acidified by addition of 0.5% (v/v) concentrated hydrochloric acid and stored at −20°C until analyzed for Se. Aliquots of unacidified urine were stored at −70°C until analyzed for 15-isoprostane F2t.

Se concentrations were measured in blood plasma, washed erythrocytes, platelets, hair, and urine by direct fluorometry or high-performance liquid chromatography (HPLC) of the fluorescent derivative formed from reaction with diaminonaphthalene after digestion in a 5:2 (v/v) nitric–perchloric acid mixture [15]. Calibration standards were prepared with National Institute of Standards and Technology (NIST) Standard Reference Material (SRM) 3149-Selenium Standard Solution, and the performance of each analytical run was validated by analysis of NIST SRM 1577a-Bovine Liver (certified value, 0.71 ± 0.07 μg/g; mean ± SD, 0.705 ± 0.075 μg/g) and duplicate samples of frozen pooled human plasma [within-run relative standard deviation (RSD), 4.0%; between-run RSD, 8.1%]. Analytical runs that included urine or hair included validation samples of NIST SRM 2670-Freeze-dried Urine (certified value, 0.46 ± 0.03 μg/ml; mean ± SD, 0.442 ± 0.022 μg/ml) or Certified Reference Material 397-Human Hair (Community Bureau of Reference, Commission of the European Communities, Brussels, Belgium; certified value, 2.0 ± 0.08 μg/g; mean ± SD, 1.78 ± 0.11 μg/g), respectively.

Glutathione peroxidase activity in blood plasma, erythrocytes and platelets was determined using the glutathione reductase-coupled assay [16] with 1.36 mM glutathione and 0.5 mM cumene hydroperoxide as substrates, and one unit of activity was defined as oxidation of 1 μmol of NADPH per minute. Frozen erythrocytes were diluted 1:19 in phosphate-buffered saline for glutathione peroxidase assays, and hemoglobin in the diluted samples was determined spectrophotometrically with Drabkin’s reagent [17]. Protein in plasma and platelet suspensions was measured by the Lowry method [18]. Oxidative damage was assessed by measuring free 15-isoprostane F2t in unacidified urine samples with a commercial enzyme-linked immunoassay (EA-85, Oxford Biomedical Research, Oxford, MI, USA).

Statistical analysis

The Box-Cox approach was used to estimate a power transformation that would stabilize the variances among groups and times for those variables with heterogeneity of variance [19]. In some cases, a power transformation was unsuccessful in stabilizing the variances, so the rank transformation was used. The SAS (version 9.1.3) proc mixed was used to fit a repeated-measures model [20]. Group, time and the interaction, and for dietary intake, variables day of year and daily maximum temperature as covariates, were the fixed effects. Subject within groups was the random effect. A first-order autoregressive covariance structure was used to account for the dependencies among the repeated measures. Single degree of freedom contrasts were used to compare the baseline averages with the treatment and follow-up measurements separately for the two groups, and the Bonferroni adjustment for multiplicity was applied [21]. Adjusted probabilities less than 0.05 were considered significant.


Subjects and protocol compliance

Forty-two subjects (22 in the high-Se yeast group and 20 in the placebo yeast group) finished the 48 weeks treatment period and completed all study procedures. The results presented herein represent the data from those 42 subjects. Seven subjects dropped out of the placebo group, and five subjects dropped out of the high-Se group. None of the 12 subjects that dropped out of the study quit for reasons related to the treatment. Approximately 20 subjects in the high-Se group and 18 subjects in the placebo group returned for the post-treatment visits at 72 and 96 weeks. The initial anthropometric characteristics and Se status of the 42 subjects who completed the study are shown in Table 1. The groups were well-balanced, and there were no significant differences between the groups with respect to these characteristics.
Table 1

Initial Characteristics of Subjects at Beginning of Run-in Period


Placebo yeasta

High-Se yeastb





Age (year)

31 ± 8.1


31 ± 9.4


Height (cm)

177 ± 7.6


180 ± 7.8


Weight (kg)

77.5 ± 12.4


76.0 ± 9.5


Body mass index (kg/m2)

24.6 ± 3.0


23.5 ± 2.2


Blood plasma [Se], (μg/l; 1 μg Se = 12.7 nmol)

143 ± 18


142 ± 19


Dietary Se intake (μg/day)

138 ± 39


137 ± 42


Dietary energy intake (MJ/day)

10.8 ± 2.0


11.3 ± 3.0


an = 20

bn = 22

cValues are means ± SD

Only 26 subjects brought the unused tablets to every visit, and 35 subjects brought unused tablets to some of the visits. Based on these partial counts of unused tablets, compliance with the daily dosing regimen was 93 ± 5.3% (range 77–100%). Three subjects in the placebo group had hair Se concentrations exceeding 10 μg/g (127 nmol/g) during the treatment period, indicating use of Se-containing shampoos, and their hair Se measurements exceeding 2 μg/g (25 nmol/g) were therefore excluded. None of the subjects in the high-Se group had hair Se concentrations exceeding 2 μg/g during the treatment period. One subject in each group had excessive hair Se concentrations, indicating Se shampoo usage during the post-treatment period, and those measurements were excluded.

There were no significant changes or group differences during the treatment period in any of the dimensions of health status measured with the SF-36 (data not shown). However, the norm-based score for “General Health” in the entire cohort declined from a mean of 57.8 during the treatment period to 53.9 during the post-treatment follow-up period (p = 0.035), and the score for “Mental Health” declined from 53.2 during treatment to 50.2 post-treatment (p = 0.050). There was no significant effect of Se on serum sodium, potassium, chloride, carbon dioxide, urea nitrogen, creatinine, glucose, calcium, protein, albumin, alkaline phosphatase, aspartate aminotransferase, bilirubin, total cholesterol, LDL cholesterol, HDL cholesterol, or triglycerides; nor was plasma Se correlated with any of these blood parameters (data not shown).

Se status

The average pre-study Se concentration in blood plasma in our subjects was about 14% higher than the average serum Se concentration reported for US men in the Third National Health and Nutrition Examination Survey (NHANES; 124.5 μg/l) [22]. There were no significant changes in any of the measures of Se status in the placebo group during or after the treatment period (Table 2 and Figs. 1, 2, 3, 4, and 5). In the high-Se group, Se increased in plasma, erythrocytes, platelets, hair, and urine, and appeared to plateau by 24 weeks of treatment. Glutathione peroxidase activity was not affected by Se supplementation. Urinary Se excretion increased most rapidly, reaching a plateau by 6 weeks, followed by plasma Se and platelet Se that plateaued at 12 weeks, whereas hair Se and erythrocyte Se continued to increase until 24 weeks. The change in response of each Se status measure in the supplemented group was divided by the pooled standard error (similar to a t statistic) to compare their relative sensitivity to Se intake, an important determinant of their usefulness in population studies. Erythrocyte Se (t = 8.5) and hair Se (t = 8.0) were the most discriminating indicators using this criterion, followed by plasma Se (t = 6.1), platelet Se (t = 5.8), and urine Se (t = 3.6). Platelet glutathione peroxidase activity did not respond significantly to Se supplementation. Tissue and urinary Se in the supplemented subjects decreased during the post-treatment period and returned to levels indistinguishable from the placebo group by 96 weeks.
Fig. 1

Plasma Se. Points represent the mean concentration (±SEM) for subjects treated with placebo yeast (filled square) or high-Se yeast (filled circle). Asterisks designate the time points at which the group means were significantly different (1 mg Se = 12.7 μmol)

Fig. 2

Erythrocyte Se. Points represent the mean concentration (±SEM) for subjects treated with placebo yeast (filled square) or high-Se yeast (filled circle). Asterisks designate the time points at which the group means were significantly different (1 mg Se = 12.7 μmol)

Fig. 3

Hair Se. Points represent the mean concentration (±SEM) for subjects treated with placebo yeast (filled square) or high-Se yeast (filled circle). Asterisks designate the time points at which the group means were significantly different (1 μg Se = 12.7 nmol)

Fig. 4

Platelet Se. Points represent the mean concentration (±SEM) for subjects treated with placebo yeast (filled square) or high-Se yeast (filled circle). Asterisks designate the time points at which the group means were significantly different (1 μg Se = 12.7 nmol)

Fig. 5

Urinary Se excretion. Points represent the mean daily Se excretion in urine (±SEM) for subjects treated with placebo yeast (filled square) or high-Se yeast (filled circle). Asterisks designate the time points at which the group means were significantly different (1 μg Se = 12.7 nmol)

Table 2

Se Status and Lipid Peroxidation


Placebo yeast (mean ± SD, n = 20)

High-Se yeast (mean ± SD, n = 22)

Se effecta (p)


48 weeksc

96 weeksd


48 weeks

96 weeks

Plasma Se (μg/l; 1 μg Se = 12.7 nmol)

146 ± 19

141 ± 18

148 ± 22

142 ± 19

228 ± 63

158 ± 32


Erythrocyte Se (μg/l)

264 ± 49

262 ± 36

268 ± 43

261 ± 35

524 ± 141

307 ± 69


Platelet Se (μg/g protein)

3.42 ± 0.58

3.28 ± 0.59

3.12 ± 0.48

3.38 ± 0.45

4.80 ± 1.06

3.32 ± 0.50


Hair Se (μg/g)

0.65 ± 0.10

0.67 ± 0.13

0.70 ± 0.12

0.64 ± 0.08

1.00 ± 0.19

0.66 ± 0.13


Urine Se (μg/day)

53 ± 21

59 ± 28

62 ± 33

69 ± 48

160 ± 109

65 ± 35


Plasma GPX (U/ml)

2.91 ± 0.43

2.99 ± 0.50

3.03 ± 0.64

3.16 ± 0.37

3.22 ± 0.69

3.04 ± 0.56


Erythrocyte GPX (U/g hemoglobin)

431 ± 93

424 ± 96

425 ± 105

427 ± 132

421 ± 119

412 ± 122


Platelet GPX (U/mg protein)

5.1 ± 1.1

5.2 ± 1.2

4.9 ± 1.1

5.2 ± 1.2

5.7 ± 1.6

4.7 ± 1.3


Urinary 15-isoprostane F2t (μg/day)

8.5 ± 8.5

6.1 ± 4.3


5.8 ± 5.1

5.6 ± 4.2



GPX Glutathione peroxidase

aRepeated measures analysis of variance: Se main effect or Se × time interaction

bValues at the end of the run-in period before starting supplementation

cValues at the end of the 48 weeks supplementation period

dValues 48 weeks after the end of supplementation

Glutathione peroxidase activity was not correlated with Se concentration in plasma or erythrocytes at any time in the study. Platelet glutathione peroxidase activity was not significantly correlated with platelet Se during the run-in period, but from 6 to 48 weeks of treatment, the correlation was significant (r2 = 0.06, p < 0.001; Fig. 6).
Fig. 6

Platelet glutathione peroxidase specific activity and platelet Se concentration from 6 weeks to 48 weeks of treatment. The line represents the regression of GPX on Se (GPX = 0.441 × Se + 3.52, r2 = 0.096, p < 0.001) and the dashed lines represent the 95% confidence intervals (1 μg Se = 12.7 nmol)

Dietary Se intakes

The average estimated Se intake from the subjects’ self-selected diets (134.8 μg/day; Table 3) was similar to typical US intakes of men aged 20–39 years (134.6 μg/day) and 40–59 years (137.2 μg/day) measured in the 1999–2000 National Health and Nutrition Examination Survey [23]. Forty-two percent of dietary Se came from grains, 39% came from meat, fish and eggs, 11% came from dairy products, and 6% came from fruits and vegetables. Urinary Se excretion was the only measure of Se status that was significantly related to dietary Se intake estimated from food records (Fig. 7). Regression of urinary Se excretion against dietary Se intake within the placebo group produced a slope of 0.134 and an intercept of 40 μg/day (508 nmol/day; r2 = 0.07, p = 0.023). Regression of urinary Se excretion against total Se intake (diet plus supplements) for the high-Se subjects yielded a similar slope and intercept of 0.158 and 35 μg/day (445 nmol/day), respectively (r2 = 0.07, p = 0.014). This relationship was stronger when all subjects were included, yielding a slope of 0.152 and an intercept of 37 μg/day (470 nmol/day; r2 = 0.14, p < 0.001).
Fig. 7

Urinary Se excretion and dietary Se intake. Open circles represent measurements from subjects in the placebo group and filled circles represent the high-Se group. The regression line included all points (Urine Se = 0.152 × Diet Se + 37.3 μg, r2 = 0.14, p < 0.001; 1 μg Se = 12.7 nmol)

Table 3

Dietary Intake of Fat and Antioxidant Nutrients

Diet component, intake/day

Placebo yeast (mean ± SD)

High-Se yeast (mean ± SD)

Se effecta (p)

Baselinebn = 20

48 weekscn = 20

96 weeksdn = 16

Baseline n = 21

48 weeks n = 22

96 weeks n = 20

Saturated fat (g)

23 ± 7.6

29 ± 11

29 ± 9.9

31 ± 13

34 ± 15

32 ± 20


Monounsaturated fat (g)

29 ± 10

32 ± 12

35 ± 18

34 ± 14

34 ± 13

31 ± 15


Polyunsaturated fat (g)

15 ± 5.3

16 ± 7.7

18 ± 12

19 ± 9.6

18 ± 7.8

15 ± 5.3


Vitamin A (IU)

8239 ± 6514

7724 ± 6390

7436 ± 6065

10671 ± 9142

7852 ± 5847

7251 ± 4423


Vitamin C (mg)

245 ± 403

171 ± 187

156 ± 128

169 ± 121

100 ± 64

183 ± 243


Vitamin E (α-tocopherol equivalent; mg)

43 ± 90

52 ± 156

13 ± 8.2

16 ± 6.7

13 ± 7.1

19 ± 14


Pantothenic acid (mg)

10.3 ± 22.0

7.6 ± 10.3

6.3 ± 3.3

6.9 ± 3.5

5.6 ± 3.2

8.5 ± 6.0


Methionine (g)

2.15 ± 0.75

2.21 ± 0.80

2.08 ± 0.89

1.98 ± 0.57

2.00 ± 0.77

1.75 ± 0.62


Cystine (g)

1.31 ± 0.49

1.24 ± 0.46

1.24 ± 0.46

1.20 ± 0.30

1.19 ± 0.37

1.08 ± 0.34


Iron (mg)

21.7 ± 11.3

17.6 ± 7.9

18.6 ± 8.0

20.0 ± 5.2

19.0 ± 7.9

23.0 ± 11.3


Zinc (mg)

14.8 ± 5.6

13.2 ± 6.0

11.7 ± 4.3

14.2 ± 5.6

12.1 ± 5.5

16.0 ± 9.6


Copper (mg)

1.70 ± 0.74

1.60 ± 0.75

1.74 ± 0.76

1.84 ± 0.69

1.63 ± 0.76

2.13 ± 1.11


Se (μg; 1 μg Se = 12.7 nmol)

135 ± 47

143 ± 71

130 ± 41

144 ± 63

127 ± 43

122 ± 42


aRepeated measures analysis of variance: Se main effect or Se × time interaction

bValues at the end of the run-in period before starting supplementation

cValues at the end of the 48 weeks supplementation period

dValues 48 weeks after the end of supplementation

Lipid peroxidation

Urinary excretion of 15-isoprostane F2t was not different between groups and did not change significantly in either group during treatment (Table 2). Additionally, urinary excretion of 15-isoprostane F2t was not correlated with dietary intake of Se (diet plus supplements), polyunsaturated fat, arachidonic acid, vitamin E, vitamin C, vitamin A, ß-carotene, iron, zinc, or copper. Likewise, urinary excretion of 15-isoprostane F2t was not correlated with urinary excretion of Se, creatinine, or nitrogen (data not shown), nor was the change in urinary excretion of 15-isoprostane F2t excretion correlated with the change in plasma Se.


Blood plasma Se concentration in our subjects was higher than the national average Se concentration in serum even though their estimated Se intakes did not differ from the US average. The concentrations of Se in serum and plasma are similar [24], and a direct comparison of serum and plasma prepared from the same blood samples found no difference in Se concentration [25]. The higher plasma Se in our subjects may have been due to the relatively high Se content of foods [26] and water [27] in the Sacramento area, which was not taken into account in the nutrient database used to estimate Se intakes. Furthermore, multi-day food records tend to under-report food intake compared to 24 h recall surveys [28], which were used in NHANES. In addition, the dietary energy intakes in Table 1 are 16% lower than the subjects’ estimated energy requirement of 13.1 MJ/day [29], another indication that subjects under-reported their intakes. The relatively high initial Se status of our subjects is consistent with the fact that 24% (10 of 42) of the subjects had pre-study plasma Se concentrations exceeding the 95th percentile observed in NHANES. Therefore, the actual Se intake of our subjects was higher than we report, and thus, higher than the national average.

The total Se intake of the subjects taking the high-Se yeast was approximately 435 μg/day (5.5 nmol/day), which exceeded the UL. However, no sign of Se toxicity (garlic odor of breath, abnormal growth or loss of hair or nails) was observed or reported by the subjects. The highest individual blood Se observed in this study was 673 μg/l (8.5 μM; estimated from plasma and erythrocyte Se concentrations assuming a hematocrit of 50%), considerably below the toxic threshold of 1,000 μg/l (12.7 μM) that was considered by the Dietary Reference Intakes Committee [1]. During treatment, the highest hair Se concentration in the supplemented subjects was 1.87 μg/g, far below the average hair Se of 21 μg/g observed in the selenosis-endemic area of China [30]. Consistent with the lack of signs of Se toxicity, none of the clinical blood chemistries (sodium, potassium, chloride, carbon dioxide, urea nitrogen, creatinine, glucose, calcium, protein, albumin, alkaline phosphatase, aspartate aminotransferase, bilirubin, total cholesterol, LDL cholesterol, HDL cholesterol, or triglycerides) were outside the normal ranges (data not shown). Only two trials have studied higher doses of selenized yeast in humans. In the Watchful Waiting Study [31], doses of 800 μg/day (10 μmol/day) have been administered for up to 3 years to men diagnosed with prostate cancer with no toxic side effects, which tends to confirm the validity of the twofold safety factor included in the UL. Likewise, an 8-month trial of 600 μg/day (7.6 μmol/day) of high-Se yeast in male and female arthritis sufferers that achieved an average blood Se concentration above 600 μg/l (7.6 μM) reported no adverse effects [32].

The average urinary Se excretion was 45.5% of the estimated dietary intake when subjects were getting all of their Se from the diet and dropped to 37.4% when taking high-Se yeast supplements. These percentages are higher than we observed in our previous study of high-Se foods in men confined to a metabolic research unit [33], presumably because Se intake was under-estimated, as noted above. Nevertheless, in both studies, the fraction of Se excreted in urine was larger when Se intake was lower and decreased when Se intake increased. Considering only the increase in urinary Se excretion during supplementation, 31% of the Se from the supplements was excreted in urine, the same fraction as was excreted in urine after 90 days of acclimation to high-Se foods in our previous study [33]. Regression of urinary Se excretion on dietary Se intake in the placebo group, the supplemented group, or all subjects combined gave estimates of the intercept from 35 to 40 μg/day, suggesting this may represent a basal urinary excretion rate. This is similar to an estimate that 33 μg/day is lost in the urine of healthy North American men, based on a balance study in healthy men [34].

Absorption of 77Se from intrinsically labeled SelenoPrecise™ yeast was 90% based on measurements in 12 men [35]. This yeast contains approximately 81% of its Se as selenomethionine [13]. Selenomethionine is metabolized by the same pathways as methionine, and the amount in excess of nutritional Se requirements is incorporated into proteins in place of methionine [36]. Measurement of Se in the bile fluid of accident victims indicates that enterohepatic reabsorption of Se from the bile is the same order of magnitude as dietary intakes [24], which may account for the relatively small fraction excreted in urine. Much of the retained Se is incorporated into skeletal muscle, where it accounts for about half the whole-body Se [24]. The time course of Se accumulation in plasma from high-Se yeast was similar to Se from high-Se foods in our previous study [33]. Plasma Se took approximately 35 d to reach half its maximum concentration when high-Se yeast supplements were administered in the present study, whereas 32 days were required to reach half-maximal plasma Se concentration when men were fed high-Se rice and beef. The similarities in plasma kinetics and urinary excretion suggest that the absorption, distribution, and elimination of Se from high-Se yeast are similar to Se from conventional foods.

Our failure to observe any change in urinary 15-isoprostane F2t excretion is consistent with our observation that glutathione peroxidase activities in blood did not change and with the absence of reports that Se supplementation decreases lipid peroxidation in healthy, well-nourished individuals. Some animal studies have found an effect of dietary Se on lipid peroxidation, but only in Se deficiency [37, 38]. A study in home parenteral nutrition patients with low Se status found no relation between breath pentane, an indicator of in-vivo lipid peroxidation, and plasma Se levels [39]. It appears that any health benefits that may be associated with Se supplementation are unlikely to be due to antioxidant protection.

When considered in terms of both speed of response and ability to discriminate between Se intakes, plasma Se displayed the best combination of desirable characteristics for Se status assessment. Plasma Se has the further advantage that it closely tracks with muscle Se [33], which represents about half the whole-body Se and therefore reflects both usual Se intake and whole-body Se stores. Erythrocyte Se was slightly better at discriminating between intakes but took about twice as long to reach a steady-state after a change in intake. Thus, the expected advantage of measuring Se in a cellular compartment such as erythrocytes is outweighed by plasma Se’s faster response and its similar kinetics to muscle Se. Platelet Se increased faster than plasma Se, but its greater variability and more complex sample preparation make it less attractive for status assessment. Platelets had the highest glutathione peroxidase specific activity, as previously reported [40], but the enzyme activity was not useful for status assessment at these high intakes. The greater variability of platelet glutathione peroxidase activity compared to tissue Se has been noted previously [41].

In contrast to our previous study in men confined to a metabolic research unit, where urinary Se excretion was the most discriminating indicator of Se status [33], we found that urine was far less discriminating than blood Se in free-living men. This was likely due to the more stringent controls in a confined study and the difficulty of obtaining reliable 24-h urine samples from free-living subjects [42]. In the present study, urinary Se excretion was the only measure of Se status that was significantly correlated with dietary Se intake. Presumably, this was because (a) the urine collections were made shortly after dietary Se intake was assessed, (b) the collections were relatively complete (creatinine excretion was 1.5 ± 0.9 g/day in the placebo group and 1.5 ± 1.3 g/day in the high-Se group) and (c) urinary Se excretion is closer metabolically to Se absorption from food than Se incorporation into tissues. A study of Se status indicators in the high-Se area of South Dakota and Wyoming similarly found that 24 h urinary Se excretion was more tightly correlated with dietary intake than was blood Se concentration [43]. In contrast to our results using 3-day food records, Longnecker et. al. observed highly significant correlations between dietary Se intake and serum Se, undoubtedly due to their use of duplicate food plates that were directly analyzed for Se.

In this study, hair Se levels were not a reliable indicator of Se intake because of contamination of hair samples by use of Se-containing anti-dandruff shampoos. However, when the contaminated samples were excluded, hair Se was among the most discriminating measures of Se intake. Hair Se responded slowly to changes in Se intake and thus may be useful for assessment of Se intakes integrated over a long time, particularly in rural populations and undeveloped areas where Se-containing shampoo usage is not prevalent and preservation of blood samples is problematic. Indeed, hair Se has been applied successfully in studies in China [44, 45] and the Czech Republic [46].



US Department of Agriculture CRIS Project no. 5306-51530-009-00D supported this research. The authors gratefully acknowledge the excellent technical assistance of Manuel Tengonciang, Jerome Crawford, Sue Littlefield, Leslie Woodhouse, Katherine Parker, Evelyn Holguin and the Human Studies Unit of WHNRC for their assistance with the conduct of this study. Mention of trade name, proprietary product, or specific equipment does not constitute a guarantee or warranty by the US Department of Agriculture, nor does it imply approval to the exclusion of other products that may be suitable. The opinions expressed herein represent those of the authors and do not necessarily represent those of the US Department of Agriculture.


  1. 1.
    Panel on Dietary Antioxidants and Related Compounds, Subcommittee on Upper Reference Levels of Nutrients, Subcommittee on Interpretation and Uses of DRIs, Standing Committee on the Scientific Evaluation of Dietary Reference Intakes, Food and Nutrition Board (2000) Dietary Reference Intakes for Vitamin C, Vitamin E, Selenium, and Carotenoids. National Academy Press, Washington, DC, USAGoogle Scholar
  2. 2.
    Kryukov GV, Castellano S, Novoselov SV, Lobanov AV, Zehtab O, Guigo R, Gladyshev VN (2003) Characterization of mammalian selenoproteomes. Science 300:1439–1443PubMedCrossRefGoogle Scholar
  3. 3.
    Swanson CA, Longnecker MP, Veillon C, Howe SM, Levander OA, Taylor PR, Mcadam PA, Brown CC, Stampfer MJ, Willett WC (1990) Selenium intake age gender and smoking in relation to indices of selenium status of adults residing in a seleniferous area. Am J Clin Nutr 52:858–862PubMedGoogle Scholar
  4. 4.
    Longnecker MP, Taylor PR, Levander OA, Howe SM, Veillon C, Mcadam PA, Patterson KY, Holden JM, Stampfer MJ et al (1991) Selenium in diet, blood and toenails in relation to human health in a seleniferous area. Am J Clin Nutr 53:1288–1294PubMedGoogle Scholar
  5. 5.
    Combs GF Jr, Combs SB (1986) The role of selenium in nutrition. Academic Press, San Diego, CA, USAGoogle Scholar
  6. 6.
    Clark LC, Combs GF Jr, Turnbull BW, Slate EH, Chalker DK, Chow J, Davis LS, Glover RA, Graham GF, Gross EG, Krongrad A, Lesher JL Jr, Park HK, Sanders BB Jr, Smith CL, Taylor JR (1996) Effects of selenium supplementation for cancer prevention in patients with carcinoma of the skin. A randomized controlled trial. J Am Med Assoc 276:1957–1963CrossRefGoogle Scholar
  7. 7.
    Blot WJ, Li JY, Taylor PR, Guo WD, Dawsey S, Wang GQ, Yang CS, Zheng SF, Gail M, Li GY, Yu Y, Liu BQ, Tangrea J, Sun YH, Liu FS, Fraumeni JF, Zhang YH, Li B (1993) Nutrition intervention trials in Linxian, China—supplementation with specific vitamin/mineral combinations, cancer incidence, and disease-specific mortality in the general population. J Natl Cancer Inst 85:1483–1492PubMedCrossRefGoogle Scholar
  8. 8.
    Blot WJ, Li JY, Taylor PR, Guo WD, Dawsey SM, Li B (1995) The Linxian trials: mortality rates by vitamin–mineral intervention group. Am J Clin Nutr 62:S1424–S1426Google Scholar
  9. 9.
    Yu SY, Zhu YJ, Li WG (1997) Protective role of selenium against hepatitis B virus and primary liver cancer in Qidong. Biol Trace Elem Res 56:117–124PubMedCrossRefGoogle Scholar
  10. 10.
    Geiger D (2004) Selenium Supplements. Newsday, May 15Google Scholar
  11. 11.
    Friel H, Lederman H (2006) A nutritional supplement formula for influenza A (H5N1) infection in humans. Med Hypotheses 16:578–587CrossRefGoogle Scholar
  12. 12.
    Daniells S (2006) The supplements—bird flu debate: selenium., February 26Google Scholar
  13. 13.
    Larsen EH, Hansen M, Paulin H, Moesgaard S, Reid M, Rayman M (2004) Speciation and bioavailability of selenium in yeast-based intervention agents used in cancer chemoprevention studies. J AOAC Int 87:225–232PubMedGoogle Scholar
  14. 14.
    U.S. Department of Agriculture (2005) USDA National Nutrient Database for Standard Reference, Release 18. U.S. Department of Agriculture, Agricultural Research Service, Nutrient Data LaboratoryGoogle Scholar
  15. 15.
    Hawkes WC, Kutnink MA (1996) High-performance liquid chromatographic-fluorescence determination of traces of selenium in biological materials. Anal Biochem 241:206–211PubMedCrossRefGoogle Scholar
  16. 16.
    Paglia DE, Valentine WN (1967) Studies on the quantitative and qualitative characterization of erythrocyte glutathione peroxidase. J Lab Clin Med 70:158–169PubMedGoogle Scholar
  17. 17.
    Drabkin D (1949) The standardization of hemoglobin measurement. Am J Med Sci 217:710–711Google Scholar
  18. 18.
    Lowry OH, Rosebrough NJ, Farr AL, Randall RJ (1951) Protein measurement with the Folin phenol reagent. J Biol Chem 193:265–275PubMedGoogle Scholar
  19. 19.
    Box GEP, Hunter WG, Hunter JS (1978) Statistics for experimenters: An introduction to design, data analysis, and model building. Wiley, New York, pp 232–235Google Scholar
  20. 20.
    Littell RC, Milliken GA, Stroup WW, Wolfinger RD (1996) SAS System for Mixed Models. SAS Institute, Cary, NC, USAGoogle Scholar
  21. 21.
    Bonferroni CE (1937) Teoria statistica delle classi e calcolo delle probabilita. In: Volume in Onore di Ricarrdo dalla Volta. Florence, Italy: Universita di Firenza, pp. 1–62Google Scholar
  22. 22.
    Kafai MR, Ganji V (2003) Sex, age, geographical location, smoking, and alcohol consumption influence serum selenium concentrations in the USA: Third National Health and Nutrition Examination Survey, 1988–1994. J Trace Elem Med Biol 17:13–18PubMedCrossRefGoogle Scholar
  23. 23.
    Ervin RB, Wang CY, Wright JD, Kennedy-Stephenson J (2004) Dietary intake of selected minerals for the United States population: 1999–2000. Adv Data 27:1–5Google Scholar
  24. 24.
    Oster O, Schmiedel G, Prellwitz W (1988) The organ distribution of selenium in German adults. Biol Trace Elem Res 15:23–45PubMedCrossRefGoogle Scholar
  25. 25.
    Rose A, Holbrook J, Lewis S, Craft N, Seidel K, Smith JC Jr (1986) Serum vs. plasma the effect of two anticoagulants on the concentration of selected nutritional and biochemical indices. 70th Annual Meeting Of The Federation Of American Societies For Experimental Biology, St. Louis, Mo., USA, Apr, 45:840Google Scholar
  26. 26.
    Harris T (1988) Selenium: The Poisoning of America. Sacramento Bee. Sacramento, CA,pp. A1–A28Google Scholar
  27. 27.
    Fan AM, Book SA, Neutra RR, Epstein DM (1988) Selenium and human health implications in California’s San Joaquin valley USA. J Toxicol Environ Health 23:539–560PubMedGoogle Scholar
  28. 28.
    Buzzard IM, Faucett CL, Jeffery RW, McBane L, McGovern P, Baxter JS, Shapiro AC, Blackburn GL, Chlebowski RT, Elashoff RM, Wynder EL (1996) Monitoring dietary change in a low-fat diet intervention study: advantages of using 24-hour dietary recalls vs food records. J Am Diet Assoc 96:574–579PubMedCrossRefGoogle Scholar
  29. 29.
    Harris JA, Benedict FG (1919) Biometric studies of basal metabolism in man. Publication #297. Carnegie Institute of Washington, Washington, DC, USAGoogle Scholar
  30. 30.
    Liu BS, Li SS (1987) Endemic selenosis and fluorosis. In: Combs GF Jr, Levander OA, Spallholz JE, Oldfield JE (eds) Selenium in biology and medicine, vol. B. Van Norstrand Reinhold, New York, pp 708–711Google Scholar
  31. 31.
    Stratton MS, Reid ME, Schwartzberg G, Minter FE, Monroe BK, Alberts DS, Marshall JR, Ahmann FR (2003) Selenium and inhibition of disease progression in men diagnosed with prostate carcinoma: study design and baseline characteristics of the ‘Watchful Waiting’ Study. Anticancer Drugs 14:595–600PubMedCrossRefGoogle Scholar
  32. 32.
    Aaseth J, Haugen M, Forre O (1998) Rheumatoid arthritis and metal compounds—perspectives on the role of oxygen radical detoxification. Analyst 123:3–6PubMedCrossRefGoogle Scholar
  33. 33.
    Hawkes WC, Alkan FZ, Oehler L (2003) Absorption, distribution and excretion of selenium from beef and rice in healthy North American men. J Nutr 133:3434–3442PubMedGoogle Scholar
  34. 34.
    Levander OA, Sutherland B, Morris VC, King JC (1981) Selenium balance in young men during selenium depletion and repletion. Am J Clin Nutr 34:2662–2669PubMedGoogle Scholar
  35. 35.
    Sloth JJ, Larsen EH, Bugel SH, Moesgaard S (2003) Determination of total selenium and 77Se in isotopically enriched human samples by ICP-dynamic reaction cell-MS. J Anal At Spectrom 18:317–322CrossRefGoogle Scholar
  36. 36.
    Rayman MP (2004) The use of high-selenium yeast to raise selenium status: how does it measure up? Br J Nutr 92:557–573PubMedCrossRefGoogle Scholar
  37. 37.
    Kivits GA, Ganguli-Swarttouw MA, Christ EJ (1982) Influence of a low-selenium diet on erythrocyte glutathione peroxidase and alkane production in exhaled air of rats as a measure of lipid peroxidation in vivo during growth. Biochim Biophys Acta 719:329–333PubMedGoogle Scholar
  38. 38.
    Glass M, Sutherland MW, Forman HJ, Fisher AB (1985) Selenium deficiency potentiates paraquat-induced lipid peroxidation in isolated perfused rat lung. J Appl Physiol 59:619–622PubMedGoogle Scholar
  39. 39.
    Lemoyne M, Van Gossum A, Kurian R, Jeejeebhoy KN (1988) Plasma vitamin E and selenium and breath pentane in home parenteral nutrition patients. Am J Clin Nutr 48:1310–1315PubMedGoogle Scholar
  40. 40.
    Matsuda A, Kimura M, Itokawa Y (1997) Selenium level and glutathione peroxidase activity in plasma, erythrocytes and platelets of healthy Japanese volunteers. J Nutr Sci Vitaminol (Tokyo) 43:497–504Google Scholar
  41. 41.
    Brown KM, Pickard K, Nicol F, Beckett GJ, Duthie GG, Arthur JR (2000) Effects of organic and inorganic selenium supplementation on selenoenzyme activity in blood lymphoctyes, granulocytes, platelets and erythrocytes. Clin Sci (Lond) 98:593–599CrossRefGoogle Scholar
  42. 42.
    Wood GP (1977) Urinary estrogens during late pregnancy: observations on the validity of “24-hour” urine collections and the estrogen/creatinine ratio. South Med J 70:959–962PubMedGoogle Scholar
  43. 43.
    Longnecker MP, Taylor PR, Levander OA, Flack V, Veillon C, Mcadam PA, Patterson KY, Holden J, Stampfer MJ, Et AL (1991) Estimation of selenium intake from selenium in serum whole blood toenails or urine. 75th Annual Meeting Of The Federation Of American Societies For Experimental Biology, Atlanta, Georgia, USA, April, 5Google Scholar
  44. 44.
    Wang M (1990) Geographical differentiation of selenium concentration in hair from children and youngsters in China. J Environ Sci 2:99–106Google Scholar
  45. 45.
    Yang G, Gu L, Zhou R, Yin S (1988) Studies of human maximal and minimal safe intake and requirement of selenium. In: Wendel A (ed) Selenium in Biology and Medicine. Springer, Berlin, pp 223–228Google Scholar
  46. 46.
    Kvicala J, Zamrazil V, Cerovska J, Bednar J, Janda J (1995) Evaluation of selenium supply and status of inhabitants in three selected rural and urban regions of the Czech Republic. Biol Trace Elem Res 47:365–375PubMedCrossRefGoogle Scholar

Copyright information

© U.S. Government 2007

Authors and Affiliations

  • Wayne Chris Hawkes
    • 1
  • B. Diane Richter
    • 1
  • Zeynep Alkan
    • 1
  • Elaine C. Souza
    • 1
  • Monique Derricote
    • 1
  • Bruce E. Mackey
    • 2
  • Ellen L. Bonnel
    • 1
  1. 1.Western Human Nutrition Research Center, Agricultural Research Service, United States Department of AgricultureUniversity of California at DavisDavisUSA
  2. 2.Western Regional Research CenterARS, USDAAlbanyUSA

Personalised recommendations