Serum a-Tocopherol Status in the United States Population

American Journal of EpidemiologyCopyright © 1999 by The Johns Hopkins University School of Hygiene and Pubfic HealthAll rights reserved

Vol. 150, No. 3Printed In USA.

Serum a-Tocopherol Status in the United States Population: Findings fromthe Third National Health and Nutrition Examination Survey

Earl S. Ford1 and Anne Sowell2

Despite the role vitamin E may have in protecting against various chronic conditions, little is known aboutvitamin E status in the US population. Using data from the Third National Health and Nutrition ExaminationSurvey (1988-1994), the authors examined the distribution and correlates of serum a-tocopherol among 16,295US adults aged 18 or more years. The mean concentration of a-tocopherol was 26.8 umol/liter (geometric mean,25.0 umol/liter). The 25th, 50th, and 75th percentiles were 19.6, 24.1, and 30.4 ujnol/liter, respectively. The meana-tocopherol/cholesterol ratio was 5.1 (geometric mean, 4.9); the 25th, 50th, and 75th percentiles were 4.1, 4.7,and 5.5 (10"3), respectively. About 27% of the US population had a low a-tocopherol concentration (<20umol/liter). After age standardization, 29% of the men, 28% of the women, 26% of the Whites (men, 27%, andwomen, 26%), 4 1 % of the African Americans (men, 42%, and women, 40%), 28% of the Mexican Americans(men, 29%, and women, 27%), and 32% of the other participants (men, 36%, and women, 29%) had this lowconcentration. For all participants, age, educational attainment, serum cholesterol, and several serum vitaminsand carotenoids were directly related to and high density lipoprotein cholesterol was inversely related to seruma-tocopherol concentration in multiple linear regression analysis. Men had a higher concentration than didwomen, and African Americans had the lowest concentration of any racial or ethnic group. These results showthat important proportions of US adults have a low serum a-tocopherol concentration, which may increase theirrisk for chronic diseases in which low dietary intake or blood concentration of a-tocopherol have been implicated.Am J Epidemiol 1999; 150:290-300.

education; ethnic groups; health surveys; tocopherol

Vitamin E, a fat-soluble vitamin and an essentialnutrient, is a potent antioxidant that occurs in a vari-ety of foods, particularly plant and seed oils. Fourisomers of tocopherol are recognized (a-, (3-, y-, and5-tocopherol); a-tocopherol has the highest relativebiologic activity. The antioxidant properties of vita-min E have led researchers to examine whether thisvitamin could help prevent or delay the onset of dis-eases in which oxidation plays a key role in thepathogenesis. Chief among these conditions are car-diovascular disease (1) and cancer (2). Roles for vita-min E have also been postulated in the pathogenesis

Received for publication June 17, 1998, and accepted for publi-cation November 20, 1998.

Abbreviations: HDL cholesterol, high density lipoprotein choles-terol; MET, metabolic equivalent; NHANES III, Third National Healthand Nutrition Examination Survey.

1 Division of Nutrition, National Center for Chronic DiseasePrevention and Health Promotion, Centers for Disease Control andPrevention, Atlanta, GA.

2 Division of Laboratory Sciences, National Center forEnvironmental Health, Centers for Disease Control and Prevention,Atlanta, GA.

Reprint requests to Dr. Earl S. Ford, Division of Nutrition, NationalCenter for Chronic Disease Prevention and Health Promotion,Centers for Disease Control and Prevention, 4770 Buford Highway,MS K26, Atlanta, GA 30341.

or treatment of diabetes (3), immune disorders (4),acquired immunodeficiency syndrome (5), muscledamage in exercise (6), skin pathology (7), woundhealing (8), Parkinson's disease (9), neurologicabnormalities (10, 11), eye diseases (12, 13), pul-monary disease (14), liver disease (15), andmyelodysplastic disorders (16).

Despite the important role vitamin E may play in pre-venting or controlling chronic disease, few population-based studies of distributions and correlates of bloodconcentrations of vitamin E have been performed, andlittle is known about the epidemiology of serum vita-min E concentration in the US population. Most suchstudies had small sample sizes and used selected popu-lations. Except for the Hispanic Health and NutritionExamination Survey, previous national health surveysin the United States have not included vitamin E in thebiochemical analyses. The inclusion of a-tocopherol inNational Health and Nutrition Examination Survey HI(NHANES HI) presents a unique opportunity to studythe epidemiology of this vitamin. Therefore, we exam-ined the distribution of a-tocopherol in the US popula-tion as well as correlates of serum a-tocopherol con-centration.

290

Downloaded from https://academic.oup.com/aje/article-abstract/150/3/290/55283by gueston 12 April 2018

a-Tocopherol Status in the US Population 291

MATERIALS AND METHODS

To obtain information about the health of the US pop-ulation, a representative sample of the noninstitutional-ized US population is periodically asked to participatein NHANES. The latest in this series, NHANES HI, wasstarted in 1988 and completed in 1994. Household inter-views were conducted with selected participants, and atthe conclusion of the interview, adult participants hadtheir blood pressure measured. Participants who thenattended a health examination were asked to completeadditional questionnaires, had blood drawn, and under-went a series of further examinations. An in-homeexamination was conducted for some participants whowere unable, for health reasons, to attend the examina-tion center. Using a multistage, stratified samplingdesign, oversampling of children ages 2 months to 5years, persons aged 60 years and older, AfricanAmericans, and Mexican Americans was performed.Details of the plan and operations of NHANES HI havebeen published (17).

The serum concentration of a-tocopherol was mea-sured using high-performance liquid chromatography atCenters for Disease Control and Prevention laboratories(18). A detailed description of the laboratory methodol-ogy to measure a-tocopherol has been published (19).

Variables included in this analysis were age; sex;race (White, African American, Mexican American,and other); years of education; smoking status (never,former, or current); serum cotinine concentration; sys-tolic blood pressure; serum cholesterol concentration;high density lipoprotein cholesterol (HDL cholesterol)concentration; body mass index; leisure-time physicalactivity; self-reported diabetes mellitus; alcohol use;vitamin and mineral supplement use during the previ-ous 24 hours; fruit and vegetable consumption; cerealconsumption; dietary vitamin E from a single 24-hourrecall; birth control pill use; hormone replacement use;serum folate concentration; red blood cell folate con-centration; and concentrations of vitamin B12, vitaminA, vitamin C, and carotenoids. Serum cotinine con-centration was determined using an enzyme-linkedimmunoassay. We created a metabolic equivalent(MET)-frequency scale for leisure-time physical activ-ity by summing the products of the frequency of phys-ical activity and MET levels for all of the reportedactivities. One MET is the energy expenditure ofapproximately 3.5 ml oxygen/kg body weight/minuteor 1 kcal/kg body weight/hour. Alcohol use, fruit andvegetable intake, and cereal consumption wereobtained from a food frequency questionnaire.Because food portion sizes were not collected, wesummed the reported frequencies of use of the variousfoods to create the fruit and vegetable scale and thecereal scale. Vitamin and mineral use was derived

from the question: "Have you taken any vitamins orminerals during the past 24 hours?" Women whoreported that they were taking birth control pills at thetime of the interview were considered to be currentusers. We defined current users of hormone replace-ment therapy as women who reported that they wereusing estrogen or female hormones (pills, vaginalcream, suppositories, injections, or skin^atches) at thetime of the interview.

To calculate age-standardized prevalence estimates,we adjusted the data to the age distribution of the 1980US population. Rank correlation coefficients wereobtained by calculating Pearson correlation coefficientsof the ranks of continuous variables. We were unable tocalculate appropriate p values, however. Linear regres-sion analyses were performed using log-transformedserum a-tocopherol concentration. Collinearity diag-nostics were performed with Statistical AnalysisSystems software. Logistic regression analyses ofdichotomized serum a-tocopherol were performed toexamine whether factors were independently associ-ated with a concentration of serum a-tocopherol of 20umol/liter or less. Concentrations of serum vitamin Ebelow this point may be associated with increased mor-bidity and mortality due to cardiovascular disease (20,21). All analyses except the correlation analyses wereperformed with SUDAAN to account for the complexsampling design of the survey (22). We used the sam-pling weights for the participants who had had a healthexamination to calculate weighted estimates. Werestricted analyses involving oral contraceptive use towomen who were younger than age 45 years and analy-ses involving hormone replacement therapy to womenwho were age 50 years or older.

RESULTS

Background information on NHANES HI partici-pants aged 18 to over 90 years who attended the mobileexamination center is listed in tables 1 and 2.Significance testing is not provided because these areunadjusted values, and many comparisons are possible.

Among the participants who attended the medicalexamination clinic, 16,295 had a value for serum a-tocopherol concentration, and 16,272 had a value forthe a-tocopherol/cholesterol ratio. The mean concen-tration of a-tocopherol was 26.8 u.mol/liter (geometricmean, 25.0 (xmol/liter) with a range of 0.65-232.18u.mol/liter. The 25th, 50th, and 75th percentiles were19.6, 24.1, and 30.4 u.mol/liter, respectively. The meana-tocopherol/cholesterol ratio was 5.1 (geometricmean, 4.9); the 25th, 50th, and 75th percentiles were4.1, 4.7, and 5.5, respectively.

About 27 percent of the United States population hada serum a-tocopherol concentration of less than 20

Am J Epidemiol Vol. 150, No. 3, 1999


292 Ford and Sowell

TABLE 1. Unadjusted baseline characteristics of men aged 518 years, by sex and race, National Health and NutritionExamination Survey III, 1988-1991

Age (years)Education (years)Current smokeY (%)Serum cotlnine (ng/ml)Systolic blood pressure (mmHg)Serum cholesterol (mmol/liter)HDL cholesterolf (mmol/liter)Body mass index (kg/m2)Leisure-time physical activity

(frequency-METf)History of diabetes (%)Alcohol use (drinks/month)Vitamin and mineral supplement

use(%)Fruit and vegetable intake

(frequency/month)Cereal intake (frequency/month)Dietary vitamin E (a-tocopherol

equivalents)

I W .

3,2313,2203,2823,0923,1273,1273,1043,231

3,2313,2293,228

3,226

3,2293,227

3,121

White

Mean*(SE)t

44.2 (0.5)12.9(0.1)30.6(1.2)99.0 (4.0)

122.3 (0.5)5.2 (<0.1)1.2 (<0.1)

26.6(0.1)

128.8 (4.4)4.8 (0.3)

12.4(0.6)

21.3(0.9)

124.2(1.6)10.4(0.3)

11.7(0.3)

African American

kjrtno.

2,2242,2002,2882,0122,1242,0262,0172,217

2,2242,2232,224

2,180

2,2242,224

2,132

Mean*(SE)

40.0 (0.5)11.5(0.1)37.7(1.1)

119.6(4.4)125.6 (0.5)

5.1 (<0.1)1.4 (<0.1)

26.4(0.1)

156.6(5.6)5.4 (0.5)

13.6(0.7)

14.9(1.1)

120.5(2.1)7.9 (0.3)

10.0(0.2)

Mexican American

no.

2,3562,3312,4302,2432,2492,2712,2482,347

2,3562,3522,356

2,343

2,3562,356

2,265

Mean*(SE)

35.6 (0.5)9.1 (0.2)

29.7 (1.2)39.7(2.1)

119.7(0.5)5.1 (<0.1)1.2 (<0.1)

26.7 (0.2)

108.3 (5.2)4.0 (0.4)

10.2(0.4)

12.8(1.2)

158.3(3.3)7.3 (0.4)

10.8 (0.3)

300297305283295284280298

300300300

289

300299

284

Other

Mean*(SE)

38.7 (0.7)11.4 (0.4)31.7(3.5)64.6 (9.7)

119.7(1.0)5.1 (0.1)1.2 (<0.1)

25.3 (0.3)

133.1 (10.6)2.3 (0.8)

10.1 (1.4)

15.2(2.7)

149.7 (4.9)6.1 (0.7)

9.0 (0.3)

* Unless otherwise noted, ail values are means.t SE, standard error; HDL cholesterol, high density lipoprotein cholesterol; MET, metabolic equivalent.

TABLE 2. Unadjusted baseline characteristics of women aged 518 years, by sex and race, National Health and NutritionExamination Survey III, 1988-1991

Age (years)Education (years)Current smoker (%)Serum cotinine (ng/ml)Systolic blood pressure (mmHg)Serum cholesterol (mmol/liter)HDL cholesterolf (mmol/liter)Body mass Index (kg/m2)Leisure-time physical activity

(f requency-M ETt)History of diabetes (%)Alcohol use (drinks/month)Vitamin and mineral supplement

use (%)Fruit and vegetable intake

(frequency/month)Cereal intake (frequency/month)Dietary vitamin E (a-tocopherol

equivalents)Oral contraceptive use among

women aged 18—45 years (%)Hormone replacement therapy

use among women aged 250years (%)

No.

3,7423,7233,7983,5483,6313,6003,5703,737

3,7423,7353,740

3,721

3,7413,741

3,623

1,489

2,104

White

Mean*(SE)t

46.0 (0.7)12.6(0.1)26.6(1.0)62.7 (3.0)

117.7(0.6)5.4 (<0.1)1.4 (<0.1)

25.9 (0.2)

100.8 (4.2)5.3 (0.5)5.8 (0.4)

31.4(1.1)

138.1 (1.8)10.3 (0.3)

8.3 (0.2)

19.1 (1.0)

23.5(1.4)

African American

No.

2,6822.6692,7532,4332,5442,4492,4342,674

2,6822,6802,682

2,653

2,6822,682

2,590

1,676

1,001

Mean*(SE)

41.1 (0.5)11.7(0.1)26.8(1.2)89.8 (4.6)

120.9(0.5)5.2 (<0.1)1.5 (<0.1)

28.6 (0.2)

77.9 (3.2)8.6 (0.7)4.4 (0.3)

19.2(0.9)

122.3(1.9)8.0 (0.3)

7.5 (0.2)

15.9(1.8)

11.0(1.3)

Mexican American

No.

2,3662,3502,4352,2192,2322,2582,2492,346

2,3662,3612,365

2,367

2,3662,365

2,303

1,510

813

Mean*(SE)

37.3 (0.4)9.1 (0.2)

13.9(1.0)18.3(2.0)

113.9(0.4)5.1 (<0.1)1.4 (<0.1)

27.8 (0.2)

65.6 (3.0)9.1 (0.5)2.3(0.1)

18.1 (1.1)

163.7(2.4)9.0 (0.3)

8.7 (0.4)

16.6(1.0)

14.3(1.7)

No.

394388413367385374372393

394394394

389

394394

371

209

147

Other

Mean*(SE)

40.8(1.3)11.0 (0.4)13.2(2.2)29.5 (6.2)

113.7(1.3)5.2(0.1)1.3 (<0.1)

25.9 (0.4)

81.1 (6.7)5.4(1.4)1.7(0.3)

17.8(1.9)

157.4(5.1)7.0 (0.7)

7.3 (0.4)

12.7(3.1)

9.3 (4.5)

• Unless otherwise noted, all values are means.t SE, standard error; HDL cholesterol, high density lipoprotein cholesterol; MET, metabolic equivalent




(imol/liter. After age standardization, 29 percent of themen, 28 percent of the women, 26 percent of theWhites, 41 percent of the African Americans, 28 percentof the Mexican Americans, and 32 percent of the partic-ipants of another race had this low serum a-tocopherolconcentration. For White, African-American, Mexican-American, and other men, 27, 42, 29, and 36 percent,respectively, had this low concentration. The corre-sponding values for women were 26,40,27, and 29 per-cent, respectively.

After adjustment for age, women had a mean con-centration of serum a-tocopherol similar to that ofmen in each of the racial or ethnic groups (table 3).Whites had significantly higher concentrations thandid African Americans or Mexican Americans.

The strongest positive unadjusted correlations wereobserved for the ranks of serum a-tocopherol concen-tration and the ranks of age and serum cholesterolconcentration. Other strong correlates of serum a-tocopherol concentration were concentrations of redblood cell folate, vitamin A, vitamin C, andcarotenoids except lycopene (tables 4 and 5. The rankorder of the size of the correlation coefficients wasfairly similar across the various sex and racial or eth-nic groups. The most notable inverse correlationswere between a-tocopherol concentrations and HDLcholesterol (especially among men) and serum coti-nine concentration.

Because of possible interrelations between potentialpredictors of serum a-tocopherol concentration, weperformed multiple linear regression analyses (table6). For all participants, age, sex, race or ethnicity, edu-cational attainment, smoking status, serum lipids, vita-min and mineral supplement use, and serum vitaminsand carotenoids were all significant predictors of a-tocopherol concentration. Serum a-tocopherol con-centration was higher with increasing age; educationalattainment; and serum concentrations of cholesterol,vitamins, and carotenoids except a-carotene. Serum a-tocopherol concentration was higher among womenthan among men, among Whites and MexicanAmericans than among African Americans, amongusers of vitamin and mineral supplements than amongnonusers, and among former smokers than amongnever smokers. Alcohol consumption, leisure-timephysical activity, cereal consumption, and fruit andvegetable consumption were not significant predictors.Removing them from the model did not significantlychange the multiple R2.

We also explored whether predictors of serum a-tocopherol concentration differed by sex by includinginteraction terms between sex and the other covariatesin linear regression models. Linear regression resultsstratified by sex are shown in table 6. The regression

coefficient for HDL cholesterol concentration (p <0.001) was larger for women than for men, whereas theregression coefficients for serum cholesterol concentra-tion (p < 0.001), body mass index (p < 0.001), serumfolate concentration (p = 0.013), serum vitamin A con-centration (p < 0.001), and serum lycopene concentra-tion (p = 0.040) were larger for men than for women.

Results from the logistic regression models werevery similar to those of the linear regression models.African Americans were more likely to have a seruma-tocopherol concentration of less than 20 (imol/literthan were Whites (odds ratio = 2.35, 95 percent confi-dence interval: 1.86, 2.96). Use of vitamin and mineralsupplements was inversely associated with the odds ofhaving a low serum a-tocopherol concentration (oddsratio = 0.27, 95 percent confidence interval: 0.19,0.39). In addition, age; educational attainment; serumcholesterol concentration; and concentrations of serumfolate, serum vitamin A, (i-carotene, cryptoxanthin,lutein/zeaxanthin, and lycopene were all inverselyrelated to and HDL cholesterol was positively relatedto a low concentration of a-tocopherol.

DISCUSSION

In a representative sample of the US population,serum a-tocopherol concentration varied considerablyacross sociodemographic groups. Serum a-tocopherolconcentration was higher among Whites than amongAfrican Americans. Because increasing evidence sup-ports an important role for vitamin E in several diseaseprocesses in which oxidative damage is a key feature,it is important to understand whether certain popula-tions are at increased risk for these conditions becauseof low blood a-tocopherol concentration.

Few other studies have reported on racial or ethnicdifferences in serum a-tocopherol concentration. In acase-control study of cancers, African-American con-trols had a lower lipid-adjusted vitamin E concentra-tion than did White controls (23). The low concentra-tion among African Americans may be especiallyrelevant in light of the relatively high mortality fromcardiovascular disease and cancer they experience(24). In 1994, the age-adjusted mortality rate fromtotal heart disease was 37 percent higher amongAfrican-American men than among White men and63 percent higher among African-American womenthan among White women (24). Mortality from coro-nary heart disease among African Americans exceedsthat among Whites. In addition, age-adjusted cancerincidence and mortality rates during 1990-1994 werehigher among African Americans than among Whites(25). Because recent studies suggest that vitamin Esupplementation is associated with a reduced risk forprostate cancer (26), our observation that serum a-



294 Ford and Sowell

TABLE 3. Mean serum concentrations of a-tocopherol and a-tocopherol/cholesterol ratio amongparticipants aged £18 years, National Health and Nutrition Examination Survey III, 1988-1994

Total

SexMenWomen

Race or ethnicityWhiteAfrican AmericanMexican AmericanOther

Sex and race or ethnicityMen

WhiteAfrican AmericanMexican AmericanOther

WomenWhiteAfrican AmericanMexican AmericanOther

Total

SexMenWomen

Race or ethnicityWhiteAfrican AmericanMexican AmericanOther

Sex and raceMen

WhiteAfrican AmericanMexican AmericanOther

WomenWhiteAfrican AmericanMexican AmericanOther

No.

16,295

7,6668,629

6,6844,4584,498

655

3,1052,0182,260

283

3,5792,4402,238

372

16,272

7,6548,618

6,6774,4464,494

655

3,1022,0112,258

283

3,5752,4352,236

372

Unadjustedmean(SE)t

a-Tocopherol

26.8 (0.2)

26.2 (0.3)27.3 (0.3)

27.6 (0.3)22.7 (0.2)*24.8 (0.2)*25.5 (0.6)*

26.8 (0.3)22.4 (0.3)*24.8 (0.3)*25.4 (0.7)

28.3 (0.3)22.8 (0.2)*24.7 (0.2)*25.6 (0.8)*

a- Tocopherol/cholesterol

5.1 (<0.1)

5.0(0.1)5.1 (<0.1)

5.2(0.1)4.4(<0.1)*4.8(<0.1)*4.9(0.1)*

5.1 (0.1)4.4(<0.1)*4.8(<0.1)*5.0(0.1)

5.3 (0.1)4.4(<0.1)»4.9(<0.1)«4.9 (0.1)*

Geometricmean(SE)

25.0 (0.2)

24.5 (0.2)25.4 (0.2)

25.7 (0.3)21.5(0.1)*23.4 (0.2)*23.9 (0.4)*

25.1 (0.3)21.3(0.2)*23.4 (0.2)*23.6 (0.5)*

26.3 (0.3)21.7(0.2)*23.5 (0.2)*24.1 (0.6)*

4.9(<0.1)

4.8(<0.1)4.9(<0.1)

5.0(<0.1)4.3(<0.1)*4.7(<0.1)*4.8(0.1)*

4.9(0.1)4.3(<0.1)*4.7(<0.1)*4.8 (0.1)

5.0(0.1)4.3(<0.1)*4.7(<0.1)*4.8(0.1)*

Age-adjustedmean(SE)

26.6 (0.2)

26.1 (0.3)27.1 (0.3)

27.1 (0.3)23.2 (0.2)*25.9 (0.2)*26.2 (0.7)

26.4 (0.3)22.9 (0.3)*25.8 (0.3)*26.3 (0.9)

27.7 (0.3)23.3 (0.2)*25.9 (0.2)*26.0 (0.8)

5.1 (<0.1)

5.0(<0.1)5.1 (<0.1)

5.1 (<0.1)4.5(<0.1)»4.9(<0.1)*5.0(0.1)*

5.1 (0.1)4.5(<0.1)»4.9(<0.1)*5.1 (0.1)

5.2 (0.1)4.5(<0.1)»4.9(<0.1)*4.9(0.1)*

* p < 0.05 versus Whites.t SE, standard error.

tocopherol concentration is relatively low amongAfrican Americans may be highly relevant forAfrican-American men, who suffer disproportion-

ately from prostate cancer. Thus, African-Americanmen might especially benefit from increased vitaminE intake.




TABLE 4. Unadjusted rank correlation coefficients between serum a-tocopherol concentration and a-tocopherol/cholesterolratio and various covariates among men aged 218 years, by sex and race, National Health and Nutrition Examination Survey III,1988-1991

AgeEducationSerum cot nineSystolic Hood pressureSerum cholesterolHDL chotesteroltBody mass indexLeisure-time physical activityAlcohol useFruit and vegetable IntakeDtetary vitamin ESerum MateRed Hood cell MateVitamin B12Vitamin AVitamin Ca-Carotenep-CaroteneCryptoxanthlnLutein/zeaxanthlnLycopeneVitamin use|

YesNo

Fruit and vegetableconsumption per month}:

<150S150

White

a-Tocopherol

0.350.10

-0.200.220.64

-0.140.240.07

-0.030.170.050.380.350.080.480.250.330.370.330.440.16

33.8 (0.7)*24.8 (0.3)

26.1 (0.4)*28.7 (0.5)

a-Tocopherol/cholesterol

0.230.15

-0.240.100.03

-0.170.130.15

-0.040.200.120.450.430.120.360.370.290.300.280.300.01

6.44 (0.12)*4.75 (0.05)

4.96 (0.07)*5.48 (0.09)

African American

a-Tocopherol

0.330.08

-0.130.200.64

-0.030.23

-0.04-0.03

0.140.010.260530.110.380.200.350.380.300.43017

28.3 (0.9)*21.3(0.2)

21.9(0.3)*24.0 (0.5)

a-TocopheroVcholesterol

0.130.14

-0.130.01

-0.12-0.090.040.08

-0.030.190.090.320.270.130.160.330.290.230.250.220.03

5.62 (0.15)*4.21 (0.03)

4.33 (0.05)*4.71 (0.07)

Mexican American

a-Tocopherol

0.430.01

-0.130.180.69

-0.190.28

-0.030.010.090.010.210.220.100.400.120.290.250560.390.16

31.7(1.0)*23.8 (0.2)

24.4 (0.4)25.3 (0.4)


0.240.02

-0.130.070.04

-0530.120.00

-0.040.100.070.250.240.080530520530.200.200.31

-0.03

5.94(0.12)*4.68 (0.04)

4.74 (0.06)4.91 (0.06)

Other

a-Tocopnerol

0.480.08

-0.290.250.73

-0.05053

-0.02-0.01

0500.080.410.380.290.450.240.390.420.310.390.11

32.5 (2.6)'23.7 (0.7)

24.7(1.1)26.4 (0.8)

a-TocopheroVcholesterol

0580.09

-0.340.110.13

-0.150.080.03

-0.040540.050.450.400.200.280.340.280.260540550.03

6.32 (0.52)*4.65(0.11)

4.77 (0.18)t5.23(0.12)

• p < 0.05.t HOL cholesterol, high density llpoprotein cholesterol.t Mean umoles/iiter (standard error) for serum a-tocopherol concentration and mean (standard error) for a-tocopherol/cholesterol.

What accounts for this racial gap in serum a-tocopherol concentration is unclear. The crude andadjusted odds ratios for a low serum a-tocopherol con-centration were very similar. Thus, the covariates thatwe examined did not account for the difference.Difference in dietary intake of vitamin E, which we didnot include for reasons outlined below, may accountfor much of the difference in serum concentration. Infact, the mean vitamin E consumption estimated fromthe 24-hour dietary recall is lower among AfricanAmericans than among Whites. Furthermore, the vita-min supplementation use variable we used may havebeen too crude to explain adequately the portion of theracial difference attributable to differences in vitaminsupplement use.

Population-based studies of vitamin E concentra-tions are few. The median concentrations of vitamin Efor 12 Monitoring Determinants and Trends ofCardiovascular Disease (MONICA) centers that par-ticipated in a vitamin substudy ranged from about 20to 30 |i.mol/liter (27). Particularly high concentrationswere observed at sites in Spain and Switzerland. Inanother study, vitamin E concentrations were 8.6mg/liter (20 (Jjnol/liter) among 824 Finnish men and10.5 mg/liter (24.4 nmol/liter) among 549 Finnish

women aged 40-79 years (28). In comparison, themedian serum a-tocopherol concentration in theUnited States was 24.1 (imol/liter. However, theMONICA populations were older and were not neces-sarily representative of each country's population.

The strongest predictors of serum concentration ofa-tocopherol in the NHANES III data were age, cho-lesterol, and various serum vitamin and carotenoidconcentrations. Previously, strong relations have beenobserved for cholesterol (20, 21, 28-37) and triglyc-erides (20, 21, 33, 35-38) . Weaker, but significant,relations have been observed for high density lipopro-tein cholesterol (34, 38, 39). Significant relations havealso been described between serum or plasma vitaminE concentrations and age (28, 36, 37, 40-42), sex (28,38, 43-46), body mass index (34, 35, 41, 47), waistcircumference (46), supplemental vitamin use (36, 38,41, 48), physical activity (37) , alcohol consumption(38) , blood pressure (28), serum retinol (28, 36, 49),and f}-carotene (28).

Other studies have failed to find significant relationsbetween blood concentrations of vitamin E and age(20, 33, 37, 45, 50, 51), sex (36, 41, 52), body massindex (20, 21, 28, 33, 34, 36, 38, 50), waist-hip ratio(21), high density lipoprotein cholesterol (20, 21, 34,



296 Ford and Sowell

TABLE 5. Unad|usted rank correlation coefficients between serum a-tocopherol concentration and a-tocopherol/cholesterolratio and various covarlates among women aged £18 years, by sex and race, National Health and Nutrition Examination Survey III,1988-1991

AgeEducationSerum cotinineSystolic blood pressureSerum cholesterolHDL cholesteroltBody mass IndexLeisure-time physical activityAlcohol useFrutt and vegetable IntakeDietary vitamin ESerum folateRed blood cell folateVitamin B12Vitamin AVitamin Ca-CaroteneP-CaroteneCryptoxanthlnLuteln/zeaxanthinLycopeneVitamin usef

YesNo

Fruit and vegetableconsumption per month*

<150£150

Oral contraceptive use amongwomen aged <45 years}:

YesNo

Hormone replacement therapyamong women aged £50years*

YesNo

White

ot-Tocopheroi

0.510.00

-0.280.360.640.080.170.06

-0.130.250.020.470.450.240.440.380.360.430.370.400.06

35.7 (0.6)*24.8 (0.3)

26.7 (0.4)*31.0(0.5)

23.2 (0.6)23.8 (0.3)

35.2(1.5)*33.7 (0.4)


0.290.13

-0.270.150.050.08

-0.050.160.010.250.020.540.480.290.280.450.360.400.31029

-0.03

6.49 (0.09)*4.67 (0.04)

5.03 (0.05)'5 64 (0.08)

4.74 (0.07)4.88 (0.06)

6.64 (0.22)'5.59 (0.07)

African American

cc-Tocopherol

0.46-0.03-0.12

0590.680.120.07

-0.05-0.13

0.15-0.01

0.340.320.210.340.190.390.420.340.440.11

28.2 (0.5)*21.5 (0.2)

22.3 (0.3)*24.1 (0.4)

21.0(0.6)20.5 (0.2)

31.0(1.8)27.9 (0.5)


0.170.10

-0.150.08

-0.10-0.01-0.05

0.01-0.06

0.140.090.360.350.150.130.320.280.230.240.210.01

5.33(0.10)'4.20 (0.03)

4.35 (0.04)*4.60 (0.06)

4.27 (0 10)4.29 (0.03)

5.23 (0.22)4.75 (0.08)

Mexican American

o-Tocopherol

0.40-0.11-0.16

0570.700.120.140.01

-0.070.140.020.300.260.080.360.220.320 270.300.350.01

30.1 (0.6)*23.5 (0.2)

23.7 (0.3)*25.7 (0.4)

22.2 (0.4)*23.2 (0.2)

34.6(1.7)*30.0 (0.6)


0.19-0.06-0.17

0.070.010.01

-0.010.04

-0.070.140.090.360.300.130.140.310.320.270.300.35

-0.01

5.71 (0.06)*4.66 (0.04)

4.73 (0.04)*4.97 (0.04)

4.53 (0.05)*4.78 (0.04)

5.76 (0.23)*5.16(0.08)

Other

a-Tocopheroi

0.40-0.05-0.12

0.340.70

-0.040.140.050.020.04

-0.010.200.260.130.380.090.180550.240.350.23

34.9 (2.8)*23.4 (0.5)

25.3 (0.7)25.8(1.0)

22.7(1.3)23.5 (1.0)

32.7(1.6)30.3(2.1)


0.130.05

-0.150.08

-0.04-0.11-0.09

0.030.070.120.040.180.210.120.160520.120.110130.160.08

6.40 (0.44)*4.59 (0.07)

4.82(0.11)5.07(0.18)

4.71 (0.11)4.90(0.17)

556 (0.08)5.07 (0.24)

• p < 0.05.t HDL cholesterol, high density lipoproteln cholesterol.t Mean nmotes/iiter (standard error) for serum a-tocopherol concentration and mean (standard error) for a-tocopherol/cholesterol.

37) , and alcohol use (32, 33, 41, 45, 50). Adipose a-tocopherol showed significant inverse correlationswith body mass index and waist-hip ratio but no sig-nificant associations with age, smoking status, andalcohol consumption, and y-tocopherol was only sig-nificantly associated with waist-hip ratio (46).

Blood vitamin E concentration has been found tocorrelate positively with dietary intake of vitamin E orfoods rich in vitamin E (28, 33, 35, 36, 38, 50, 53).Other studies have found dietary intake to be poor pre-dictors of vitamin E concentration (36, 41, 54). Mostdietary studies used food frequency questionnaires andfound reasonable correlations between diet and bloodconcentrations of vitamin E. The results of the single24-hour dietary recall did not correlate well withserum concentrations in NHANES HI, and evenaccounting for attenuation of the correlation coeffi-cient due to misclassification would not alter the inter-pretation much. The food frequency questionnaire of

NHANES El, unfortunately, included a relatively shortlist of foods and did not include portion sizes for thefoods. This made it difficult to estimate accurately theamount and type of fats consumed that influence vita-min E intake. Our food frequency scales for fruits andvegetables and for cereal were not significant predic-tors of vitamin E concentrations. Decreasing blood a-tocopherol concentration with increased energy intakehas also been described (33, 41).

Although some studies have reported a significantassociation between blood vitamin E concentration andsmoking status (34, 35, 38, 55, 56), most have failed tofind one (20, 32, 34, 37, 41, 45, 50, 52, 57-62). In ourunivariate analysis, current smokers had lower serum a-tocopherol concentrations than did nonsmokers. Afteradjustment for the covariates, however, the differencewas eliminated, which suggests that some of the covari-ates explain the association between smoking status andserum a-tocopherol concentration. Vitamin E concen-




TABLE 6. Linear regression of various predictors on log-transformed serum a-tocopherol concentration, National Hearth andNutrition Examination Survey III, 1988-1994

Independent vanabte

Age (years)

Sex

Race or ethnlcttyOtherMexican AmericanAfrican AmericanWhite (reference)

Education (years)

Smoking statusCurrentFormerNever (reference)

Serum cholesterol(mmol/llter)

Serum HDL cholesterol*(mmoMlter)

Body mass Index (kg/m*)

History of diabetes

Vitamin/mineral use Inprevious 24 hours

Serum totate (nmol/liter)

Serum vitamin A (umol/liter)

Serum vitamin C (umol/liter)

Serum a-caroteoe Oimol/lrter)

Serum p-carotene (junol/liter)

Serum p-cryptoxanttiln(umol/llter)

Serum luteln/zeaxanthln(umol/liter)

Serum lycopene (umot/lrter)

Multiple FP

P

0.0024

-0.0254

-0.00670.0288

-0.0770

0.0058

0.00160.0168

0.1419

-0.0844

0.0007

0.0123

0.1687

0.0034

0.0947

0.0011

-0.0873

0.0798

0.0645

0.1518

0.0583

Total (n= 15,082)

SE*

0.0002

0.0067

0.01500.00950.0068

0.0009

0.00530.0074

0.0044

0.0101

0.0006

00116

0.0134

0.0006

0.0066

0.0002

0.0467

0.0150

0.0315

0.0210

0.0147

0.64

Pvalue

<0.001

<0.001

0.6580.004

<0.001

<0.001

0.7640.028

<0.001

<0.001

0.243

0.293

<0.001

<0.001

<0.001

<0.001

0.067

<0.001

0.046

<0 001

<0.001

P

0.0016

-0.02330.0172

-0.0644

0.0042

0.00680 0197

0.1522

-0.1308

0.0029

-0.0044

01439

0.0052

0.1073

0.0009

-0.0580

0.0722

0.0812

0.1630

0.0563

Men (n = 7,059)

SE

0.0003

0.01720.01500.0107

0.0012

0.00820.0138

0.0057

0 0149

0.0009

0.0170

0.0193

0.0009

0.0094

0.0002

0.0652

0 0275

0.0525

0.0315

0.0246

0.65

Pvalue

<0.001

0.2020.258

<0.001

0.002

0.4140.158

<0.001

<0.001

0.003

0.800

<0.001

<0.001

<0.001

<0.001

0.378

0.011

0.128

<0.001

0.027

P

0.0031

0.01310.0423

-0.0787

0.0068

0.00100.0134

01324

-0.0458

-0.0003

0.0296

01851

00026

0.0837

0.0011

-0.0996

0.0831

0.0433

01285

0.0505

Women (n = 8,023)

SE

0.0003

0.01920.01010.0079

0.0014

0 00810.0075

0.0048

0.0112

0.0007

0.0123

0.0137

0.0006

0.0075

0.0003

0.0587

0.0161

0.0354

0.0239

0.0163

0.65

Pvalue

<0.001

0.499<0.001<0.001

<0.001

0.9010.078

<0.001

<0.001

0.632

0.020

<0.001

<0.001

<0.001

<0.001

0.096

<0.001

0.228

<0.001

0.003

1 SE, standard error; HDL cholesterol, high density llpoprotein cholesterol.

tration has been found to be higher among former smok-ers than among current or never smokers (20). Theresults from our analysis agree with this finding.

Several researchers have examined correlates ofserum or plasma vitamin E separately by gender (28,33,50). In the Finnish study, body mass index and use ofvitamin supplements predicted serum oc-tocopherolssimilarly among men and women (28). Age, smokingstatus, blood pressure, serum retinol, and several othervariables predicted serum a-tocopherol concentrationsdifferently in men and women. In a study of 330 maleand female melanoma study controls aged 18-79 years,no sex differences in the relations of vitamin E intake,plasma cholesterol and triglycerides, and body massindex to plasma a-tocopherol concentration wereobserved (33). However, the relations between plasmaa-tocopherol concentration and energy intake, age, cig-arette use, and alcohol use differed between men andwomen. In another study of 121 men and 186 women,

the relations between plasma a-tocopherol concentra-tion and age, plasma cholesterol and triglyceride con-centrations, body mass index, energy intake, and vita-min E intake were similar between the sexes (50). In ouranalysis of the NHANES IE data, we found the relationsbetween serum a-tocopherol and various predictors tobe fairly similar between men and women except forHDL cholesterol, serum cholesterol concentration, bodymass index, serum folate concentration, serum vitaminA concentration, and serum lycopene concentration.

We were able to explain about 65 percent of the totalvariance in serum tocopherol concentration. Otherstudies have reported an R1 or an adjusted R2 rangingfrom 48 to 73 percent (33, 36, 38).

The lower limit of normal for vitamin E concentra-tion ranges is 11.15-17.84 |imol/liter (0.5-0.8 mg/dl)(63). These lower limits are based on more acute man-ifestations of vitamin E deficiency (e.g., shortened redblood cell life span, dermatitis, and peroxide hemoly-



298 Ford and Sowell

sis) and not on studies of prevention of chronic dis-ease. We chose a serum oc-tocopherol cutpoint of 20(xmol/liter because the literature suggests an increasedrisk for cardiovascular disease below this concentra-tion (20, 21). An ecologic study from the MONICAproject suggested that the relation between ischemicheart disease mortality and vitamin E concentrationmay be reasonably linear over a range of 20-30(imol/liter (27). The authors suggested that a lipid stan-dardized concentration in excess of more than 27.5-30(imol lipid standardized vitamin E/liter, correspondingto an unadjusted vitamin E concentration of about 29Hmol/liter, would be needed to reduce the risk forischemic heart disease (27).

The Alpha-Tocopherol, Beta-Carotene CancerPrevention Trial failed to demonstrate a benefit ofadministrating 50 mg of vitamin E daily on coronaryheart disease mortality (64), an intake generally belowthat suggested to be beneficial by epidemiologic stud-ies. The Cambridge Heart Antioxidant Study showed areduction of primarily nonfatal myocardial infarctionamong participants taking 400-800 IU per day (65).Therefore, daily doses of more than 50 mg of vitaminE may be necessary to reduce the risk for ischemicheart disease.

Vitamin E is thought to prevent ischemic heart dis-ease by increasing the resistance of low densitylipoprotein to oxidative modification (66). Oxidizedlow density lipoprotein displays several deleteriouseffects that contribute to the formation of atheroscle-rotic plaques. These include being chemotactic tomonocytes, promoting binding of monocytes to thevascular endothelium, inhibiting macrophage motility,and being cytotoxic to endothelial cells (1). In addi-tion, vitamin E improves vascular function, inhibitsleukocyte adhesion to endothelial cells, reduces cellu-lar oxidative injury, and inhibits platelet activation andsmooth muscle proliferation (1). Various antioxidants,including vitamin E, have been shown to limit oxida-tion of low density lipoprotein (67-73).

The pathways through which vitamin E may reducethe risk for various cancers are less well understood.The antioxidant properties of vitamin E are thought toprotect against damage by free radicals, such as thosegenerated by lipid peroxidation. Vitamin E may be par-ticularly important in a population with high intakes ofpolyunsaturated fatty acids for whom the potential forlipid peroxidation may be especially high (28). Inaddition, vitamin E may limit the formation of N-nitroso compounds in the stomach (74) and protectsthe antioxidant activity of selenium.

Countries with low rates of cardiovascular disease(e.g., Switzerland, Spain, France, and Italy) were alsoamong the countries that had the highest medians of

vitamin E concentration in the MONICA study.Compared with the MONICA populations, the mean ormedian serum oc-tocopherol concentration of the USpopulation would rank roughly in the middle, corre-sponding roughly also to its international ranking forcardiovascular disease (75). This leads one to specu-late about whether efforts to increase serum vitamin Econcentrations in the United States through increasedintake of vitamin E could have a favorable impact oncardiovascular disease mortality. Since 1968, coronaryheart disease mortality has decreased in the UnitedStates. However, the rate of decline may have slowedin recent years for reasons that are not clear (24). If thelink between vitamin E intake or physiologic concen-trations of vitamin E and cardiovascular disease mor-bidity and mortality can be further substantiated, pub-lic health programs to promote the intake of vitamin Ecould help further reduce the death toll from cardio-vascular disease.

In summary, we describe serum a-tocopherol statusin the United States. Perhaps the most significant find-ing of our study is that African Americans, who havehigh rates of cardiovascular disease and cancer mortal-ity, had the lowest concentrations of oc-tocopherol ofthe major race-ethnic groups we studied. Thus, waysmay need to be found to increase the intake of vitaminE by African Americans. Hopefully, studies currentlyunderway will be able to definitively demonstrate theutility of vitamin E in reducing chronic disease, partic-ularly cardiovascular disease. Should vitamin E proveto be of major benefit in lowering the risk for cardio-vascular disease and other conditions, the optimalintake and safety profile of this vitamin will need to beestablished. With better understanding of the role ofvitamin E in the pathogenesis of various chronic con-ditions, an increased recommended daily allowancefor vitamin E may need to be considered. In addition,optimal ranges for concentrations of serum vitamin Ethat currently range in excess of 11.6-18.6 (imol/litermay need to be revised in the future.

REFERENCES

1. Diaz MN, Frei B, Vita JA, et al. Antioxidants and atheroscle-rotic heart disease. N Engl J Med 1997;337:4O8-16.

2. Byers T, Perry G. Dietary carotenes, vitamin C, and vitamin Eas protective antioxidants in human cancers. Annu Rev Nutr1992; 12:139-59.

3. Salonen JT, Nyyssdnen K, Tuomainen T, et al. Increased riskof non-insulin dependent diabetes mellitus at low plasma vita-min E concentrations: a four year follow up study in men. BMJ1995;311:1124-7.

4. Kubena KS, McMurray DN. Nutrition and the immune sys-tem: a review of nutrient-nutrient interactions. J Am DietAssoc 1996;96:1156-64.

5. Liang B, Chung S, Araghiniknam M, et al. Vitamins and




immunomodulation in AIDS. Nutrition 1996;12:l-7.6. Dekkers JC, van Doornen LJ, Kemper HC. The role of antiox-

idant vitamins and enzymes in the prevention of exercise-induced muscle damage. Sports Med 1996,21:213-38.

7. Nachbar F, Korting HC. The role of vitamin E in normal anddamaged skin. J Mol Med 1995;73:7-17.

8. Martin A. The use of antioxidants in healing. Dennatol Surg1996;22:156^60.

9. Ebadi M, Srinivasan SK, Baxi MD. Oxidative stress andantioxidant therapy in Parkinson's disease. Prog Neurobiol1996;48:1-19.

10. Cadet JL. Movement disorders: therapeutic role of vitamin E.Toxicol Ind Health 1993;9:337^t7.

11. Tanyel MC, Mancano LD. Neurologic findings in vitamin Edeficiency. Am Fam Physician 1997;55:197-201.

12. Snodderly DM. Evidence for protection against age-relatedmacular degeneration by carotenoids and antioxidant vitamins.Am J Clin Nutr 1995;62 (Suppl. 6):1448S-61S.

13. Varma SD, Devamanoharan PS, Morris SM. Prevention ofcataracts by nutritional and metabolic antioxidants. Crit RevFood Sci Nutr 1995;35:lll-29.

14. Hatch GE. Asthma, inhaled oxidants, and dietary antioxidants.Am J Clin Nutr 1995;61 (Suppl. 3):625S-30S.

15. Britton RS, Bacon BR. Role of free radicals in liver diseasesand hepatic fibrosis. Hepatogastroenterology 1994;41:343-8.

16. McCarley DL. New uses for old vitamins. The treatment ofmyelodysplastic disorders. Med Clin North Am 1993;77:919-29.

17. Centers for Disease Control and Prevention. Plan and opera-tion of the Third National Health and Nutrition ExaminationSurvey, 1988-1994. Hyattsville, MD: National Center forHealth Statistics, 1994. (Vital and Health Statistics, Series 1,no. 32)

18. Sowell AL, Huff DL, Yeager PR, et al. Retinol, a-tocopherol,lutein/zeaxanthin, P-cryptoxanthin, lycopene, a-carotene,trans-P-carotene, and four retinyl esters in serum determinedsimultaneously by reversed-phase HPLC with multiwave-length detection. Clin Chem 1994;40:411-16.

19. Gunter EW, Lewis SM, Koncikowski SM. Laboratory proce-dures used for the Third National Health and NutritionExamination Survey (NHANES m), 1988-1994. Atlanta, GA:Centers for Disease Control and Prevention, 1996.

20. Riemersma RA, Wood DA, Macintyre CCA, et al. Risk ofangina pectoris and plasma concentrations of vitamins A, C,and E and carotene. Lancet 1991;337:l-5.

21. Singh RB, Ghosh S, Niaz MA, et al. Dietary intake, plasmalevels of antioxidant vitamins, and oxidative stress in relationto coronary artery disease in elderly subjects. Am J Cardiol1995;76:1233-8.

22. Shah BV, Barnwell BG, Bieler GS. SUDAAN user's manual,version 7.5. Research Triangle Park, NC: Research TriangleInstitute, 1997.

23. Willett WC, Polk BF, Underwood BA, et al. Relation of serumvitamins A and E and carotenoids to the risk of cancer. N EnglJMed 1984;310:430-4.

24. National Heart, Lung, and Blood Institute. Morbidity & mor-tality: 1996 chartbook on cardiovascular, lung, and blood dis-eases. Bethesda, MD: National Heart, Lung, and BloodInstitute, 1996.

25. National Cancer Institute. SEER cancer statistics review,1973-1994. Bethesda, MD: National Cancer Institute, 1997.(NIH publication no. 97-2789).

26. The Alpha-Tocopherol, Beta-Carotene Cancer PreventionStudy Group. The effect of vitamin E and beta carotene on theincidence of lung cancer and other cancers in male smokers. NEngl J Med 1994;330:1029-35.

27. Gey KF, Moser UK, Jordan P, et al. Increased risk of cardio-vascular disease at suboptimal concentrations of essentialantioxidants: an epidemiologic update with special attention tocarotene and vitamin C. Am J Clin Nutr 1993;57 (Suppl.):787S-97S.

28. Knekt P, Seppanen R, Aaran RK. Determinants of serum

alpha-tocopherol in Finnish adults. Prev Med 1988; 17:725-35.29. Desai I, Lee M. Plasma vitamin E and cholesterol relationship

in Western Canadian Indians. Am J Clin Nutr 1974;27:334-8.30. Yang NY, Desai ID. Effect of high levels of vitamin E on liver

and plasma lipids and fat soluble vitamins in rats. J Nutr1977;107:1418-26.

31. Vessby B, Lithell H, Boberg J. Supplementation with vitaminE in hyperlipidemic patients treated with diet and clofibrate.Effects on serum lipoprotein concentrations, plasma fatty acidcomposition and adipose tissue lipoprotein lipase activity. AmJ Clin Nutr 1977;30:517-22.

32. Stahelin HB, ROsel F, Buess E, et al. Cancer, vitamins, andplasma lipids: Prospective Basel Study. J Natl Cancer Inst1984;73:1463-8.

33. Stryker WS, Kaplan LA, Stein EA, et al. The relation of diet,cigarette smoking, and alcohol consumption to plasma beta-carotene and alpha-tocopherol levels. Am J Epidemiol 1988;127:283-96.

34. Riemersma RA, Oliver M, Elton RA, et al. Plasma antioxi-dants and coronary heart disease: vitamins C and E, and sele-nium. Eur J Clin Nutr 1990;44:143-50.

35. Bolton-Smith C, Casey CE, Gey KF, et al. Antioxidant vitaminintakes assessed using a food-frequency questionnaire: corre-lation with biochemical status in smokers and nonsmokers. BrJNutrl991;65:337^6.

36. Hebert JR, Hurley TG, Hsieh J, et al. Determinants of plasmavitamins and lipids: the Working Well Study. Am J Epidemiol1994;140:132-47.

37. Kitamura Y, Tanaka K, Kiyohara C, et al. Relationship of alco-hol use, physical activity and dietary habits with serumcarotenoids, retinol and alpha-tocopherol among maleJapanese smokers. Int J Epidemiol 1997;26:307-14.

38. Herbeth B, Chavance M, Musse N, et al. Dietary intake andother determinants of blood vitamins in an elderly population.Eur J Clin Nutr 1989;43:175-86.

39. Salonen JT, Salonen R, Seppanen K, et al. Relationshipbetween serum selenium and antioxidants to plasma lipopro-teins, platelet aggregability and prevalent ischemic heart dis-ease in Eastern Finnish men. Atherosclerosis 1988;70:155-60.

40. Wald NJ, Boreham J, Hayward JL, et al. Plasma retinol, p-carotene and vitamin E levels in relation to, the future risk ofbreast cancer. Br J Cancer 1984;49:321^.

41. Goodman GE, Thomquist M, Kestin M, et al. The associationbetween participant characteristics and serum concentrationsof beta-carotene, retinol, retinyl palmitate, and alpha-toco-pherol among participants in the Carotene and RetinolEfficacy Trial (CARET) for prevention of lung cancer. CancerEpidemiol Biomarkers Prev 1996;5:815-21.

42. Winklhofer-Roob BM, van't Hof MA, Shmerling DH.Reference values for plasma concentrations of vitamin E andA and carotenoids in a Swiss population from infancy to adult-hood, adjusted for seasonal influences. Clin Chem 1997;43:146-53.

43. Lehmann J, Marsall MW, Slover HT, et al. Influence of dietaryfat level and dietary tocopherols on plasma tocopherols onhuman subjects. J Nutr 1977;107:1006-15.

44. Behrens WA, Thompson JN, Mad6 R. Distribution of alpha-tocopherol in human plasma lipoproteins. Am J Clin Nutr1982;35:691-6.

45. Buiatti E, Munoz N, Kato I, et al. Determinants of plasma anti-oxidant vitamin levels in a population at high risk for stomachcancer. Int J Cancer 1996;65:317-22.

46. Virtanen SM, van't Veer P, Kok F, et al. Predictors of adiposetissue tocopherol and toenail selenium levels in nine countries:the EURAMIC study. European Multicentre Case-ControlStudy on Antioxidants, Myocardial Infarction, and Cancer ofthe Breast. Eur J Clin Nutr 1996;50:599-606.

47. Sinha R, Patterson BH, Mangels AR, et al. Determinants ofplasma vitamin E in healthy males. Cancer EpidemiolBiomarkers Prev 1993;2:473-9.

48. Willett WC, Stampfer MJ, Underwood BA, et al. Vitamins A,



300 Ford and Sowell

E, and carotene: effects of supplementation on their plasmalevels. Am J Clin Nutr 1983;38:559-66.

49. Salonen JT, Salonen R, Penttiia I, et al. Serum fatty acids,apolipoproteins, selenium and vitamin antioxidants and therisk of death from coronary artery disease. Am J Cardiol1985;56:226-31.

50. Asherio A, Stampfer Ml, Colditz GA, et al. Correlations ofvitamin A and E intakes with the plasma concentrations ofcarotenoids and tocopherols among American men andwomen. J Nutr 1992;122:1792-801.

51. Olivieri O, Stanzial AM, Girelli D, et al. Selenium status, fattyacids, vitamins A and E, and aging: the Nove Study. Am J ClinNutr 1994;60:510-17.

52. Miyamoto H, Araya Y, Ito M, et al. Serum selenium and vita-min E concentrations in families of lung cancer patients.Cancer 1987;60:1159-62.

53. Romieu I, Stampfer MJ, Stryker WS, et al. Food predictors ofplasma beta-carotene and alpha-tocopherol: validation of afood frequency questionnaire. Am J Epidemiol 1990;131:864-76.

54. Marshall JR, Lanza E, Bloch A, et al. Indexes of food andnutrient intakes as predictors of serum concentrations of nutri-ents: the problem of inadequate discriminant validity. Am JClin Nutr 1997;65 (Suppl):1269S-74S.

55. Fernandez-Banares F, Gine JJ, Cabre E, et al. Factors associ-ated with low values of biochemical vitamin parameters inhealthy subjects. Int J Vitam Nutr Res 1994;64:68-74.

56. Munro LH, Burton G, Kelly FJ. Plasma RRR-tocopherol con-centrations are lower in smokers than in non-smokers afteringestion of a similar oral load of this antioxidant vitamin.Clin Sci 1997;92:87-93.

57. Chow KK, Thacker RR, Changchit C, et al. Lower levels ofvitamin C and carotenes in plasma of cigarette smokers. J AmColl Nutr 1986;5:305-12.

58. Comstock GW, Menkes MS, Schober SE, et al. Serum levelsof retinol, beta-carotene, and alpha-tocopherol in older adults.Am J Epidemiol 1988; 127:114-23.

59. Looker AC, Underwood BA, Wiley J, et al. Serum a-tocopherol levels of Mexican Americans, Cubans, and PuertoRicans aged 4-74 y. Am J Clin Nutr 1989;50:491-6.

60. Ito Y, Shima Y, Ochiai J, et al. Effects of the consumption ofcigarettes, alcohol and foods on serum concentrations ofcarotenoids, retinol, and tocopherols in healthy inhabitantsliving in rural area of Hokkaido. Jpn J Hyg 1991;46:874-81.

61. Sanders TAB, Haines AP, Wormald R, et al. Essential fattyacids, plasma cholesterol, and fat-soluble vitamins in subjectswith age-related maculopathy and matched control subjects.

Am J Clin Nutr 1993;57:428-33.62. Marangon K, Herbeth B, Lecomte E, et al. Diet, antioxidant

status, and smoking habits in French men. Am J Clin Nutr1998;67:231-9.

63. Berkow R, Fletcher AJ, eds. The Merck manual. 16th ed.Rahway, NJ: Merck Research Laboratories, 1992.

64. Rapola JM, Virtamo J, Ripatti S, et al. Randomised trial of a-tocopherol and P-carotene supplements on incidence of majorcoronary events in men with previous myocardial infarction.Lancet 1997;349:1715-20.

65. Stephens NG, Parsons A, Schofield PM, et al. Randomisedcontrolled trial of vitamin E in patients with coronary disease:Cambridge Heart Antioxidant Study (CHAOS). Lancet1996:347:781-6.

66. Steinberg D, Parthasarathy S, Carew TE, et al. Beyond cho-lesterol: modifications of low-density lipoprotein that increaseits atherogenicity. N Engl J Med 1989;320:915-24.

67. Dieber-Rotheneder M, Puhl H, Waeg G, et al. Effect of oralsupplementation with D-a-tocopherol on the vitamin E con-tent of human low density lipoproteins and resistance to oxi-dation. J Lipid Res 1991;32:1325-32.

68. Jialal L Grundy S. The effect of dietary supplementation withalpha-tocopherol on the oxidative modification of low densitylipoprotein. J Lipid Res 1992:33:899-906.

69. Princen HM, van Poppel G, Vogelezang C, et al.Supplementation of vitamin E but not beta-carotene in vivoprotects low-density lipoprotein from lipid peroxidation invitro. Arterioscler Thromb Vase Biol 1992; 12:554-62.

70. Kearney JF Jr, Gaziano JM, Xu A, et al. Dietary anti-oxidants preserve endothelium-dependent vessel relaxationin cholesterol-fed rabbits. Proc Natl Acad Sci U S A 1993;90:11880-4.

71. Reaven PD, Khouw A, Beltz WF, et al. Effect of dietaryantioxidant combinations in humans. Arterioscler ThrombVase Biol 1993;13:59O-6OO.

72. Jialal I, Grundy SM. Effect of combined supplementation witha-tocopherol, ascorbate, and beta carotene on low-densitylipoprotein oxidation. Circulation 1993;88:2780-6.

73. Steward-Lee AL, Forster LA, Nourooz-Zadeh J, et al. VitaminE protects against impairment of endothelium-mediated relax-ations in cholesterol-fed rabbits. Arterioscler Thromb1994; 14:494-9.

74. Mirvish SS. Inhibition by vitamins C and E of in vivo nitrosa-tion and vitamin C occurrence in the stomach. Eur J CancerPrevl996;l (Suppl. 5): 131-6.

75. American Heart Association. Cardiovascular diseases. Dallas,TX: American Heart Association, 1998.



Documents

Serum a-Tocopherol Status in the United States Population