Abstract

Background: Selenium and α-tocopherol, the major form of vitamin E in supplements, appear to have a protective effect against prostate cancer. However, little attention has been paid to the possible role of γ-tocopherol, a major component of vitamin E in the U.S. diet and the second most common tocopherol in human serum. A nested case–control study was conducted to examine the associations of α-tocopherol, γ-tocopherol, and selenium with incident prostate cancer. Methods: In 1989, a total of 10 456 male residents of Washington County, MD, donated blood for a specimen bank. A total of 117 of 145 men who developed prostate cancer and 233 matched control subjects had toenail and plasma samples available for assays of selenium, α-tocopherol, and γ-tocopherol. The association between the micronutrient concentrations and the development of prostate cancer was assessed by conditional logistic regression analysis. All statistical tests were two-sided. Results: The risk of prostate cancer declined, but not linearly, with increasing concentrations of α-tocopherol (odds ratio highest versus lowest fifth = 0.65; 95% confidence interval = 0.32–1.32; Ptrend = .28). For γ-tocopherol, men in the highest fifth of the distribution had a fivefold reduction in the risk of developing prostate cancer than men in the lowest fifth (Ptrend = .002). The association between selenium and prostate cancer risk was in the protective direction with individuals in the top four fifths of the distribution having a reduced risk of prostate cancer compared with individuals in the bottom fifth (Ptrend = .27). Statistically significant protective associations for high levels of selenium and α-tocopherol were observed only when γ-tocopherol concentrations were high. Conclusions: The use of combined α- and γ- tocopherol supplements should be considered in upcoming prostate cancer prevention trials, given the observed interaction between α-tocopherol, γ-tocopherol, and selenium.

A protective effect of selenium and α-tocopherol, the major form of vitamin E in supplements, against prostate cancer has been observed in clinical trials (1,2) designed to test the efficacy of these micronutrients against skin and lung cancers, respectively. These findings have prompted the design of a randomized trial specifically designed to test the efficacy of selenium and α-tocopherol for the prevention of prostate cancer (3). Additional support for the use of selenium as a preventive agent against prostate cancer comes from observational studies (46) that found higher serum or toenail concentrations of selenium to be associated with a reduced risk of cancer at several sites, including prostate cancer.

Results of studies (711) of serum levels of α- and γ-tocopherol have, however, been inconsistent. An observational study (12) of health professionals reported a protective effect of supplemental vitamin E against metastatic or fatal prostate cancer, but only among current or recent smokers. Few studies have examined the combined effect of selenium and the tocopherols against prostate cancer. Selenium has often been included in studies of the relationships of antioxidant micronutrients with cancer, primarily because it is a component of glutathione peroxidase, which has antioxidant activity (13). Multiple selenium compounds may have a role in inhibiting carcinogenesis, but which specific selenium compounds are involved and how they might act are still uncertain (14). The role of glutathione as an antioxidant may be related to its ability to interact with vitamin E to protect cells against oxygen-reactive radicals (15,16). In the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study (ATBC) (7) among Finnish smokers, there was no association of selenium intake at baseline with subsequent prostate cancer among the groups not given α-tocopherol; however, among those with α-tocopherol supplementation, higher intakes of selenium were associated with slightly lower risks of prostate cancer, although the reduction was not statistically significant.

Little attention has been paid to the possibility that γ-tocopherol, a major component of tocopherol in the U.S. diet and the second most common tocopherol in human serum, might also be involved in the pathogenesis of prostate cancer. To examine whether it has a protective role against incident prostate cancer in combination with α-tocopherol and selenium, a nested case–control study was conducted among participants of a blood specimen bank in Washington County, MD.

Subjects and Methods

Study Population

In 1989, a total of 10 456 male residents of Washington County donated blood for a specimen bank in a campaign nicknamed “CLUE II,” based on the campaign slogan “Give us a clue to cancer and heart disease.” Brief medical histories, including medication use in the previous 48 hours, smoking history, self-report of height and weight at blood donation, and age 21 years, and blood samples were collected by trained personnel housed in large trailers that were moved to many locations in the county to be sure that all adult males had an opportunity to participate. On the basis of the 1990 U.S. Bureau of the Census (Suitland, MD) data, it was estimated that 29% of the resident males more than 45 years of age did so. (Only one case subject in this study was <45 years old.) Blood was drawn into 20-mL heparinized Vacutainers™ (Bectin Dickinson, Rutherford, NJ) and immediately refrigerated at 4 °C until centrifuged at 1000g for 30 minutes at room temperature; divided into aliquots of plasma, buffy coat, and a portion of the red blood cells; and frozen at −70 °C within a few hours of collection. Specimens have been kept at this temperature until removed for dividing into aliquots for this study. At the time of blood draw, information was obtained on all supplements and medication taken within 48 hours.

Each participant was asked to mail in a nail clipping from the big toe. Eighty-six percent of male participants over the age of 45 years returned nail clippings. After having been classified as large, medium, or small, toenail clippings were stored at room temperature in brown paper envelopes.

In 1996, participants were mailed a questionnaire that included questions on family history of cancer, self-report of cancer, and vasectomy. Sixty-nine percent of the cohort responded to the follow-up questionnaire.

Cases of prostate cancer were identified by linking the list of CLUE II participants who were residents of Washington County with the Washington County Cancer Registry and, since 1995, also with the Maryland Cancer Registry. In 1997 (the latest reporting year available from the State Registry), the total number of cancer cases in Maryland was similar to the estimated number of cases based on American Cancer Society estimates (101.4% of estimated cases). About 20% of prostate cancer cases were unstaged compared with 13% in Surveillance, Epidemiology, and End Results Program (SEER)1(17). In 1997, the Washington County Registry had 86 cases of prostate cancer reported; 84 of these were also in the Maryland State Registry. The cohort is stable with few (<1% per year) lost to follow-up. During the period from January 1990 through September 1996, 145 men were identified who had pathologic confirmation of prostate cancer as their first diagnosis of cancer. Of these, 119 men had submitted a toenail clipping in 1989.

For each case subject, two control subjects were selected who were Washington County residents, participants in CLUE II, and nearest in age to the case subject to be matched, with one control subject being younger and the other being older than the case subject. Control subjects were also matched to case subjects with respect to race, date of participation in the CLUE II program within 3 weeks, and to the extent possible, the size of the toenail clipping.

Written informed consent was obtained at the time of blood donation. The study was approved by the Committee on Human Research, The Johns Hopkins School of Hygiene and Public Health, Baltimore, MD.

Laboratory Methods

Plasma samples were thawed in ice water under dim yellow light and were sent under dry ice to the Department of Biomedical Research, Our Lady of Mercy Medical Center, Bronx, NY, for assays of α-tocopherol and γ-tocopherol by high-performance liquid chromatography (18). Toenail clippings were sent to the Research Reactor Facility, University of Missouri, Columbia, where, after being cleaned, they were assayed for selenium by neutron activation analysis (19). Both types of specimens were submitted to the laboratories in sets containing one case and two control specimens. The position of the case specimen within each set of three was random. Sets were given a serial number, and specimens within each set were identified only by the CLUE II study number. Laboratories were requested to assay sets in set number order and to be sure that the specimens in each set were assayed on the same day and with the same reagents.

Intra-assay coefficients of variation for α- and γ-tocopherol, determined from quality-control plasma specimens interspersed randomly throughout the sets, were 3% and 6%, respectively. The assay for γ-tocopherol is able to detect less than 5 ng of γ-tocopherol per injection (1/10th fraction of 100 μL extracted). The coefficient of variations for selenium by neutron activation analyses was 5%. Levels of cholesterol and triglycerides were measured by enzymatic colorimetric method by use of the Beckman™ (Beckman Instruments, Inc., Fullerton, CA) autoanalyzer.

Assays were not possible for specimens from two case subjects and one control subject, reducing the number of case subject–control subject sets for analysis to 117 sets: 116 sets with one case subject and two control subjects and one set with one case subject and one control subject.

Statistical Analyses

Concentrations of selenium and α- and γ-tocopherols were compared with the use of the Wilcoxon signed-rank test taking the average of the control subject values. Conditional logistic regression models were used to assess the association between micronutrient concentrations and the risk of developing prostate cancer. Concentrations of α- and γ-tocopherols and selenium were categorized into fifths on the basis of the distributions in the control subject groups. Odds ratios (ORs) and the corresponding 95% confidence intervals (CIs) for each fifth were calculated with the use of the lowest fifth as the referent category. Findings were adjusted for the effects of education, body mass index (BMI) at age 21 years, and hours since last meal. BMI was calculated as kilograms per meter square based on self-report of height and weight at blood donation. Tests for trend were assessed by entering a categoric variable in the conditional logistic regression model, with the median value of each fifth of the distribution among the control subjects as the score. Analyses were performed stratifying by stage at diagnosis and year from blood donation to year of diagnosis. All P values were based on two-tailed tests and were considered to be statistically significant at P<.05.

The association between prostate cancer and the combined effect of selenium and the tocopherols was assessed by dichotomizing the nutrients according to the median control subject value and stratifying by combinations of high/low values for the three nutrients. Two-way and three-way interactions among selenium and α- and γ-tocopherols were assessed. Interaction effects were assessed by the likelihood ratio test for the interaction terms added to a model that already contained the dichotomous variables for micronutrients.

The analyses for α- and γ-tocopherols were done with and without adjustment for blood lipids. Because tocopherols reside mainly in low-density lipoproteins and biomembranes, individuals who have higher lipid levels tend to have higher tocopherol levels. To account for the correlation between lipids and tocopherols, we first regressed tocopherols on total lipids. The residuals from this regression model, which represent the lipid-adjusted tocopherol levels, were log transformed and entered into the logistic regression models (20).

Results

Comparisons of case and control subjects with respect to their characteristics at study baseline in 1989 are shown in Table 1. The great majority of the study population was white, as expected from the 1990 racial composition of Washington County, in which only 1.9% of persons more than age 45 years were black. Case subjects had more years of schooling than control subjects. Case subjects were more likely to be in the upper four fifths of the control subject BMI distribution at age 21 years but in the lower fifth of the control subject BMI distribution at baseline in 1989. The two groups were similar with respect to age, cigarette smoking, use of vitamin supplements during the 48 hours prior to blood donation, and hours from last meal to the time of blood draw. No case or control subjects reported taking selenium supplements. Seventy-eight percent of the case subjects were diagnosed more than 2 years after blood donation. Staging data were available for 68% of the case subjects (21). Among those with staging information, 86% of case subjects had disease of pathologic stage II or higher at the time of diagnosis.

Median concentrations of α-tocopherol and γ-tocopherol were lower among the prostate cancer case subjects than among the control subjects; the difference was statistically significant for γ-tocopherol (Table 2). Median concentrations of selenium in toenail clippings of case and control subjects were almost identical. The differences in concentrations of γ-tocopherol did not differ by age at diagnosis or by stage of disease (data not shown).

Table 3 shows the association between α-tocopherol, γ-tocopherol, and selenium with subsequent prostate cancer according to fifths of the control subject distribution. The risk of prostate cancer was lower among men with higher concentrations of α-tocopherol, γ-tocopherol, and selenium. The risk of prostate cancer declined with increasing concentrations of α-tocopherol but not in a linear fashion (OR highest to lowest fifth = 0.65; 95% CI = 0.32–1.32; Ptrend = .28). The strongest association was observed for γ-tocopherol. Compared with men in the lowest fifth, men in the highest fifth of the distribution of γ-tocopherol had a fivefold reduction in the risk of developing prostate cancer; the trend in this association was not monotonic but was nevertheless highly statistically significant (Ptrend = .002). The association between γ-tocopherol and prostate cancer was strongest among men diagnosed with stage II or higher disease. The association of selenium and prostate cancer was in the protective direction, with individuals in the top four fifths of the distribution having a lower risk of prostate cancer; no monotonic trend was observed (Ptrend = .27). Adjustments for effects of total lipids, education, BMI at age 21 years, and hours since last meal made no important differences in the estimates of risk associated with serum concentrations of the tocopherols or with selenium concentrations in toenails. Results were similar when individuals who took vitamin supplements within 48 hours of blood donation or who were diagnosed with prostate cancer within 2 years following blood draw were excluded.

For α-tocopherol and selenium, protective associations with prostate cancer risk were observed when γ-tocopherol concentrations were above the control subject-based median value. γ-Tocopherol concentrations above the median value were associated with a 50% reduction in prostate cancer risk when α-tocopherol or selenium levels were also above the control subject–subject median values (P = .09 and .11, respectively), whereas the associations were relatively weak when α-tocopherol or selenium levels were below the medians. Joint associations between α-tocopherol and selenium and subsequent prostate cancer are shown in Table 4. Compared with individuals with low concentrations of all three micronutrients, concentrations of selenium and α-tocopherol above the median control subject distribution were associated with a statistically significant decreased risk of prostate cancer only in the presence of concentrations of γ-tocopherol above the median level. The strengths of these interactions approach statistical significance (P = .07 for the two-way and three-way interaction terms).

Discussion

The major form of vitamin E in supplements is α-tocopherol, whereas γ-tocopherol is the main form in the diet and is becoming increasingly important. During the last 20 years, the U.S. diet has changed to include soybeans, and soybean oil contains an 8 : 1 ratio of γ-tocopherol : α-tocopherol (22,23). In this prospective study, higher concentrations of plasma γ-tocopherol were associated with a statistically significant lower risk of developing prostate cancer. The risk of prostate cancer in the highest fifth compared with the lowest fifth of α-tocopherol was 0.65, similar to the risk reduction associated with α-tocopherol supplementation observed in the ATBC (2,7). Our findings suggest that a higher magnitude of risk reduction may occur if concentrations of γ-tocopherol are increased. On the basis of our results, a stronger protective association with selenium should also be observed in the presence of high γ-tocopherol concentrations.

In support of our finding, laboratory studies suggest γ-tocopherol to be a better inhibitor in vitro of electrophilic mutagens than α-tocopherol. Both α-tocopherol and γ-tocopherol have similar ability to quench singlet oxygen (24), while an in vitro study (25) has suggested that the α- and γ- isomers may have different antioxidant activities. That report (25) found that peroxynitrite-induced lipid peroxidation and lipid hydroperoxide formation in liposomes exposed to peroxynitrite were inhibited more effectively by γ- than by α-tocopherol. γ-Tocopherol has also been shown to inhibit neoplastic transformation in a murine fibroblast transformation assay (26).

We have not been able to identify other prospective studies that simultaneously investigated the association between concentrations of α-tocopherol, γ-tocopherol, and selenium and the risk of prostate cancer. Of interest, in a study of baseline intake of nutrients and association with subsequent prostate cancer among participants of the ATBC trial, a protective association for γ-tocopherol intake and selenium intake was seen only among participants who received α-tocopherol supplementation (7). Nomura et al. (8) investigated serum α- and γ-tocopherol concentrations and the development of prostate cancer among a cohort of Japanese-American men. A statistically nonsignificant lower risk of prostate cancer was observed only among men in the highest fourth of γ-tocopherol distribution. Low serum vitamin E concentrations (presumably α-tocopherol, since most assays of vitamin E were actually α-tocopherol) were associated with subsequent prostate cancer mortality, particularly among smokers, among participants of the Basel study (10,27). In a prospective study of toenail selenium levels and advanced prostate cancer [stage C or D (28) ] conducted among health professionals, higher selenium levels were associated with a lower risk of advanced-stage disease (6). Similar to our findings, the patterns of association appeared to be more compatible with a threshold effect than with a linear dose–response trend.

Replication of these findings in other populations is encouraged, particularly given the implication of the results for a proposed chemoprevention trial of prostate cancer using selenium and α-tocopherol as the intervention agents (3). We observed a statistically significant protective association against prostate cancer only with γ-tocopherol. Moreover, protective associations between selenium and α-tocopherol concentrations and subsequent prostate cancer were observed only in the presence of higher concentrations of γ-tocopherol. Since supplementation with α-tocopherol may lower γ-tocopherol concentrations in plasma and tissues (29,30), consideration should be given to supplementation with combined α- and γ-tocopherols in future prostate cancer prevention trials.

Table 1.

Mean ages and percentage distributions of characteristics of prostate cancer case and control subjects (Washington County, MD, 1989)*

Characteristic Case subject (n = 117) Control subject (n = 233) Matched OR (95% CI) 
*OR = odds ratio, CI = confidence interval, and BMI = body mass index. 
†Obtained at baseline by self-report of weight and height at year 21. 
‡Information obtained from 1996 follow-up questionnaire. 
Age at blood donation, y 66.4 ± 7.4 66.3 ± 7.5 — 
Age at diagnosis, y 70.5 ± 6.9 — — 
Educational level, y, %    
    <12 23.9 38.2 1.00 (referent) 
    12 41.0 36.9 1.85 (1.03–3.33) 
    >12 35.0 24.9 2.32 (1.27–4.24) 
Cigarette smoking, %    
    Never smoker 41.0 42.1 1.00 (referent) 
    Former smoker 50.4 50.6 1.01 (0.64–1.62) 
    Current smoker 8.6 7.3 1.22 (0.50–2.98) 
Hours since last meal, %    
    <1 11.1 11.3 1.00 (referent) 
    1 to <3 29.9 27.3 1.05 (0.48–2.32) 
    3 to <8 43.6 49.8 0.86 (0.41–1.82) 
    ≥8 15.4 11.7 1.28 (0.53–3.09) 
    Unknown 0.0 0.7 — 
BMI at baseline, kg/m2, %    
    17.2–24.1 29.1 19.7 1.00 (referent) 
    24.2–25.4 15.4 20.2 0.53 (0.26–1.06) 
    25.5–26.9 14.5 20.2 0.50 (0.25–1.01) 
    27.0–28.9 17.1 20.6 0.57 (0.28–1.14) 
    29.0–37.7 23.9 19.3 0.84 (0.44–1.60) 
BMI at age 21 y, kg/m2, %†    
    14.8–19.7 12.0 19.7 1.00 (referent) 
    19.8–21.2 21.4 22.3 1.68 (0.78–3.64) 
    21.3–22.8 24.8 18.0 2.34 (1.07–5.12) 
    22.9–24.9 19.7 19.7 1.59 (0.75–3.38) 
    25.0–32.1 21.4 20.2 1.82 (0.83–4.01) 
Vitamin supplement use within 48 h before blood donation, %    
    No 71.8 75.1 1.00 (referent) 
    Yes 26.5 23.2 1.21 (0.72–2.02) 
    Unknown 1.7 1.7  
Family history, father and brothers, %‡    
    No 65.8 72.1 1.00 (referent) 
    Yes 7.7 6.9 1.19 (0.49–2.86) 
    Missing 26.5 21.0  
Vasectomy, %‡    
    No 41.0 59.7 1.00 (referent) 
    Yes 7.7 6.9 1.56 (0.64–3.80) 
    Missing data 51.3 33.5  
Stage, %    
    0 1.7   
    I 7.7   
    II 35.9   
    III 17.9   
    IV 5.1   
    Missing data 31.6   
Year of diagnosis, %    
    1990–1991 22.2   
    1992–1993 36.8   
    1994–1996 41.0    
Characteristic Case subject (n = 117) Control subject (n = 233) Matched OR (95% CI) 
*OR = odds ratio, CI = confidence interval, and BMI = body mass index. 
†Obtained at baseline by self-report of weight and height at year 21. 
‡Information obtained from 1996 follow-up questionnaire. 
Age at blood donation, y 66.4 ± 7.4 66.3 ± 7.5 — 
Age at diagnosis, y 70.5 ± 6.9 — — 
Educational level, y, %    
    <12 23.9 38.2 1.00 (referent) 
    12 41.0 36.9 1.85 (1.03–3.33) 
    >12 35.0 24.9 2.32 (1.27–4.24) 
Cigarette smoking, %    
    Never smoker 41.0 42.1 1.00 (referent) 
    Former smoker 50.4 50.6 1.01 (0.64–1.62) 
    Current smoker 8.6 7.3 1.22 (0.50–2.98) 
Hours since last meal, %    
    <1 11.1 11.3 1.00 (referent) 
    1 to <3 29.9 27.3 1.05 (0.48–2.32) 
    3 to <8 43.6 49.8 0.86 (0.41–1.82) 
    ≥8 15.4 11.7 1.28 (0.53–3.09) 
    Unknown 0.0 0.7 — 
BMI at baseline, kg/m2, %    
    17.2–24.1 29.1 19.7 1.00 (referent) 
    24.2–25.4 15.4 20.2 0.53 (0.26–1.06) 
    25.5–26.9 14.5 20.2 0.50 (0.25–1.01) 
    27.0–28.9 17.1 20.6 0.57 (0.28–1.14) 
    29.0–37.7 23.9 19.3 0.84 (0.44–1.60) 
BMI at age 21 y, kg/m2, %†    
    14.8–19.7 12.0 19.7 1.00 (referent) 
    19.8–21.2 21.4 22.3 1.68 (0.78–3.64) 
    21.3–22.8 24.8 18.0 2.34 (1.07–5.12) 
    22.9–24.9 19.7 19.7 1.59 (0.75–3.38) 
    25.0–32.1 21.4 20.2 1.82 (0.83–4.01) 
Vitamin supplement use within 48 h before blood donation, %    
    No 71.8 75.1 1.00 (referent) 
    Yes 26.5 23.2 1.21 (0.72–2.02) 
    Unknown 1.7 1.7  
Family history, father and brothers, %‡    
    No 65.8 72.1 1.00 (referent) 
    Yes 7.7 6.9 1.19 (0.49–2.86) 
    Missing 26.5 21.0  
Vasectomy, %‡    
    No 41.0 59.7 1.00 (referent) 
    Yes 7.7 6.9 1.56 (0.64–3.80) 
    Missing data 51.3 33.5  
Stage, %    
    0 1.7   
    I 7.7   
    II 35.9   
    III 17.9   
    IV 5.1   
    Missing data 31.6   
Year of diagnosis, %    
    1990–1991 22.2   
    1992–1993 36.8   
    1994–1996 41.0    
Table 2.

Median serum concentrations for α-tocopherol, γ-tocopherol, and toenail concentrations for selenium by case–control subject status (Washington County, MD, 1989)

 Median (range)   
 Case subjects Control subjects % difference Two-sided P
*P value; Wilcoxon signed-rank test. 
α-Tocopherol, mg/dL 1.27 (0.78–10.39) 1.31 (0.65–4.76) −3.1 .07 
γ-Tocopherol, mg/dL 0.24 (0.03–0.67) 0.28 (0.04–1.05) −14.3 .0002 
Selenium, ppm 0.77 (0.07–2.27) 0.79 (0.48–1.98) −2.5 .26  
 Median (range)   
 Case subjects Control subjects % difference Two-sided P
*P value; Wilcoxon signed-rank test. 
α-Tocopherol, mg/dL 1.27 (0.78–10.39) 1.31 (0.65–4.76) −3.1 .07 
γ-Tocopherol, mg/dL 0.24 (0.03–0.67) 0.28 (0.04–1.05) −14.3 .0002 
Selenium, ppm 0.77 (0.07–2.27) 0.79 (0.48–1.98) −2.5 .26  
Table 3.

Association between α-tocopherol, γ-tocopherol, and selenium and subsequent prostate cancer according to fifths of the distribution in control subjects (Washington County, MD, 1989)*,†

 Low 1 High 5 Two-sided Ptrend 
*OR = odds ratio; CI = confidence interval, and BMI = body mass index. 
†Quintiles: α-tocopherol, mg/dL—1.05, 1.21, 1.39, and 1.74; γ-tocopherol, mg/dL—0.17, 0.25, 0.32, 0.41; and selenium, ppm—0.69, 0.75, 0.81, and 0.91. 
‡Adjusted for total lipids, BMI at age 21 years, education, and hours since last meal. 
§Adjusted for BMI at age 21 years, education, and hours since last meal. 
α-Tocopherol       
    No. of case/control subjects 29/49 24/44 20/58 26/45 18/47  
    Matched OR (95% CI) 1.00 (referent) 0.91 (0.46–1.78) 0.67 (0.32–1.41) 0.93 (0.47–1.83) 0.65 (0.32–1.32) .28 
    Adjusted‡ 1.00 (referent) 0.79 (0.38–1.65) 0.32 (0.13–0.75) 0.83 (0.40–1.73) 0.64 (0.30–1.36) .37 
γ-Tocopherol       
    No. of case/control subjects 26/45 34/48 30/52 20/44  7/44  
    Matched OR (95% CI) 1.00 (referent) 1.30 (0.66–2.58) 1.00 (0.50–1.98) 0.68 (0.32–1.46) 0.19 (0.07–0.56) .002 
    Adjusted‡ 1.00 (referent) 1.06 (0.51–2.22) 1.20 (0.60–2.41) 0.75 (0.34–1.64) 0.25 (0.09–0.68) .013 
Selenium       
    No. of case/control subjects 32/45 20/48 21/46 24/47 20/47  
    Matched OR (95% CI) 1.00 (referent) 0.57 (0.28–1.15) 0.63 (0.31–1.26) 0.72 (0.37–1.40) 0.58 (0.29–1.18) .27 
    Adjusted§ 1.00 (referent) 0.41 (0.18–0.93) 0.55 (0.26–1.17) 0.66 (0.33–1.33) 0.38 (0.17–0.85) .12  
 Low 1 High 5 Two-sided Ptrend 
*OR = odds ratio; CI = confidence interval, and BMI = body mass index. 
†Quintiles: α-tocopherol, mg/dL—1.05, 1.21, 1.39, and 1.74; γ-tocopherol, mg/dL—0.17, 0.25, 0.32, 0.41; and selenium, ppm—0.69, 0.75, 0.81, and 0.91. 
‡Adjusted for total lipids, BMI at age 21 years, education, and hours since last meal. 
§Adjusted for BMI at age 21 years, education, and hours since last meal. 
α-Tocopherol       
    No. of case/control subjects 29/49 24/44 20/58 26/45 18/47  
    Matched OR (95% CI) 1.00 (referent) 0.91 (0.46–1.78) 0.67 (0.32–1.41) 0.93 (0.47–1.83) 0.65 (0.32–1.32) .28 
    Adjusted‡ 1.00 (referent) 0.79 (0.38–1.65) 0.32 (0.13–0.75) 0.83 (0.40–1.73) 0.64 (0.30–1.36) .37 
γ-Tocopherol       
    No. of case/control subjects 26/45 34/48 30/52 20/44  7/44  
    Matched OR (95% CI) 1.00 (referent) 1.30 (0.66–2.58) 1.00 (0.50–1.98) 0.68 (0.32–1.46) 0.19 (0.07–0.56) .002 
    Adjusted‡ 1.00 (referent) 1.06 (0.51–2.22) 1.20 (0.60–2.41) 0.75 (0.34–1.64) 0.25 (0.09–0.68) .013 
Selenium       
    No. of case/control subjects 32/45 20/48 21/46 24/47 20/47  
    Matched OR (95% CI) 1.00 (referent) 0.57 (0.28–1.15) 0.63 (0.31–1.26) 0.72 (0.37–1.40) 0.58 (0.29–1.18) .27 
    Adjusted§ 1.00 (referent) 0.41 (0.18–0.93) 0.55 (0.26–1.17) 0.66 (0.33–1.33) 0.38 (0.17–0.85) .12  
Table 4.

Combined association of α-tocopherol, γ-tocopherol, and selenium with the subsequent development of prostate cancer, (Washington County, MD, 1989)*, †

1
Editor's note: SEER is a set of geographically defined, population-based, central cancer registries in the United States, operated by local nonprofit organizations under contract to the National Cancer Institute (NCI). Registry data are submitted electronically without personal identifiers to the NCI on a biannual basis, and the NCI makes the data available to the public for scientific research.
Supported by Public Health Service grant CA94028 from the NCI, National Institutes of Health, Department of Health and Human Services, and by grant DAMD1–94-J-4265 from the Department of Defense.
These data were supplied in part by the Maryland Cancer Registry, Department of Health and Mental Hygiene, Baltimore.
The Department of Health and Mental Hygiene specifically disclaims responsibility for any analyses, interpretations, or conclusions.

We thank all of the participants of the cohort studies.

References

1
Clark LC, Dalkin B, Krongrad A, Combs GF Jr, Turnbull BW, Slate EH, et al. Decreased incidence of prostate cancer with selenium supplementation: results of a double-blind cancer prevention trial.
Br J Urol
 
1998
;
81
:
730
–4.
2
Heinonen OP, Albanes D, Virtamo J, Taylor PR, Huttunen JK, Hartman AM, et al. Prostate cancer and supplementation with α-tocopherol and β-carotene: incidence and mortality in a controlled subject trial.
J Natl Cancer Inst
 
1998
;
90
:
440
–6.
3
Taylor PR, Albanes D. Selenium, vitamin E, and prostate cancer—ready for prime time? [editorial].
J Natl Cancer Inst
 
1998
;
90
:
1184
–5.
4
Comstock GW, Bush TL, Helzlsouer K. Serum retinol, beta-carotene, vitamin E, and selenium as related to subsequent cancer of specific sites.
Am J Epidemiol
 
1992
;
135
:
115
–21.
5
Willett WC, Polk BF, Morris JS, Stampfer MJ, Pressed S, Rosner B, et al. Prediagnostic serum selenium and risk of cancer.
Lancet
 
1983
;
2
:
130
–4.
6
Yoshizawa K, Willett WC, Morris SJ, Stampfer MJ, Spiegelman D, Rimm EB, et al. Study of prediagnostic selenium level in toenails and the risk of advanced prostate cancer.
J Natl Cancer Inst
 
1998
;
90
:
1219
–24.
7
Hartman TJ, Albanes D, Pietinen P, Hartman AM, Rautalahti M, Tangrea JA, et al. The association between baseline vitamin E, selenium, and prostate cancer in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study.
Cancer Epidemiol Biomarkers Prev
 
1998
;
7
:
335
–40.
8
Nomura AM, Stemmermann GN, Lee J, Craft NE. Serum micronutrients and prostate cancer in Japanese Americans in Hawaii.
Cancer Epidemiol Biomarkers Prev
 
1997
;
6
:
487
–91.
9
Hsing AW, Comstock GW, Abbey H, Polk BF. Serologic precursors of cancer. Retinol, carotenoids, and tocopherol and risk of prostate cancer.
J Natl Cancer Inst
 
1990
;
82
:
941
–6.
10
Eichholzer M, Stahelin HB, Gey KF, Ludin E, Bernasconi F. Prediction of male cancer mortality by plasma levels of interacting vitamins: 17-year follow-up of the prospective Basel study.
Int J Cancer
 
1996
;
66
:
145
–50.
11
Gann PH, Ma J, Giovannucci E, Willett W, Sacks FM, Hennekens CH, et al. Lower prostate cancer risk in men with elevated plasma lycopene levels: results of a prospective analysis.
Cancer Res
 
1999
;
59
:
1225
–30.
12
Chan JM, Stampfer MJ, Ma J, Rimm EB, Willett WC, Giovannucci EL. Supplemental vitamin E intake and prostate cancer risk in a large cohort of men in the United States.
Cancer Epidemiol Biomarkers Prev
 
1999
;
8
:
893
–9.
13
Rotruck JT, Pope AL, Ganther HE, Swanson AB, Hafeman DG, Hoekstra WG. Selenium: biochemical role as a component of glutathione peroxidase.
Science
 
1973
;
179
:
588
–90.
14
Medina D. Mechanisms of selenium inhibition of tumorigenesis.
Adv Exp Med Biol
 
1986
;
206
:
465
–72.
15
Mills GC. Hemoglobin catabolism. I. Glutathione peroxidase, an erythrocyte enzyme which protests hemoglobin from oxidative breakdown.
J Biol Chem
 
1957
;
229
:
189
–97.
16
Levander OA, Burk RF. Selenium. In: Shils ME, Olson JA, Shike M, editors. Modern nutrition in health and disease. 8th ed. Philadelphia (PA): Lea & Febiger; 1994.
17
Cancer incidence in Maryland: 1997. Maryland Cancer Registry. Community and Public Health Administration. Baltimore (MD): Maryland Department of Health and Mental Hygiene; 1997.
18
Sowell AL, Huff DL, Yeager PR, Caudill SP, Gunter EW. Retinol, α-tocopherol, lutein/zeaxanthin, beta-carotene and four retinyl esters in serum determined simultaneously by reversed-phase HPLC with multiwave-length detection.
Clin Chem
 
1994
;
40
:
411
–6.
19
McKown DM, Morris JS. Rapid measurement of selenium in biological samples using instrumental neutron activation analysis.
J Radioanal Chem
 
1978
;
43
:
411
–20.
20
Willett W, Stampfer MJ. Total energy intake: implications for epidemiologic analyses.
Am J Epidemiol
 
1986
;
124
:
17
–27.
21
Beahrs OH, Henson DE, Hutter RV, Kennedy BJ, editors. Handbook for staging of cancer. Philadelphia (PA): Lippincott; 1993.
22
Bieri JG, Evarts RP. Gamma tocopherol: metabolic biological activity and significance in human vitamin E nutrition.
Am J Clin Nutr
 
1974
;
27
:
980
–6.
23
Machlin LS. Vitamin E. In: Machlin LS, editor. Handbook of vitamins. 2nd ed. New York (NY): Marcel Dekker; 1991. p. 99–144.
24
Di Mascio P, Murphy ME, Sies H. Antioxidant defense systems: the role of carotenoids, tocopherols, and thiols.
Am J Clin Nutr
 
1991
;
53
(1 Suppl):
194S
–200S.
25
Christen S, Woodall AA, Shigenaga MK, Southwell-Keely PT, Duncan MW, Ames BN. Gamma-tocopherol traps mutagenic electrophiles such as NO(X) and complements alpha-tocopherol: physiological implica-tions.
Proc Natl Acad Sci U S A
 
1997
;
94
:
3217
–22.
26
Cooney RV, Franke AA, Harwood PJ, Hatch-Pigott V, Custer LJ, Mordan LJ. Gamma-tocopherol detoxification of nitrogen dioxide: superiority to α-tocopherol.
Proc Natl Acad Sci U S A
 
1993
;
90
:
1771
–5.
27
Eichholzer M, Stahelin HB, Ludin E, Bernasconi F. Smoking, plasma vitamins C, E, retinol, and carotene, and fatal prostate cancer: seventeen-year follow-up of the prospective Basel study.
Prostate
 
1999
;
38
:
189
–98.
28
Gittes RF. Carcinoma of the prostate.
N Engl J Med
 
1991
;
324
:
236
–45.
29
Lehmann J, Rao DD, Canary JJ, Judd JT. Vitamin E and relationships among tocopherols in human plasma, platelets, lymphocytes, and red blood cells.
Am J Clin Nut
 
1988
;
47
:
470
–4.
30
Vatassery GT, Morley JE, Kuskowski MA. Vitamin E in plasma and platelets of human diabetic patients and control subject.
Am J Clin Nutr
 
1983
;
37
:
641
–4.