We examined the validity of using the selenium level in a single biological specimen as a surrogate measure of usual intake. We used data from 77 free-living adults from South Dakota and Wyoming. Subjects provided multiple 1-day duplicate-plate food composites, repeated specimens of blood and toenails, and 24-hour urine collections. We developed a statistical calibration method that incorporated measurement error correction to analyze the data. The Pearson correlation coefficients between selenium intake and a single selenium status measure, after deattenuation to adjust for the effect of within-person variation in intake, were: 0.78 for whole blood, 0.74 for serum, 0.67 for toenails, and 0.86 for urine. We present formulas to estimate the intake of individuals, based on selenium levels in a single specimen of blood, toenails, or urine. In these data, the concentration of selenium in a single specimen of whole blood, serum, or toenails served reasonably well as a measure for ranking subjects according to long-term selenium intake but provided only a rough estimate of intake for each subject.

To determine whether high dietary selenium intake was associated with adverse effects, selenium in diet, blood, and toenails was studied in relation to human health in adults residing in western South Dakota and eastern Wyoming. Over a 2-y period 142 subjects were recruited from households selected at random and from ranches where unusually high selenium intakes were suspected. Subjects completed health questionnaires, underwent physical examinations, provided blood samples for clinical assessment, and provided blood, urine, toenails, and duplicate-plate food collections for selenium analysis. About half of the 142 free-living subjects had selenium intakes greater than 2.54 mumol/d (200 micrograms/d) (range 0.86-9.20 mumol/d, or 68-724 micrograms/d). Physical findings characteristic of selenium toxicity were not present nor were clinically significant changes in laboratory tests or frequency of symptoms related to selenium in the blood, toenails, or diet. We found no evidence of toxicity from selenium in subjects whose intake was as high as 9.20 mumol/d (724 micrograms/d).

Duplicate meals, serum, whole blood, and toenails were collected every 3 mo for 1 y from a group of 44 free-living adults residing in high-selenium areas of South Dakota and Wyoming to assess the relation of selenium intake to indices of selenium status. The average selenium values for the group were as follows: dietary intake, 174 +/- 91 micrograms/d (mean +/- SD), 2.33 +/- 1.08 micrograms/kg body wt; serum, 2.10 +/- 0.38 mumol/L; whole blood, 3.22 +/- 0.79 mumol/L; and toenails, 15.2 +/- 3.0 nmol/g. Selenium intake (micrograms/kg body wt) was strongly correlated (all values, P less than 0.01) with selenium concentration of serum (r = 0.63), whole blood (r = 0.62), and toenails (r = 0.59). Men and women had similar mean values of serum, whole blood, and toenail selenium despite higher selenium intakes in men. Smokers had lower tissue selenium concentrations than did nonsmokers due, at least in part, to lower selenium intake. Age was not associated with tissue selenium content. Of the variables examined selenium intake was clearly the strongest predictor of tissue selenium concentration.

The relationship of whole blood selenium (Se) to glutathione peroxidase (GPX) activity was examined for individuals in New Zealand, Oregon, and South Dakota who represented, respectively, populations with exposure to low, medium, and high amounts of Se. The mean (respective) blood Se levels were 60, 200, and 400 ng/ml. Intergroup differences in blood Se levels were highly significant (P less than 0.001). GPX assays were performed using two variations of an enzyme-coupled procedure to assess the equivalence of the two methods. Despite a fourfold difference in absolute activities measured by these methods, the GPX activities were highly correlated (r = .86) between procedures. Average blood GPX activity was significantly lower (P less than 0.001) for the New Zealand group compared with the other two groups, but there was no difference in GPX activities between the Oregon and South Dakota groups. Linear regression of GPX vs. Se values within each group indicated a significant correlation of these parameters only in the New Zealand group (r = .46, P less than 0.01). Comparison of these parameters for combined data from all three groups also showed a significant positive correlation (r = .60, P less than 0.001). A saturation model (In GPX = k1 + k2 (Se)-1)) fits the combined data better (r = .80, P less than 0.01) than does direct comparison of the two parameters. These results suggest that GPX activity is an appropriate indicator of human Se status only in populations with below normal exposure to Se, as activity of this enzyme is saturated at relatively low levels.

No abstract