Dose-response relations in urinary excretion of trimethylselenonium in the rat

75Se-labeled selenite was administered to fasting rats by orogastric intubation (1.5-3000 micrograms/kg body wt). Urine was collected and characterized for total radioactivity as well as for radiolabeled trimethylselenonium (TMSe). At lower doses of selenite (up to 500 micrograms/kg body wt), 30% of the administered dose was excreted. At higher doses of selenite, fractional urine excretion decreased as a function of the dose. The observed decrease in fractional urine excretion was not caused by changes in the absorption of the administered radiolabel. There was a direct relationship between the amount of the administered dose of selenite (up to 1500 micrograms/kg body wt) and the proportion of urinary [75Se] excreted as TMSe. Pretreatment with seleno compounds (10 or 100 micrograms Se/kg body wt as selenite, or selenomethionine) for 35 d before a challenge dose of [75Se]selenite did not influence the excretion of total [75Se] or of [75Se]TMSe in urine. Ingestion of a choline-deficient diet, which should deplete the availability of methyl groups, did not have any effect on excretion of total [75Se] or of [75Se]TMSe in urine after a challenge dose of [75Se]selenite (500 micrograms/kg body wt). The data presented here permit the following conclusions: 1) Production of TMSe is dose dependent, 2) production of TMSe from a single acute dose does not depend on the history of selenium intake and 3) rats fed a methyl-deficient diet are able to eliminate Se via formation of TMSe.     

Metabolism of selenite in men with widely varying selenium status.

OBJECTIVE: This study was undertaken to investigate the metabolism of selenite in men with life-long intakes of deficient, adequate and excess selenium. METHODS: Stable isotopes of selenium were infused for five hours into Chinese men living in deficient, adequate or excessive selenium areas, and 24-hour urine and blood samples were collected daily for the next seven days. Stable isotopic selenium excretion was determined in urine and in whole plasma and plasma fractions. RESULTS: Even though there was a positive correlation of selenium intake with the urinary excretion of this element, this relationship was not linear over the entire range (deficient, adequate, excessive) of selenium intake. When the urine excretion was normalized internally within each group, a sharp increase in the slope of this relationship was found when long-term intake increased to adequate amounts, but the slope reached a plateau when the daily intake exceeded the adequate group. The plasma selenoprotein P fraction was labeled initially, but the incorporation in the glutathione peroxidase fraction subsequently increased by a small amount. A two-month dietary restriction of selenium of the subjects from the excess area did not result in a reduction of urinary excretion of infused selenite. CONCLUSION: A complex relationship exists between long-term intake of selenium and selenium status, and subjects living in the excess area are more saturated with selenium than anticipated. More than two months of depletion are required to affect urinary excretion of selenium.     

Effect of chronic selenite supplementation on selenium excretion and organ accumulation in rats

We examined the effect of chronic selenite supplementation on whole body and selected organ selenium (Se) accumulation, urine excretion of total Se and trimethylselenonium ion, and Se balance in adult male rats. Animals were housed in metabolic cages and given either deionized water or water containing 4 micrograms of Se/mL as selenite for 30 d. Absorption of selenite was nearly complete, with only approximately 10% of ingested Se appearing in feces. There was a rapid rise in urinary Se that reached a plateau within a few days and accounted for 54 +/- 2% of the intake. Excretion of trimethylselenonium ion (TMSe) in urine increased rapidly, representing 35-40% of urinary Se in the supplemented animals compared with only 2% for the control group. In one experiment, rats were killed at 30 d and total carcass Se was measured using isotope dilution analysis. Supplemented rats had only a modest increase in whole body Se (94 +/- 4 micrograms Se vs. 66 +/- 3 in controls). Calculation of Se balance in the supplemented rats showed that approximately 35% of ingested Se could not be accounted for by urine plus fecal losses combined with the portion retained in the carcass. The results from this study demonstrate that under the condition of supplementation at 4 micrograms of Se/mL of drinking water, pathways other than urinary and fecal excretion may account for a substantial portion of Se loss.     

Kinetics of a single administration of 74Se-selenite by oral and intravenous routes in adult humans

The purpose of this study was to explore the fate of a single dose of labeled selenium as determined by its route of administration. Thus, the appearance of a stable isotope of selenium, administered as 74-Se-selenite, was measured in plasma, urine, and feces, with neutron activation analysis, following a 81.7 micrograms dose of 74Se-selenite given either intravenously or orally in two groups (n = 4) of healthy, young adult men, who were otherwise maintained on a diet providing a constant and adequate selenium intake. From these isotopic data, measurable parameters of urine excretion, total body retention and selenite-exchangeable metabolic pool (Se-EMP) were defined to provide a quantitative assessment of selenium metabolism in these subjects. The initial 24-hr urine excretion of the label was higher for the intravenously administered label (18.2 +/- 2.1% of dose) compared to the oral dose (11.7 +/- 2.6% absorbed dose). Thereafter, the excretion of isotope was the same for both groups. For equivalent entry of Se into the body, measured total body retention and Se-EMP were the same for both groups. These initial kinetic data suggest that the overall utilization of selenium from a single administration of selenite is comparable for the two routes of intake and that the host's selenium requirement can probably be met adequately via the intravenous administration of selenite.     

The retention and distribution by healthy young men of stable isotopes of selenium consumed as selenite, selenate or hydroponically-grown broccoli are dependent on the isotopic form

Twenty-seven healthy young men were randomly assigned to diets that supplied low (32.6 microg/d) or high (226.5 microg/d) levels of selenium for a 105-d study. After consuming the diets for 85 d, subjects were fed a test meal that contained 74Se in the form of selenite or selenate and 82Se incorporated into hydroponically-raised broccoli. Urine, fecal and blood samples were collected daily. Isotope absorption was not different (P > 0.05) for selenate and Se in broccoli; Se absorption from selenite was highly variable and was not included in statistical analyses. Significantly more isotope was absorbed by subjects fed the high Se diet (P = 0. 015). Urinary isotope excretion was greater when selenate was fed than when broccoli was fed (P = 0.0001), and consequently more Se from broccoli (as compared to selenate) was retained (59.2 +/- 2.4 and 36.4 +/- 4.6% for Se in broccoli and selenate, respectively; P = 0.0001). Despite the higher retention, less isotope from broccoli than from selenate was present in the plasma. Plasma proteins separated by gel permeation chromatography showed that most of the isotopes were distributed between two medium molecular weight peaks. Less isotope was found in plasma proteins of subjects fed the high Se diet, but the form of Se had no effect on isotope distribution. These results show that dietary Se intake alters the retention of stable isotopes of Se and that humans retain and distribute Se from broccoli in a different manner than Se from inorganic salts.     

Urinary and fecal excretions and absorption of a large supplement of selenium: superiority of selenate over selenite

A correction needs to be made to the form of selenium used in earlier studies; what was believed to be selenite-Se in solution is now known to have been selenate-Se. In the present study, excretion of Se was followed in 13 women after ingestion of 1 mg Se as selenite or selenate in solution. Fecal excretion of selenate-Se was less than for selenite-Se reflecting a higher apparent absorption [94 +/- 4% (SD), 62 +/- 14%, respectively]. Peak excretion of Se occurred 3 h earlier for selenate-Se than for selenite-Se and was 6 times higher. Total urinary excretion of selenate-Se was 3 times that of selenite-Se and still 2 times as high when expressed as % absorbed dose. Total recovery of Se in urine and feces was similar for both forms. There was remarkable agreement between these results and those reported earlier for selenate-Se (Selovet-1) and selenite-Se.     

Absorption and retention of 75SeO3 in relation to soybean protein intake in the rat

The absorption and retention of ⁷⁵Se, given as an oral dose of ⁷⁵SeO₃, was studied in young male rats receiving different levels and sources of soy protein, with and without selenite and methionine supplementation. Rats fed a protein-free diet had a higher cumulative urine ⁷⁵Se excretion and a sligtly lower ⁷⁵Se absorption and ⁷⁵Se retention than rats fed diets containing 10% protein supplied as soya flour. Results indicated that supplementation with selenite decreased the fractional absorption and retention of selenium, but the overall effect was a marked increase in the total amount of selenium ingested, absorbed and retained. Methionine supplementation of a diet based on soya increased growth and PER: it also decreased slightly cumulative feces ⁷⁵Se excretion and increased ⁷⁵Se absorption, but only in rats fed diets supplemented with selenium. The present findings are consistent with the view that selenium homeostasis in the rat is maintained largely through changes in the urinary excretion of selenium and they show that an inadequate protein diet reduces the efficiency of retention of absorbed selenite.     

Effects of chemical form and dosage on the incorporation of selenium into tissue proteins in rats

We investigated the incorporation of Se into the proteins of liver and muscle, the two main Se pools, during replenishment of Se-deficient rats with normal or large doses of 75Se-labeled selenite and selenomethionine, doses equivalent to the amounts ingested from a diet with 0.2 or 2 mg Se/kg. With the higher intake, Se levels were elevated. More Se was retained from selenomethionine than from selenite. After separation of the labeled proteins, it was apparent that the higher tissue Se contents were mainly due to nonspecific incorporation into a large number of proteins. We observed no differences between the two chemical forms with regard to the formation of the specific selenoproteins. The 10-fold increase in the Se supply led to a relatively small rise in the levels of these compounds. The results indicate that after ingestion of normal amounts of selenite nearly all of the element is present in the specific selenoproteins. With increasing doses a part is also incorporated nonspecifically into numerous other proteins. In the case of selenomethionine, a part of the element follows the same metabolic pathways, but a percentage is also deposited directly and nonspecifically into proteins in place of methionine.     

Ascorbic acid-selenite interactions in humans studied with an oral dose of 74SeO3(2-)

The interaction between dietary ascorbic acid at extremes of ascorbic acid intake and selenium in young adult male humans was investigated with a stable-isotope approach using 74Se-selenite. Measurements were made of 74Se in plasma, urine, and feces with neutron-activation analysis after oral administration of 74SeO3(2-). Urine excretion and total body retention of isotope and the selenite-exchangeable metabolic pool (Se-EMP) were calculated. Limiting dietary ascorbic acid to about 20 mg/d appeared to reduce the time-related retention of absorbed selenite and the size of Se-EMP. Compared with a diet providing 1 g ascorbic acid/d the low ascorbic acid intake was associated with a lower fractional absorption of the isotope, a reduced retention of the label, and a smaller Se-EMP. These data and those previously obtained in subjects with more usual ascorbic acid intakes point to a possible important role for ascorbic acid in the maintenance of Se homeostasis.     

On supplementing the selenium intake of New Zealanders. 2. Prolonged metabolic experiments with daily supplements of selenomethionine, selenite and fish

1. The daily intake of selenium by three subjects was supplemented with 100 microgram Se as selenomethionine (Semet-Se) or sodium selenite (selenite-Se)/d for 10-11 weeks, or with 65 microgram Se as in mackerel (Scomber japonicus) (fish-Se)/d for 4 weeks. 2. Urinary and faecal excretion of Se was measured and also Se concentration in whole blood, plasma and erythrocytes. Measurements on blood were made at intervals after supplementation had ceased. 3. Selenite-Se was not as well absorbed (0.46 of the intake) during the first 4 weeks as Semet-Se (0.75 of the intake) and fish Se (0.66 of the intake). 4. Blood Se increased steadily with Semet-Se, from 0.08 to 0.18 microgram Se/ml, but more slowly with selenite-Se, reaching a plateau in 7-8 weeks at 0.11 microgram Se/ml. Plasma Se increased more rapidly with Semet-Se than with selenite-Se, so that initially with Semet-Se plasma Se was greater than erythrocyte Se. 5. Daily urinary excretion increased with all forms of supplement, with initially a greater proportion of absorbed selenite-Se being excreted than Semet-Se or fish-Se. A close relationship was found between plasma Se and 24 h urinary excretion. The findings suggested that there was a rapid initial excretion of presumably unbound Se then a slower excretion of residual unbound, loosely bound or bound Se. 6. Total retentions of 3.5 mg selenite-Se and 4.5 mg Semet-Se were large when compared with an estimate of body content of 6 mg Se, derived in another paper (Stewart, Griffiths, Thomson & Robinson, 1978). Retention of Semet-Se and fish-Se appeared to be reflected in blood Se, whereas for selenite-Se, blood Se reflected retention for only a short period after which Se appeared to be retained without altering the blood Se. This suggested that Semet-Se and selenite-Se were metabolized differently. 7. A double blind-dosing trail with 100 microgram Semet-Se was carried out for 12 weeks on twenty-four patients with muscular complaints in Tapanui, a low-Se-soil area. Blood Se increased in the experimental group (from 0.067 to 0.143 microgrm Se/ml); clinical findings were not conclusive and will be presented elsewhere. 8. Bood Se was measured in New Zealand residents before travelling to Europe or to North America. On return their blood Se was increased, and depending upon the period of time spent outside New Zealand some values reached concentrations found in visitors and new settlers to New Zealand. 9. The results from these studies and the earlier studies of single and multiple dosing have been used to look at the various criteria in use for assessing Se status of subjects. It is suggested that plasma Se be used in preference to 24 h urinary excretion, and in addition to whole blood Se and glutathione peroxidase (EC 1.11.1.9) activity.     

On supplementing the selenium intake of New Zealanders. 1. Short experiments with large doses of selenite or selenomethionine

1. Urinary and faecal excretion of single oral doses of 1 mg selenium or 0.1 mg Se as selenomethionine (Semet-Se) in solution were studied in two women. Most of the Se was absorbed and little was eliminated in the urine (0.05-0.22 dose). 2. The results have been compared with those from an earlier study (Thomson, 1974) on the same two women after similarly sized doses of sodium selenite (selenite-Se) in solution. Although selenite-Se was almost as well absorbed as Semet-Se more was excreted in the urine (0.41-0.85 dose). 3. Repeated dosing with 1 mg selenite-Se on five consecutive days in one of the women indicated that 1.1 mg had been retained. 4. Twenty patients with muscular complaints from Tapanui (South Otago, New Zealand), a low-Se soil area, ingested 0.5 mg selenite-Se daily for 20 d. Blood Se increased rapidly to almost twice the initial concentration but reached a plateau well below most values reported for residents outside New Zealand. No difference in blood Se concentration was found between those who did or did not report improvement. 5. Spasmodic medication with selenite-Se by some residents near Lincoln (Christchurch, New Zealand) for periods of up to 10 years or more had increased the blood Se somewhat.     

Increase in Dopamine Metabolites in Murine Striatum After Oral Exposure to Inorganic But Not Organic Form of Selenium

Selenium (Se) is an essential as well as a toxic trace element. Se intoxication has been reported in both livestock and humans. The central nervous system is sensitive to Se poisoning; exposure to Se causes blind staggers in cattle, poliomyelomalacia in pigs, and nervous system disorders in humans. Differences in neurotoxicity between inorganic and organic Se have been demonstrated. In this study, groups of five male BALB/c mice each were administered sodium selenite or selenomethionine in drinking water ad libitum at 0, 1, 3, and 9 ppm as Se for 14 days. At the end of Se exposure, their brains were removed and dissected into different regions. The concentration of norepinephrine (NE), dopamine (DA), dihydroxyphenylacetic acid (DOPAC), homovanillic acid (HVA), serotonin (5-HT), and 5-hydroxyindolacetic acid (5-HIAA) were determined in each brain region. Food and water consumption and body weight gain were significantly decreased in the group treated with the highest concentration of sodium selenite. In mice administered sodium selenite at 3 and 9 ppm, DOPAC was significantly higher in the striatum than in the control group. The striatal HVA was also increased in the group treated with 3 ppm Se; the DA showed a similar pattern, but the increase was not statistically significant. No alterations of NE, 5-HT, or 5-HIAA levels were detected in any brain region of mice treated with sodium selenite. No significant differences in any parameter among the groups treated with selenomethionine were observed indicating that inorganic Se was more neurotoxic than organic Se via drinking water. The alterations of DA metabolites by inorganic Se in DA-rich striatum suggested a Se-specific increased neural activity of dopaminergic pathways. Results may be useful in further elucidation of neurotoxicity of Se and in establishing a safe level of intake for this element. Received: 10 September 1999/Accepted: 30 November 1999     

Kinetic modeling of selenium metabolism in nonpregnant ewes

The kinetics of selenium metabolism in three nonpregnant ewes were studied by the intravenous injection of 75Se-sodium selenite and measurement of radioactivity responses in blood, tissues and excreta. Stable selenium measurements were also made to determine selenium intake, excretion in feces and urine, and mass of selenium in tissues. Immediately following tracer injection, there was a rapid disappearance of radioactivity from plasma reflecting the uptake of the element by the liver and blood cells. The decrease in plasma radioactivity ceased abruptly by 30-45 min, and was followed by an increase to a peak by 3-4 h and a more gradual biphasic decline thereafter. A kinetic model of selenium metabolism in the whole animal was constructed employing the SAAM/CONSAM computer program. The multiphasic response of plasma radioactivity during a physiological steady state was explained on the basis of rapid hepatic uptake of selenium and its subsequent reappearance in the circulation in protein-bound form followed by further metabolism and excretion of the element. The model provides reference parameter values for 75Se-sodium selenite kinetics in selenium-replete, mature nonpregnant ewes for comparison with the kinetics in animals whose selenium status may be altered.     

Dynamics of selenite metabolism in young men: studies with the stable isotope tracer method

Dynamics of selenite metabolism in young adult North American men were studied using an amino acid diet and the stable isotope tracer methodology during a short-term selenium replete-restriction phase. During the initial 10 days subjects consumed the diet providing a total daily selenium intake of 107.7 +/- 0.1 micrograms mostly as selenite. This was followed by selenium restriction at 11.4 +/- 0.1 micrograms/day (as impurities from diet components) for the next 34 days. Kinetic studies with the stable isotope tracer 74SeO32- were carried out on days 4 and 39 of the study. Kinetics of excretion of the tracer in feces and urine were followed from which body retention curves were constructed. The retention curves were resolved into two exponential decay components with half-lives of 2.4 +/- 0.3 and 162 +/- 9 days (mean +/- 1 SEM), respectively. Retention data and urine isotope enrichment curves were combined to determine dynamics of changes in the apparent body pool size for selenite (10.4 mg at t----infinity) as well as rates of turnover for this parameter.     

Comparison of the magnitude of the selenite-exchangeable metabolic pool and whole body endogenous selenium in adult rats

The quantitative relationship between the size of the selenite-exchangeable metabolic pool (WSe-EMP) and whole body endogenous selenium (Seend) was investigated in adult male rats. Two experiments based on multiple labeling with stable isotopes were performed. One focused on short-term (7 d, Expt. 1) and the other on long-term (60 d, Expt. 2) relationships. Rats were fed a Torula yeast diet and water supplemented with [76Se]selenite at 0.1 micrograms Se/mL; the in vivo [74Se]selenite tracer was administered orally. Groups of three or four animals were killed at timed intervals and whole carcass or selected organs were analyzed for the stable isotopes 74Se, 77Se and 82Se with hydride generation/inductively coupled plasma mass spectrometry. The value of WSe-EMP was determined from plasma or urine isotope ratios. In Experiment 1, with plasma as the sampling compartment, WSe-EMP at 24 h was 36.5 +/- 1.2% of the baseline value of whole body endogenous selenium (Seend) and 36.3 +/- 1.8% at 7 d. When urine was the sampling compartment, the corresponding values were 3.9 +/- 0.3% and 43.1 +/- 2.8%, respectively. In Experiment 2, WSe-EMP (plasma) was 38.9 +/- 1.3% of Seend at 7 d, increasing to 45.5 +/- 1.6% at 60 d. The corresponding values for urine as the sampling compartment were 45.5 +/- 2.0% (7 d) and 61.5 +/- 1.7% (60 d), respectively.     

Studies in community nutrition: Estimation of sodium output

A practical procedure for estimating dietary intake of salt would facilitate examination of the possible relationship between salt intake and hypertension and could provide feedback to aid persons in changing salt intake when such is recommended. This is a study with 19 subjects of a sampling procedure of estimating sodium output which requires less patient cooperation than that needed for a 24-hr urine sample. Each subject provided timed urine samples for portions of the entire 24-hr period for three consecutive days. Sodium excretion for any one of the entire 24-hr periods correlated well with the mean 24-hr excretion for the 3-day period. The mean sodium excretion of the three 9-hr overnight urine samples correlated equally well with the mean 24-hr excretion for the entire 3-day period. This latter procedure, which puts far less a burden of compliance on the subjects, is suggested as preferable to obtaining a 24-hr urine sample. For purposes of checking compliance with dietary suggestions, it is possible to replace the analysis for sodium with a simple titrator strip that measures chloride.     

Deposition of dietary organic and inorganic selenium in rat erythrocyte proteins

The deposition of selenium (Se) in erythrocyte proteins was studied in rats fed Se as sodium selenite, selenocystine, selenomethionine (Se-Met), high Se wheat or selenium-enriched yeast. Ion-exchange chromatography of acid hydrolyzates of selenium-enriched yeast, high Se wheat and intrinsically labeled 75Se wheat indicated that the majority of the Se was present as Se-Met. Gel filtration (Sephadex G-150) of erythrocyte lysates revealed that Se was deposited mainly in two proteins, glutathione peroxidase (GPx) and hemoglobin (Hb). When selenite or selenocystine was the dietary form of Se, the majority of the erythrocyte Se was present with GPx, but when the Se was supplied from either Se-Met, yeast or wheat, it was deposited to a greater extent in Hb than GPx. Injection of 75Se as either Se-Met or selenite gave results consistent with the feeding studies. 75Se-labeled selenite injection resulted in labeling of primarily GPx, but 75Se-Met injection labeled predominantly Hb. Hence, the dietary forms of Se influence the relative distribution of Se between GPx and Hb in erythrocytes, and this may be a factor contributing to the difference between human and animal erythrocytes.     

Absorption, distribution and excretion of selenium from beef and rice in healthy North American men

Previous metabolic studies of selenium used pure selenium compounds with pharmacologic activities unrelated to selenium nutrition. Healthy men were fed foods naturally high or low in selenium while confined to a metabolic research unit. Selenium intake was 47 microg/d (595 nmol/d) for 21 d while energy intakes and body weights were stabilized and selenium excretion and intake came into metabolic balance. On d 22, selenium intake was changed to either 14 microg/d (177 nmol/d, low selenium) or 297 microg/d (3.8 micromol, high selenium) for the remaining 99 d. The absorption, distribution and excretion of selenium in food were similar to selenomethionine, and distinctly different from sodium selenite. Daily urinary selenium excretion and selenium concentrations in plasma and RBC showed the largest responses to selenium intake relative to interindividual variation. Urinary selenium and plasma selenium responded most rapidly to changes in selenium intake, whereas RBC reflected longer-term selenium intake. Given the difficulty of 24-h urine collections outside a metabolic research unit, RBC and plasma selenium seem to be the most useful indicators of selenium intake. During the intervention period, the high selenium group retained 15 mg (190 micromol) of selenium, with approximately 5 mg (63 micromol) going into skeletal muscle. The low selenium group lost only 0.9 mg (11 micromol) of whole-body selenium but lost 3.3 mg (42 micromol) from muscle, indicating that selenium was redistributed from muscle to tissues that have a higher metabolic priority for selenium such as testes. Fecal excretion decreased by half, representing an important but previously underappreciated adaptation to selenium restriction.     

High dietary intake of sodium selenite induces oxidative DNA damage in rat liver

One mechanism for the cancer-chemopreventive effects of high selenium (Se) intake is hypothesized to be antioxidant protection of DNA. In this work we examine DNA oxidation in whole animals as a function of dietary Se intake and carcinogen administration. Weanling male Sprague-Dawley rats were fed a basal, Torula yeast-based, Se-deficient diet supplemented with 0, 0.15, or 2.0 ppm Se as sodium selenite for 10 wk. They were then injected with 0, 0.1, or 10 mg /kg body weight of the pro-oxidant carcinogen N-nitrosodiethylamine. High levels of carcinogen and high levels of selenite intake each increased concentration of 8-hydroxy-2'-deoxyguanosine in liver DNA. Se-dependent glutathione peroxidase I gene expression and enzyme activity were dramatically reduced by dietary Se deficiency but were unaffected by carcinogen administration. There were no significant main or interactive effects of Se or carcinogen on activity or gene expression of the DNA repair enzyme 8-oxoguanine glycosylase I. These results do not support the hypothesis that high Se intake may be cancer-preventive by inhibiting oxidative DNA damage. Rather than inhibiting oxidative DNA damage, these findings suggest that high dietary intake of inorganic Se may promote in vivo DNA oxidation.     

Serum prostate-specific antigen but not testosterone levels decrease in a randomized soy intervention among men

BACKGROUND: Low prostate cancer incidence and high soy intake in Asian countries suggest a possible protective effect of soy foods against prostate cancer. The goal of this pilot study was to evaluate the feasibility of a randomized, crossover soy trial among men and to investigate the effects of daily soy intake on serum prostate-specific antigen (PSA) and testosterone levels. METHODS: We randomized 24 men to a high or a low soy diet for 3 months. After a 1-month washout period, the men crossed over to the other treatment. During the high soy diet, the men consumed two daily soy servings; during the low soy diet, they maintained their usual diet. During the entire study each man donated four blood samples and five overnight urine samples. Dietary compliance was assessed by soy calendars, 24-h dietary recalls, and urinary isoflavone excretion measured by high-pressure liquid chromatography with photodiode array detection. Blood samples were analyzed for serum testosterone and PSA by radioimmunoassay. When necessary, variables were log transformed. Two sample t-tests compared the two groups before each study period. Mixed models incorporating the repeated measurements were used to evaluate the effect of the soy diet on urinary isoflavone excretion and serum analytes. RESULTS: Twenty-three men aged 58.7+/-7.2 years completed the study. The compliance with the study regimen was high according to self-reported soy food intake and urinary isoflavone excretion. No significant between-group and within-group differences were detected. During the high soy diet, dietary isoflavone intake and urinary isoflavone excretion increased significantly as compared to the low soy diet. A 14% decline in serum PSA levels (P=0.10), but no change in testosterone (P=0.70), was observed during the high soy diet in contrast to the low soy diet. CONCLUSION: The high adherence as shown by three measures of compliance in this pilot trial demonstrated the feasibility of an intervention based on soy foods among free-living men.