Human Nutrition and Metabolism

2 downloads 15 Views 282KB Size Report
in Serum Carotenoids and Fat-Soluble Vitamins: Baseline Results from the ... determinants of serum carotenoid and fat-soluble vitamin concentrations in a large, ...
Human Nutrition and Metabolism

Demographic, Dietary and Lifestyle Factors Differentially Explain Variability in Serum Carotenoids and Fat-Soluble Vitamins: Baseline Results from the Sentinel Site of the Olestra Post-Marketing Surveillance Study1 Cheryl L. Rock,*2 Mark D. Thornquist,† Alan R. Kristal,† Ruth E. Patterson,† Dale A. Cooper,‡ Marian L. Neuhouser,† Dianne Neumark-Sztainer** and Lawrence J. Cheskin†† *Department of Family and Preventive Medicine, University of California, San Diego, La Jolla, CA 92093-0901; †Department of Epidemiology and Department of Biostatistics, University of Washington, and Fred Hutchinson Cancer Research Center, Seattle, WA 98104; **Division of Epidemiology, University of Minnesota, Minneapolis, MN 55454; ‡The Procter & Gamble Company, Cincinnati, OH 45224; and ††Division of Gastroenterology, Johns Hopkins School of Medicine, Baltimore, MD 21224.




25-hydroxyvitamin D

Olestra, sucrose esterified with medium- and long-chain fatty acids, is a new product in the U.S. food supply, which is approved for use as a fat replacement in savory snacks such as potato chips and crackers (Federal Register 1996). Olestra can reduce the absorption of carotenoids and fat-soluble vitamins when co-consumption occurs (Jones et al. 1991a and 1991b, Koonsvitsky et al. 1997, Westrate and van het Hof 1995), although an effect on vitamin status is offset by adding these



Downloaded from by guest on June 12, 2013

ABSTRACT Biochemical measures of nutrients or other dietary constituents can be an important component of nutritional assessment and monitoring. However, accurate interpretation of the nutrient concentration is dependent on knowledge of the determinants of the body pool measured. The purpose of this study was to identify the determinants of serum carotenoid and fat-soluble vitamin concentrations in a large, community-based sample (n 5 1042). Multiple linear regression analysis was used to examine effects of demographic characteristics (age, sex, race/ethnicity, education), health-related behavior (exercise, sun exposure, smoking, alcohol consumption), and intake (diet, supplements) on serum retinol, 25-hydroxyvitamin D, a-tocopherol, phylloquinone, and carotenoid concentrations. Age, sex, race/ethnicity, vitamin A intake, and alcohol consumption were found to be determinants of serum retinol concentration. Race/ethnicity, vitamin D intake, body mass index, smoking status, and sun exposure were determinants of serum 25-hydroxyvitamin D concentration. Determinants of serum a-tocopherol were age, sex, race/ethnicity, a-tocopherol intake, serum cholesterol, percentage of energy from fat (inversely related), supplement use, and body mass index. Age, sex, phylloquinone intake, serum triglycerides, and supplement use were determinants of serum phylloquinone concentration. Primary determinants of serum carotenoids were age, sex, race/ethnicity, carotenoid intake, serum cholesterol, alcohol consumption, body mass index, and smoking status. Overall, the demographic, dietary, and other lifestyle factors explained little of the variability in serum concentrations of retinol (R2 5 0.20), 25-hydroxyvitamin D (R2 5 0.24), and the carotenoids (R2 5 0.15– 0.26); only modest amounts of the variability in serum phylloquinone concentration (R2 5 0.40); and more substantial amounts of the variability in serum a-tocopherol concentration (R2 5 0.62). J. Nutr. 129: 855– 864, 1999. humans

compounds to olestra-containing food products (Prince and Welschenbach 1998, Schlagheck et al. 1997). Fortification with carotenoids has not been recommended. As part of the approval process, the U. S. Food and Drug Administration (FDA)3 mandated further study of the consumption patterns and the effects of olestra on nutritional status associated with the introduction of olestra-containing foods into the U.S. food supply (Kristal et al. 1998). Thus, one aim of the Olestra

1 Presented in part at the Experimental Biology 96 Meeting, San Francisco, CA. [Rock, C. L., Kristal, A. R., Patterson, R. E., Neuhouser, M., NeumarkSztainer, D. & Thornquist, M. D. (1998) Determinants of serum carotenoids and fat-soluble vitamins: baseline results from the sentinel site of the Olestra PostMarketing Surveillance Study. FASEB J. 12: A544 (abs.)]. 2 To whom correspondence and reprint requests should be addressed.

3 Abbreviations used: BMI, body mass index; FDA, Food and Drug Administration; FFQ, food frequency questionnaire; HPLC, high performance liquid chromatography; NCI, National Cancer Institute; NIST, National Institute of Standards and Technology; OPMSS, Olestra Post-Marketing Surveillance Study; RIA, Radioimmunoassay; USDA, U.S. Department of Agriculture.

0022-3166/99 $3.00 © 1999 American Society for Nutritional Sciences. Manuscript received 13 October 1998. Initial review completed 18 November 1998. Revision accepted 22 December 1998. 855



MATERIALS AND METHODS The OPMSS consists of four field sites, an operations support unit at WESTAT in Rockville, MD, and a coordinating center at the Fred Hutchinson Cancer Research Center in Seattle, WA. Data for this analysis are from the Marion County (Indianapolis area), IN, site, termed the sentinel site, which recruited participants between August and December 1996, before the test-market release of olestra snacks in February 1997. Details on the design of the OPMSS and baseline results on diet and snack food consumption are provided elsewhere (Kristal et al. 1998). Subjects. Participants were adult volunteers identified through a random-digit-dial telephone survey to collect population-level data on the prevalence and patterns of savory snack use and fruit and vegetable consumption. At the completion of the telephone survey, a random sample of eligible participants were invited to join the clinical component of the study. To be eligible, participants had to be willing to visit the clinic site, complete questionnaires, provide a blood sample, and complete follow-up telephone calls. Those with a medical condition that would confound the measurement of associations between diet and serum carotenoids and fat-soluble vitamins (e.g., hemophilia, cystic fibrosis, kidney disease requiring dialysis), or

who had dietary restrictions that precluded eating savory snack foods, were excluded. Clinic visits were conducted between September 1996 and January 1997. Participants received $100 as compensation for travel expenses and personal time. Procedures for this study were approved by Institutional Review Boards of all of the organizations involved, and written informed consent was obtained from all subjects. Response rates to the telephone survey were similar to those obtained in other health-related random-digit-dial surveys (Kristal et al. 1993): the interview rate (completed interviews divided by known eligibles) was 89.2%, and the conservatively-estimated efficacy rate (completed interviews divided by known plus estimated eligibles) was 61.9%. The participation rate for the clinical component of the study reported here (clinic participants divided by invited eligible participants) was 56.9%. Procedures. Before the clinic visit, participants received detailed instructions on how to prepare for the visit, along with a food frequency questionnaire (FFQ) and snack food questionnaire. Participants were instructed to fast for 6 h preceding the visit and take no nutritional supplements on the day of the visit. Participants were asked to wear light clothing and to bring the completed FFQ and their vitamin supplements and medications to the clinic. At the clinic visit, interviewers administered questionnaires on demographic characteristics (age, sex, race/ethnicity, education) and health-related behavior in the previous month (smoking, alcohol consumption, physical activity, and sun exposure). Trained staff measured height, weight, and waist and hip circumferences, and collected information on dose and duration of all current supplements and medications. Nutritionists administered 24-h dietary recalls and reviewed FFQs for completeness. Phlebotomists collected blood samples into two 13-mL serum separating tubes, which were protected from light throughout processing and handling. Blood was allowed to clot and was then separated with refrigerated centrifugation at 2300 x g at 4°C for 10 min. Serum samples were briefly stored locally at -20°C and then shipped, packed with dry ice, twice per week to the coordinating center, where they were stored at -70°C until analysis. Dietary assessment. The self-administered FFQ, with a reference period of over the past month, consists of 122 food items, 19 items on food purchasing and preparation, and 4 items on the usual consumption of fruits, vegetables, and cooking and table fats. The snack questionnaire consists of 16 snack food items and a single item on usual consumption of all snack foods. Trained nutritionists administered 24-h recalls, following standardized protocols, which were later entered into the University of Minnesota’s Nutrient Data System (Minneapolis, MN) for analysis. Details of dietary assessment protocols are given by Kristal et al. (1998), and the FFQ and its measurement characteristics are described by Patterson et al. (1996 and 1999). Nutrient databases for both the FFQ and the 24-h recalls are primarily from the University of Minnesota Nutrition Coding Center nutrient database. This database uses the U.S. Department of Agriculture Nutrient Database for Standard Reference (1987) and its revisions as the primary data source, supplemented with information from the scientific literature and food manufacturers. Nutrient databases were augmented to include data for a-carotene, b-carotene, b-cryptoxanthin, lycopene, and lutein plus zeaxanthin from fruits and vegetables only, derived from the U.S. Department of Agriculture (USDA)-National Cancer Institute (NCI) carotenoid food composition database (Chug-Ahuja et al. 1993) and other published sources. Databases were also augmented with vitamin K (phylloquinone) content of foods, based on values derived from high performance liquid chromatography (HPLC) methods (Booth et al. 1993, Booth et al. 1996a and 1996b, Koivu et al. 1997). Servings per day of fruits and vegetables were calculated from the FFQ, following the approach used in the NCI 5-A-Day program (Havis et al. 1995, Krebs-Smith et al. 1995), as the sum of “fruit juices,” “a piece or serving of fruit (not counting juices),” “green salads,” “potatoes (not fried),” and “a serving of vegetables (not counting salads or potatoes).” With the exception of vitamin K, intakes of macronutrients, carotenoids, and vitamins were estimated from the FFQ. Vitamin K

Downloaded from by guest on June 12, 2013

Post-Marketing Surveillance Study (OPMSS) is to examine the influence of olestra consumption on serum concentrations of carotenoids and fat-soluble vitamins in a representative sample of the U.S. population, as an aspect of evaluating the effect of olestra on nutritional status. Biochemical measures of nutrients or other dietary constituents can be a valuable component of nutritional assessment and monitoring. However, many factors must be considered. For several micronutrients, the concentration of the nutrient in the circulating body pool (i.e., serum) appears to be a reasonably accurate reflection of overall status for the nutrient. In contrast, the amount of some micronutrients in the circulating pool may be homeostatically regulated when the storage pool is adequate, or may be unrelated to intake, and thus has little relationship with total body reserves or overall status. Accurate interpretation of the nutrient concentration in tissues is dependent on knowledge of the determinants of the body pool that is measured, including non-dietary influencing factors. For example, tocopherols and carotenoids are transported in the circulation nonspecifically by the cholesterolrich lipoproteins (Clevidence and Bieri 1993, Rock 1997, Romanchik et al. 1995), so higher concentrations of these lipoproteins are predictive of higher concentrations of the associated micronutrients, independent of dietary intake. Smoking and alcohol consumption may need to be considered in the interpretation of serum and other tissue concentrations of several micronutrients, particularly compounds that may be subject to oxidation (i.e., vitamin C, tocopherols, carotenoids, folate). Few data on the determinants of serum 25-hydroxyvitamin D and phylloquinone concentrations in a large, heterogeneous population have been previously collected or reported. The purpose of this study was to identify determinants of baseline serum concentrations of the carotenoids and fatsoluble vitamins in participants from the sentinel study site (Indianapolis, IN) of the OPMSS, prior to market introduction of olestra. This study identifies the demographic, dietary, and other lifestyle factors that influence serum concentrations of retinol, 25-hydroxyvitamin D, a-tocopherol, vitamin K (phylloquinone), a-carotene, b-carotene, lycopene, lutein, zeaxanthin, and b-cryptoxanthin in a large and heterogeneous group of free-living subjects. Characterization of the baseline determinants of these serum concentrations is an important first step in examining potential effects of olestra exposure.


ple was analyzed with each batch of samples. Between-day variability was 9.7 and 5.3% for the low-and high-concentration samples, respectively. Serum cholesterol and triglyceride analysis. Determinations of total cholesterol and triglycerides were performed at Quintiles Laboratories using enzymatic methods and the Boehringer Mannheim/ Hitachi 747 analyzer (Roche Lab Systems, Indianapolis, IN). Briefly, cholesterol esters in the serum are hydrolyzed quantitatively into free fatty acids and cholesterol by cholesterol esterase. In the presence of oxygen, free cholesterol is oxidized by cholesterol oxidase producing cholest-4-en-3-one and hydrogen peroxide. The hydrogen peroxide reacts in the presence of peroxidase with phenol and 4-aminophenazone to form an o-quinone imine dye. The intensity of the color formed is proportional to the cholesterol concentration and can be measured photometrically. Control samples are assayed with each batch of specimens. Precision studies were conducted using packaged reagents, pooled human serum, and control sera, and this cholesterol method meets the goal of #3% for both precision and bias. Serum triglyceride concentration was determined using an enzymatic method involving the hydrolysis of triglycerides with lipase to form glycerol and fatty acids. Glycerol was phosphorylated with adenosine triphosphate by glycerol kinase to form glycerol-3-phosphate and adenosine diphosphate. The glycerol-3-phosphate was then reacted with oxygen in the presence of glycerol kinase to form dihydroxyacetone phosphate and hydrogen peroxide. Hydrogen peroxide, 4-aminophenazone, and 4-cholorphenol were then treated with peroxidase to form a quinoneimine complex that absorbs at a primary wavelength of 505 nm. The increase in absorbance is proportional to the concentration of triglyceride in the sample. Laboratory precision and accuracy were determined as described above, achieving the same goal of #3% for both precision and bias. Statistical analysis. Participants with missing data on sex or age were excluded from analysis. Pregnant women were excluded because pregnancy can have profound effects on serum lipid and nutrient concentrations. Small numbers of participants had serum concentrations of vitamins or carotenoids that were not detectable by laboratory methods (4% for 25-hydroxyvitamin D, 3% for phylloquinone, 7% for a-carotene, and ,1% for the remaining analytes) and were assigned values at the midpoint between zero and the laboratory minimum detectable value. Some FFQ and body mass index (BMI) measurements were missing or excluded because of unreliable values [FFQ: energy intake ,3347 kJ/d (800 kcal/d) or .20,920 kJ/d (5000 kcal/d) for men (11%) or ,2510 kJ/d (600 kcal/d) or .16,736 kJ/d (4000 kcal/d) for women (9%); BMI: ,15 or .60 kg/m2 (3%)]. To include these subjects in the analysis, we used the following scheme to complete the data set. The values of the missing or excluded variables were set equal to the means of those variables among all subjects with valid values for those variables. Note that subjects with missing or excluded FFQ had all of their FFQ-derived variables imputed in this manner. Because the subjects who were excluded as outliers might differ systematically from the others (and the imputation of their missing or excluded data by the mean values may differ systematically from their true means), we added dummy variables to denote subjects whose FFQ was missing or excluded because of low or high calculated energy intake and subjects whose BMI was excluded. All models including FFQ or BMI data include the appropriate dummy variables. Linear regression was used to model the associations between each serum carotenoid and fat-soluble vitamin concentration (as a dependent variable) with demographic, dietary, and other factors. Table 1 gives details on the independent variables considered in each model. All models included demographic characteristics, total energy intake, and total intake (diet plus supplements) of the serum nutrient being modeled. Most models included either serum cholesterol or triglyceride concentration because this value reflects the lipoprotein fractions with which a micronutrient is associated. Models also included a dummy variable to indicate the use of a dietary supplement containing the nutrient being modeled, as an indicator of an effect that might occur with a very large level of nutrient intake. Models also included usual daily servings of fruits and vegetables to capture variability in carotenoid intake that may not be well assessed using

Downloaded from by guest on June 12, 2013

intake data were taken from the 24-h dietary recall because phylloquinone in serum has a very short half life, and thus, diet in the immediate past is believed to be most biologically relevant to serum concentrations (Booth and Suttie 1998). Micronutrients from vitamin supplements were collected using an inventory procedure based on abstracting information on fat-soluble vitamins and b-carotene from supplement bottle labels. Average daily dose of supplemental micronutrients was calculated by summing intakes from multivitamins and single supplements, taking into account the number of pills and frequency of use. Supplemental intakes of vitamins D and K were almost exclusively from multivitamins. Total micronutrient intakes used in this study include both supplemental and food sources. Serum carotenoid and fat-soluble vitamin analysis. Analysis of serum concentrations of retinol, 25-hydroxyvitamin D, a-tocopherol, and carotenoids (a-carotene, b-carotene, lycopene, lutein, zeaxanthin, and b-cryptoxanthin) was conducted at Quintiles Laboratories (Atlanta, GA). Analysis of serum phylloquinone concentration was conducted at the USDA Human Nutrition Research Center on Aging at Tufts University (Boston, MA). The separation and quantification of retinol, a-tocopherol, and carotenoids (a-carotene, b-carotene, lycopene, lutein, zeaxanthin, and b-cryptoxanthin) was accomplished with HPLC methodology. Serum samples were first precipitated with ethanol and extracted into hexane. The organic layer was removed, evaporated to dryness at room temperature, reconstituted, and transferred to an amber microvial for automatic injection. The HPLC method is gradient reversed phase, and the system is equipped with a photo diode array detector, a refrigerated automatic liquid sampling unit, and a Bakerbond Intermediate C18 column (Mallinckrodt Baker, Phillipsburg, NJ). The detector was set at 292 nm, 325 nm, and 451 nm for a-tocopherol, retinol, and carotenoids, respectively. The mobile phase components were acetonitrile with triethylamine (0.5 mL/L), methylene chloride with triethylamine (0.5 mL/L), and methanol with ammonium acetate (4.2 g/L) and triethylamine (0.5 mL/L), and the flow rate was 1.0 mL/min. Quantitation was performed using peak height ratios to internal standards for each of the three wavelengths. Accuracy was assessed by analysis of National Institute of Standards and Technology (NIST) Standard Reference Material SRM 986, Fat-Soluble Vitamins and by participation in the NIST Micronutrients Measurement Quality Assurance Program. The inter-assay precision for the analysis was determined at two levels ranging from low-normal to high-normal concentration. Coefficients of variation for individual analytes ranged from 2.7 to 9.8% for the low-normal sample and from 1.9 to 8.0% for the high-normal sample. 25-Hydroxyvitamin D was analyzed with the INCSTAR (INCSTAR, Stillwater, MN) 125I Radioimmunoassay (RIA) Kit, which consists of a two-step procedure. The first procedure involves the rapid extraction of 25-hydroxyvitamin D and other hydroxylated metabolites from serum with acetonitrile. Following extraction, the treated sample is then assayed using an equilibrium RIA procedure. The RIA method is based on an antibody with specificity to 25hydroxyvitamin D. Two quality control samples provided by INCSTAR (one at low-normal and one at high-normal) are analyzed with each batch of samples to assess the inter-assay precision, and coefficients of variation range from 5.7 to 9.2%. Vitamin K (phylloquinone) separation and quantification was accomplished with HPLC methodology using an internal standard (vitamin K1(25)) method of Davidson and Sadowski (1997). Serum was extracted with hexane after deproteinization with ethanol, and then the extract was dried, reconstituted, and subjected to solid-phase extraction on C18. The resulting extract was chromatographed using a BDS-Hypersil column (Keystone Scientific, Bellefonte, PA) and an HPLC system (model 510, Waters Chromatography, Milford, MA) equipped with a post-column, in-line zinc reactor to catalyze the reduction of phylloquinone to its fluorescent hydroquinone form, which was then quantified with a fluorometric detector (model 980, Applied Biosystems, Ramsey, NJ). The mobile phase was dichloromethane (0.1 L/L), methanol (0.9 L/L), water (5 mL/L), ZnCl2 (1.36 g/L), acetic acid (0.3 g/L), and sodium acetate (0.41 g/L). One low-normal and one high-normal concentration quality control sam-




TABLE 1 Independent variables in regression models predicting serum fat-soluble vitamin and carotenoid concentrations in the study population Variable Type

Variable (coding)



25-Hydroxyvitamin D




Age (per decade) Sex (female) Race/Ethnicity (non-white) Intake: diet1 & supplements (per 10% increase) FFQ2 low outlier FFQ high outlier FFQ missing Energy (per 10% increase) Supplement use % Energy from fat (per 5 percentage points) Fruits & vegetables (servings/d) Not fasting Cholesterol (per 10% increase) Triglycerides (per 10% increase) BMI4 BMI missing5 Physical activity (light, moderate, heavy) Smoking Alcohol consumption (per 15 g/d) Sun exposure (,0.5, 0.5, ,1, 1, 11 h/d)






X X X X 0 0

X X X X 0 0

X X X X 0 0


X 0


0 0 0

0 0




Serum Health behavior


X 0 0

X X X X 03 0 0 X 0

0 0


Downloaded from by guest on June 12, 2013

0 0 0 0


X 5 variable in model and results reported regardless of statistical significance. 0 5 variable only in model if statistically significant (P , 0.05). 1 Dietary intake estimated from food frequency questionnaire for retinol, a-tocopherol, vitamin D, and carotenoids. Dietary intake estimated from 24-h dietary recall for phylloquinone. 2 FFQ 5 food frequency questionnaire. 3 Supplement use assessed for b-carotene only. 4 BMI 5 body mass index calculated as weight (kg)/height (m2). 5 Included in model only if BMI also in model.

current nutrient databases. Finally, specific models included measures of factors that may uniquely affect serum nutrient concentrations, such as sun exposure, exercise, smoking, and BMI. All dependent variables (serum nutrients) were log transformed before analysis to improve normality. Thus, all regression coefficients were interpreted as percentage change in the serum nutrient associated with change in each independent variable. Most independent variables were also transformed and coded to facilitate interpretation of regression coefficients. Dietary intake measures, with the exception of percentage of energy from fat, were log transformed. The logarithms of nutrient and serum measures and BMI were divided by the logarithm of 1.10, percentage of energy from fat was divided by 5, and alcohol intake (g) was divided by 15 (thus representing number of standard drinks). Thus, after appropriate back-transformation when required, regression coefficients for these variables were interpreted as the effects of increasing exposure by 10%, 5%, and drinks per day, respectively. Exercise and sun exposure were coded as 1, 2, and 3, corresponding to none/light, moderate, or heavy and ,0.5 h/d, 0.5–1 h/d, and .1 h/d, respectively. Regression coefficients for these variables were interpreted as the percentage change in serum nutrient concentration associated with increasing one level in activity or sun exposure. Values are means 6 SD.

RESULTS Table 2 gives the demographic characteristics and healthrelated behavior of the study population. Thirty percent were smokers, 54% reported exercising less than once per week, and average fat intake was 35 6 8% of energy. Forty-five percent were overweight or obese, defined as BMI .27.3 kg/m2 for

women or .27.8 kg/m2 for men. Table 3 gives descriptive data on serum carotenoid and fat-soluble vitamin concentrations. Tables 4 and 5 present results of regression models predicting serum fat-soluble vitamin and carotenoid concentrations from demographic, dietary, and health-related characteristics. TABLE 2 Demographic characteristics of study sample (n 5 1042) Characteristic Age (y) 18–34 35–54 551 Sex Females Males Race/Ethnicity1 White African-American Hispanic Other Education,1 y #12 13–15 161



394 464 184

37.8 44.5 17.7

631 411

60.6 39.4

802 194 24 17

77.0 18.7 2.3 1.6

416 308 317

39.9 29.6 30.4

1 Cell sizes differ slightly because of missing values.



TABLE 3 Mean and distributions of serum fat-soluble vitamins and carotenoids by sex in the study population Women (n 5 631)

Men (n 5 411)

Percentiles Analyte Retinol,1 mmol/L 25-Hydroxyvitamin D, nmol/L a-Tocopherol,1 mmol/L Phylloquinone,1 nmol/L Carotenoids: a-Carotene,1 mmol/L b-Carotene,1 mmol/L Lycopene,1 mmol/L Lutein,1 mmol/L Zeaxanthin,1 mmol/L b-Cryptoxanthin,1 mmol/L





Geometric Mean (SD)




Geometric Mean (SD)

1.29 11.48 16.64 0.17

1.84 30.45 24.65 0.57

2.66 66.64 47.55 2.05

1.82 (0.55) 29.34 (20.49) 26.56 (12.03) 0.57 (0.55)

1.56 14.98 16.67 0.26

2.16 31.45 25.23 0.89

2.95 68.89 43.83 3.54

2.15 (0.56) 31.85 (19.45) 26.41 (10.21) 0.92 (0.93)

0.02 0.08 0.28 0.12 0.04 0.04

0.07 0.23 0.61 0.22 0.07 0.09

0.19 0.67 1.02 0.37 0.12 0.20

0.01 0.06 0.29 0.13 0.04 0.04

0.05 0.18 0.65 0.22 0.08 0.08

0.16 0.51 1.1 0.39 0.13 0.18

0.06 0.23 0.55 0.22 0.07 0.09

(0.06) (0.20) (0.34) (0.10) (0.04) (0.06)

0.04 0.18 0.58 0.22 0.07 0.08

(0.05) (0.15) (0.40) (0.10) (0.04) (0.06)

1 Cell sizes differ slightly because of missing values.

Downloaded from by guest on June 12, 2013

Significant determinants of serum retinol concentration were age, sex, race/ethnicity, vitamin A intake (adjusted for energy intake), and alcohol consumption, although the model explained only 20% of the variance observed. The most substantial relationship observed was between sex and serum retinol concentration, with a difference of nearly -17% attributable to female (versus male) sex. Vitamin D intake and sun exposure were both independently positively associated with serum 25-hydroxyvitamin D after adjustment for other variables in the model, and BMI was significantly inversely associated with serum 25-hydroxyvitamin D concentration. Race/ethnicity and smoking status were also found to be significant independent determinants of serum 25-hydroxyvitamin D, with nonwhite race/ethnicity and smoking associated with lower concentrations than white race/ ethnicity and nonsmokers. The most substantial relationship was between race/ethnicity and serum 25-hydroxyvitamin D, with a difference of ;244% attributable to nonwhite (versus white) race/ethnicity. Overall, ;24% of the variance in serum 25-hydroxyvitamin D concentration was explained by the variables in this model. In contrast, the model for a-tocopherol explained nearly two-thirds (62%) of the variance observed in serum a-tocopherol concentration. Serum cholesterol concentration, age, and a-tocopherol intake were all significantly and positively associated with serum a-tocopherol concentration. Serum cholesterol exerted the strongest effect; i.e., a 10% increase in cholesterol concentration was associated with a 10% increase in serum a-tocopherol concentration. Percentage energy from fat in the diet and BMI were both inversely related to serum a-tocopherol concentration. When level of physical activity and smoking status were entered into the model, only the inverse relationship with smoking tended to be significant (P 5 0.067). As a component of the total a-tocopherol intake, the vitamin E in dietary supplements was directly related to serum a-tocopherol, but supplement use (as a dichotomous variable) was inversely associated with circulating concentration. Other significant independent determinants of serum a-tocopherol were sex and race/ethnicity, with female sex and nonwhite race/ethnicity associated with lower concentrations than male sex and white race/ethnicity when adjusted for other variables in the model. In the model that explained 40% of the variance in serum

phylloquinone, age, vitamin K (phylloquinone) intake, and serum triglyceride concentration were found to be significantly directly related to serum phylloquinone concentration. Female (versus male) sex was predictive of lower serum phylloquinone concentration, accounting for a difference of ;214%. Similar to the relationship between supplement use and serum a-tocopherol concentration, vitamin K supplementation (as a dichotomous variable) was inversely associated with circulating concentration, although in this case, multivitamins were the only supplement source of the vitamin, as noted above. The relationship between race/ethnicity and serum phylloquinone was not statistically significant. As shown in Table 5, the determinants identified for the various serum carotenoids differed somewhat across the regression models for these compounds. Age was significantly and directly associated with serum a-carotene, b-carotene, and lutein concentration, but was inversely associated with serum lycopene concentration. Female (versus male) sex was predictive of higher serum a-carotene and b-carotene concentrations and significantly lower serum lutein and zeaxanthin concentrations. Race/ethnicity was significantly independently associated with serum b-cryptoxanthin, lutein, and zeaxanthin concentrations, with nonwhite (versus white) status predictive of higher concentrations of these carotenoids. Intake was directly associated with serum concentrations of all carotenoids, although only marginally significantly associated for zeaxanthin (P 5 0.06), and serum cholesterol was directly associated with all except a-carotene. Alcohol consumption was inversely associated with both serum b-carotene and lycopene concentrations, but was directly related to serum zeaxanthin concentration. Fruit and vegetable servings were also independently associated with serum b-cryptoxanthin and lutein concentrations. Smoking was inversely related to all of the serum carotenoids, and BMI, when adjusted for other variables in the model, was inversely associated with all except serum lycopene concentrations. Models for carotenoids explained 15–26% of the variance in serum carotenoid concentrations. DISCUSSION Several demographic, dietary, and other lifestyle factors were found to be determinants of serum carotenoid and fat-



TABLE 4 Predictors of serum concentrations of fat-soluble vitamins in multivariate analyses in the study population (n 5 1042) Variable (coding)

3.8 (2.6, 5.0)* 216.8 (219.8, 213.8)* 210.6 (214.0, 27.0)* 0.8 (0.5, 1.1)* 20.9 (21.3, 20.4)* 3.5 (24.5, 12.2) 1.2 (28.4, 11.9) 22.3 (215.6, 13.1) 24.2 (210.1, 2.1)

25-Hydroxyvitamin D % Change (95% CI)

a-Tocopherol % Change (95% CI)

21.9 (24.3, 0.6) 26.4 (213.4, 1.2)

6.3 (5.1, 7.6)* 23.6 (26.9, 20.1)*

3.8 (0.4, 7.3)* 214.2 (222.6, 25.0)*

27.8 (211.5, 24.1)*

2.8 (28.5, 15.5)

244.3 (248.9, 239.2)* 0.7 (0.3, 1.2)* 20.6 (21.5, 0.4) 6.1 (211.2, 26.6) 26.3 (224.7, 16.6) 19.8 (213.1, 65.3) 214.5 (225.7, 21.7)*

1.4 (1.2, 1.6)*

Phylloquinone % Change (95% CI)

2.5 (2.1, 2.9)*

21.0 (21.5, 20.6)* 3.4 (24.7, 12.3) 23.4 (212.6, 6.8) 4.8 (29.5, 21.4) 22.0 (28.0, 4.5)

6.9 (211.4, 29.0)

10.0 (9.1, 10.8)*

0.5 (22.0, 3.1) 8.5 (7.7, 9.4)*

26.0 (27.4, 24.5)*

21.0 (21.7, 20.3)*

19.6 (25.6, 51.6) 29.8 (216.7, 22.3)*

23.9 (213.8, 7.2) 23.4 (26.9, 0.2)

2.3 (0.7, 3.9)*

0.6 (29.6, 12.0) 0.7 (23.9, 5.5)

20.7 (22.7, 1.4) 6.5 (0.9, 12.4)* 0.5 (20.6, 1.7) 2.9 (21.9, 8.0)


1.0 (21.5, 3.5) 210.8 (221.4, 1.2)


22.0 (23.1, 20.8)* 27.3 (213.2, 21.0)*


Downloaded from by guest on June 12, 2013

Variables in all models Demographic Age (per decade) Sex (female) Race/Ethnicity (nonwhite) Diet Micronutrient (per 10% increase) Energy (per 10% increase) FFQ low outlier1 FFQ high outlier2 FFQ missing Serum Not Fasting Other variables Serum Cholesterol (per 10% increase) Triglycerides (per 10% increase) Health behavior Body mass index (per 10% increase) Body mass index missing Smoking Alcohol consumption (per 15 g/d) Physical activity (light, moderate, heavy) Sun exposure (,0.5, 0.5, ,1, 11 h/d) Diet % Energy from Fat (per 5%) Supplement Use Fruits and Vegetables (servings/d) Variance explained (R2)3

Retinol % Change (95% CI)

214.7 (224.8, 23.3)*


* P , 0.05. 1 Indicator variable for FFQ excluded because energy intake ,3347 kJ/d (800 kcal/d) for males or ,2510 kJ/d (600 kcal/d) for females. 2 Indicator variable for FFQ excluded because energy intake .20,920 kJ/d (5000 kcal/d) for males or .16,736 kJ/d (4000 kcal/d) for females. 3 Statistics based on model with variables used in all models and the other variables that were significant (P , 0.05).

soluble vitamin concentrations in this study involving a large, heterogeneous group of free-living individuals. However, the numerous factors that we examined were observed to explain little of the variability in serum concentrations of retinol, 25-hydroxyvitamin D, and the carotenoids and only modest amounts of the variability in serum phylloquinone concentration. In contrast, more substantial amounts of the variability in serum a-tocopherol concentration could be explained by intake, demographic characteristics, circulating cholesterol concentration, and lifestyle factors.

The usefulness of biochemical indicators of nutritional status in community-based research is built upon knowledge of the physiologic and other determinants of the measures. Knowledge of the relationship between the indicator and the risk of nutrient depletion, in addition to the responsiveness of the indicator to interventions or change (Habicht and Pelletier 1990), is also necessary. Practical considerations include the ability to conveniently access the body compartment for measurement, the procedures necessary to collect and process the sample, subject burden, and the resources for laboratory



TABLE 5 Predictors of serum concentrations of carotenoids in multivariate analyses in the study population (n 5 1042) Variable (coding)

b-Carotene % Change (95% CI)

10.4 (5.8, 15.2)* 19.9 (5.5, 36.3)*

6.7 (3.1, 10.4)* 11.7 (0.7, 23.9)*

211.9 (224.0, 2.1) 2.9 (2.4, 3.5)* 22.2 (23.7, 20.6)* 236.5 (252.8, 214.4)* 241.3 (259.4, 215.2)* 2.8 (240.0, 76.2)

1.8 (29.3, 14.1) 2.9 (2.4, 3.4)* 21.3 (22.5, 20.1)* 26.4 (226.0, 18.4) 218.2 (238.6, 9.1) 245.3 (264.1, 216.6)*

Lycopene % Change (95% CI)

Lutein % Change (95% CI)

Zeaxanthin % Change (95% CI)

b-Cryptoxanthin % Change (95% CI)

211.3 (213.6, 29.0)* 25.5 (212.7, 2.3)

3.9 (2.0, 5.8)* 27.0 (212.0, 21.8)*

0.4 (21.6, 2.4) 28.6 (214.1, 22.7)*

21.3 (23.9, 1.4) 22.6 (210.1, 5.6)

22.3 (210.8, 6.9)

19.4 (12.2, 27.0)*

26.3 (18.0, 35.3)*

26.7 (15.7, 38.7)*

1.8 (1.2, 2.5)*

0.8 (0.4, 1.3)*

20.9 (21.9, 0.1) 20.8 (21.5, 20.2)* 0.6 (216.2, 20.6) 211.1 (221.7, 0.9) 23.5 (222.7, 20.6) 25.7 (219.2, 10.1) 248.1 (262.5, 228.1)* 6.7 (215.0, 33.9)

20.9 (24.3, 52.8)

29.3 (224.5, 8.9)

217.1 (228.0, 24.4)*

2.1 (20.8, 5.0)

4.2 (1.9, 6.6)*

8.7 (6.9, 10.6)*

29.6 (211.9, 27.2)*

29.3 (211.1, 27.4)*

21.1 (22.7, 0.5)

0.3 (232.6, 49.2) 234.3 (242.5, 224.9)*

12.8 (217.3, 54.0) 222.3 (230.1, 213.5)*

23.9 (224.5, 22.2) 210.4 (217.4, 22.8)*

21.9 (27.5, 4.0)

24.8 (29.1, 20.4)*

25.9 (29.2, 22.4)*

23.8 (27.8, 0.3)

21.7 (25.0, 1.7) 24.5 (215.7, 8.2)

3.2 (0.6, 5.9)*

1.7 (22.2, 5.7)

1.2 (21.7, 4.3)

22.0 (24.2, 0.3)




0.3 (0.0, 0.7) 20.2 (21.0, 0.5) 211.7 (223.3, 1.7) 25.3 (220.3, 12.5) 27.6 (228.2, 19.0)

0.8 (0.5, 1.0)* 20.9 (21.8, 0.1) 227.2 (239.6, 212.1)* 229.2 (243.7, 211.1)* 17.1 (216.2, 63.8)

7.2 (22.8, 18.3)

7.7 (23.4, 20.2)

8.2 (26.5, 25.1)

7.3 (6.1, 8.6)*

6.8 (5.4, 8.3)*

5.7 (3.9, 7.6)*

24.4 (25.5, 23.3)*

23.4 (24.6, 22.2)*

28.2 (222.3, 8.6) 216.5 (230.7, 0.5) 28.5 (213.5, 23.2)* 27.6 (213.2, 21.5)* 1.6 (20.9, 4.2)

26.5 (28.0, 24.9)* 212.2 (231.4, 12.4) 221.0 (227.4, 214.1)*

3.0 (0.2, 5.8)*

0.6 (23.0, 4.3)

21.1 (23.0, 0.8)

1.5 (20.5, 3.6)

0.1 (22.7, 3.0)

2.0 (0.1, 4.0)*

0.2 (21.9, 2.4)

3.5 (1.1, 5.9)*




* P , 0.05. 1 Indicator variable for FFQ excluded because energy intake ,3347 kJ/d (800 kcal/d) for males or ,2510 kJ/d (600 kcal/d) for females. 2 Indicator variable for FFQ excluded because energy intake .20,920 kJ/d (5000 kcal/d) for males or .16,736 kJ/d (4000 kcal/d) for females. 3 For lutein and zeaxanthin, values used are lutein plus zeaxanthin. 4 Statistics based on model with variables used in all models and other variables that were significant (P , 0.05).

analysis. All of the micronutrients examined in this study are fat-soluble dietary constituents and could be influenced by olestra consumption. However, the concentration in the circulating body pool is variably reflective of overall status or total body reserves across this group of compounds, and a wide range of physiologic and metabolic characteristics are represented. As previously reviewed and summarized by Olson (1991 and 1996) and Underwood (1990), the amount of retinol (the predominant form of vitamin A) in the circulating body pool is highly regulated and is essentially homeostatically controlled when liver stores are adequate. However, sustained dietary inadequacy eventually results in a decline in the serum concentration, and the percentage of individuals with low values in a target group may be useful as an indicator of the vitamin A status of the population. Sex and age were significant predictors of serum retinol concentration in the present study. These factors likely represent physiologic factors that influence the rate of synthesis, release, and degradation of holo-retinol binding protein. We also found vitamin A intake, as well as race/ethnicity, to be significant determinants of serum retinol concentration. Race/ethnicity could feasibly re-

Downloaded from by guest on June 12, 2013

Variables in all models Demographic Age (per decade) Sex (female) Race/Ethnicity (nonwhite) Diet Carotenoid3 (per 10% increase) Energy (per 10% increase) FFQ low outlier1 FFQ high outlier2 FFQ missing Serum Not fasting Other variables Serum cholesterol (per 10% increase) Health behavior Body mass index (per 10% increase) Body mass index missing Smoking Alcohol consumption (per 15 g/d) Diet % Energy from fat (per 5%) Supplement use Fruits and vegetables (servings/d) Variance explained (R2)4

a-Carotene % Change (95% CI)

flect heritable physiologic or metabolic factors. Alcohol is known to promote increased oxidation of vitamin A compounds and reduced liver stores, but has also been associated with increased serum retinol concentration (Olson 1991), as was observed in the present study. Serum a-tocopherol concentration generally reflects overall status for vitamin E, although this value is best described in relation to serum lipids, or more specifically, total serum cholesterol, which reflects the lipoprotein faction with which tocopherols are transported in the circulation (Thurnham et al. 1986). In contrast to the relationship between vitamin A intake and serum retinol concentration, a-tocopherol in the circulating pool and peripheral tissues varies directly with the tocopherol intake (Machlin 1991). The regulatory mechanism that was identified for the tocopherols limits the amount of other tocopherols or isomeric forms retained by the body, but does not function to limit the uptake of a-tocopherol (Traber and Kayden 1989,Traber et al. 1990 ). The inverse relationship between serum a-tocopherol and the indicator variable of supplement use (yes or no) in the present study analysis presumably reflects a declining effect on the serum concentration with the administration of high doses because of reduced



compounds may also affect concentrations in tissues. Otherwise, carotenoids appear to be remarkably unregulated and are transported in the circulation in association with cholesterolrich lipoproteins (Clevidence and Bieri 1993). Overall, the determinants of serum carotenoid concentrations that were observed in the present study were similar to those observed in previous studies of such associations (Ascherio et al. 1992, Brady et al. 1996, Drewnowski et al. 1997). Serum cholesterol concentration was generally predictive of higher concentrations, whereas smoking, alcohol consumption, and increased BMI were generally associated with lower serum concentrations of the various carotenoids. As observed in previous studies (Ascherio et al. 1992, Rock et al. 1997), the amount of variability in serum carotenoid concentrations that could be explained by quantified dietary intake, demographic characteristics, BMI, circulating lipid concentrations, smoking, and alcohol consumption was not substantial. The capability of accurately measuring dietary intake has a powerful effect on the ability to explain the variability in the serum nutrient concentrations. Errors or inaccuracy in the measurement of intake reduce the likelihood that the variability can be explained and also increase the likelihood that surrogate variables will be identified as determinants. One important source of error is the nutrient database because the quality of the food content data, and also the variability in content of individual foods, varies across these nutrients. HPLC data on the content of phylloquinone in foods has only recently become available (Booth and Suttie 1998), and the quality of the carotenoid food content data is still considered quite limited (Mangels et al. 1993). In the present study, errors in quantifying intake are likely to have affected the identification of determinants of serum concentrations and the apparent importance of these factors. For example, more frequent consumption of lutein-rich foods, such as greens and spinach, has been observed among African-Americans compared to whites (Swanson et al. 1993), so that differences in intake are the most likely explanation for the identification of race/ ethnicity as a determinant of serum lutein. Dietary and lifestyle factors are often closely related, so that the relative importance as determinants may be inaccurately described. A biological effect of exposure to cigarette smoke is the oxidation and increased turnover of plasma tocopherols, carotenoids, and retinol (Handelman et al. 1996). However, smokers typically consume substantially smaller amounts of fruits, vegetables, and other good sources of these micronutrients (Brady et al. 1996, Marangon et al. 1998). In survey studies, smokers have also been reported to consume significantly fewer dairy products than nonsmokers (Marangon et al. 1998), so that smoking behavior may be identified as a determinant of serum 25-hydroxyvitamin D concentration because of limitations in quantifying vitamin D intake, while the relationship may actually be mediated by dietary differences. The ability to accurately describe the relationship between intake and serum concentration is also influenced by the time span that is the focus, given the differences in turnover rate and the relative amount of total body reserves. For example, serum a-tocopherol represents a relatively large body pool, and tissue turnover is slow compared to many other micronutrients, so demonstrating a relationship with average dietary intake assessed by FFQ (which captures a relatively long-term dietary pattern) is a reasonable goal. In comparison, the total body pool of vitamin K is relatively small, and tissue turnover is fast. Thus, the use of recent phylloquinone intake data in the model to predict serum concentration enabled explaining

Downloaded from by guest on June 12, 2013

absorption efficiency; i.e., the effect is decreased compared to what might be expected. Although absorption of the vitamin E compounds requires biliary and pancreatic secretions, we found an inverse relationship with percentage energy from fat in the diet and serum a-tocopherol concentration. It is possible that a lower level of dietary fat intake occurs in subjects in this population in association with other unmeasured factors that tend to increase a-tocopherol intake or serum concentrations, and even the lower-fat diets in this group presumably provide adequate dietary fat to enable tocopherol absorption. The circulating body pool of 25-hydroxyvitamin D is considered a good biochemical indicator of vitamin D status (Collins and Norman 1991), but similar to retinol, a specific protein synthesized in the liver carries this compound in the circulation. We found race/ethnicity to be the most important determinant of serum 25-hydroxyvitamin D concentration, a relationship that has been observed in smaller study populations (Harris and Dawson-Hughes 1998a). In the present study, both vitamin D intake and sun exposure were directly associated with blood concentration. All subjects were measured and queried in the fall and winter months, so the possibility of changes in the degree of influence of these factors across the seasons (Salamone et al. 1993, Webb et al. 1990) could not be addressed. Also, constraints in the depth and scope of the analysis of medication use limited our ability to identify effects of drugs on serum 25-hydroxyvitamin D concentration, which might have also contributed to variability in the concentration (Harris and Dawson-Hughes 1998b). Circulating vitamin K (phylloquinone) concentration was observed to be highly dependent on dietary intake within the previous 24 h and thus is not considered a sensitive measure of overall vitamin K status (Booth and Suttie 1998). In the present study, we confirmed the association between vitamin K intake within the preceding 24 h and serum phylloquinone concentration. The inverse relationship between the supplemental vitamin K indicator variable and serum phylloquinone concentration likely either reflects declining absorption efficiency with higher levels of intake or suggests the influence of some other characteristic of individuals who use dietary supplements. Serum triglyceride concentration, which reflects the level of triglyceride-rich lipoproteins that carry phylloquinone in the circulation in both fasting and postprandial states (Booth and Suttie 1998, Lamon-Fava et al. 1998), was identified as another important determinant. We found a positive association between age and serum phylloquinone concentration in the present study, which was observed in previous smaller studies of serum phylloquinone concentrations in healthy subjects (Sadowski et al. 1989, Sokoll and Sadowski 1996). We also found a relationship between serum phylloquinone and sex, with women having lower concentrations than men. An independent effect of sex was not consistently observed in previous studies of serum phylloquinone concentration in healthy subjects (Sokoll and Sadowski 1996); however, these relationships have been previously examined in only a limited number of subjects or groups. Similar to a-tocopherol, carotenoids in the circulating pool were shown to directly reflect intake of these compounds, increasing in response to increased intake (Albanes et al. 1997, Rock et al. 1997 and 1998) and declining with depletion (Rock et al. 1992). However, several factors influence the absorption efficiency, and wide interindividual variability in the degree of responsiveness is typically observed. Regulated mechanisms enable the conversion of provitamin A carotenoids to retinal in intestinal and hepatic tissue, and nonenzymatic (or additional enzyme-mediated) oxidation of these


ACKNOWLEDGMENTS The authors gratefully acknowledge Sarah Booth and Kenneth Davidson at Tufts University for providing vitamin K food content and analytic data and John Peters at Procter & Gamble (Cincinnati, OH) for his involvement and invaluable contributions to this project.

LITERATURE CITED Albanes, D., Virtamo, J., Taylor, P. R., Rautalahti, M., Pietinen, P. & Heinonen, O. P. (1997) Effects of supplemental b-carotene: Cigarette smoking, and alcohol consumption on serum carotenoids in the Alpha-Tocopherol, BetaCarotene Cancer Prevention Study. Am. J. Clin. Nutr. 66: 366 –372. Ascherio, A., Stampfer, M. J., Colditz, G. A., Rimm, E. B., Litin, L. & Willett, W. C. (1992) Correlations of vitamin A and E intakes with the plasma concentrations of carotenoids and tocopherols among American men and women. J. Nutr. 122: 1792–1801. Booth, S. L. & Suttie, J. W. (1998) Dietary intake and adequacy of vitamin K. J. Nutr. 128: 785–788. Booth, S. L., Pennington, J.A.T. & Sadowski, J. A. (1996a) Food sources and dietary intakes of vitamin K1 (phylloquinone) in the American diet: Data from the FDA Total Diet Study. J. Am. Diet. Assoc. 96: 149 –154. Booth, S. L., Pennington, J.A.T. & Sadowski, J. A. (1996b) Dihydro-vitamin K1: Primary food sources and estimated dietary intakes in the American diet. Lipids 31: 715–720. Booth, S. L., Sadowski, J. A., Weihrauch, J. L. & Ferland, B. (1993) Vitamin K1 (phylloquinone) content of foods: A provisional table. J. Food Compos. Anal. 6: 109 –120. Brady, W. E., Mares-Perlman, J. A., Bowen, P. & Stacewicz-Sapuntzakis, M. (1996) Human serum carotenoid concentrations are related to physiologic and lifestyle factors. J. Nutr. 126: 129 –137. Briefel, R., Sowell, A., Huff, D., Hodge, C., Koncikowski, S., Wright, J. & Gunter, E. (1996) The distribution of serum carotenoids in the U.S. population,. (1988) -1994: Results from the Third National Health and Nutrition Examination Survey (NHANES III). FASEB J. 10: A4700 (abs). Chug-Ahuja, J. K., Holden, J. M., Forman, M. R., Mangels, A. R., Beecher, G. R. & Lanza, E. (1993) The development and application of a carotenoid database for fruits, vegetables, and selected multicomponent foods. J. Am. Diet. Assoc. 93: 318 –323. Clevidence, B. A. & Bieri, J. G. (1993) Association of carotenoids with human plasma lipoproteins. Methods Enzymol. 214: 33– 46. Collins, E. D. & Norman, A. W. (1991) Vitamin D. In: Handbook of Vitamins (Machlin, L. J., ed.), 2nd ed., pp. 59 –98. Marcel Dekker, New York, NY. Crespo, C. J., Keteyian, S. J., Heath, G. W. & Sempos, C. T. (1996) Leisure-time physical activity among US adults. Results from the Third National Health and Nutrition Examination Survey. Arch. Intern. Med. 156: 93–98. Davidson, K. W. & Sadowski, J. A. (1997) Determination of vitamin K compounds in plasma or serum by high performance liquid chromatography (HPLC) using post-column chemical reduction and fluorimetric detection. Methods Enzymol. 282: 408 – 421.

Drewnowski, A., Rock, C. L., Henderson, S. A., Shore, A. B., Fischler, C., Galan, P., Preziosi, P. & Hercberg, S. (1997) Serum b-carotene and vitamin C as biomarkers of vegetable and fruit intakes in a community-based sample of French adults. Am. J. Clin. Nutr. 65: 1796 –1802. Federal Register (1996) Part III, Department of Health and Human Services, Food and Drug Administration, 21 CFR Part 172. Food additives permitted for direct addition to food for human consumption: olestra, final rule. 62: 3118 – 3173. Habicht, J. P. & Pelletier, D. L. (1990) The importance of context in choosing nutritional indicators. J. Nutr. 120 (suppl.): 1519 –1524. Handelman, G. J., Packer, L. & Cross, C. E. (1996) Destruction of tocopherols, carotenoids, and retinol in human plasma by cigarette smoke. Am. J. Clin. Nutr. 63: 559 –565. Harris, S. S. & Dawson-Hughes, B. (1998a) Seasonal changes in plasma 25hydroxyvitamin D concentrations of young American black and white women. Am. J. Clin. Nutr. 67: 1232–1236. Harris, S. S. & Dawson-Hughes, B. (1998b) The association of oral contraceptive use with plasma 25-hydroxyvitamin D levels. J. Am. Coll. Nutr. 17: 282–284. Havas, S., Heimendinger, J., Damron, D., Nicklas, T. A., Cowan, A., Beresford, S. A., Sorensen, G., Buller, D., Bishop, D., Baranowski, T. & Reynolds., K. (1995) 5-a-day for better health-nine community research projects to increase fruit and vegetable consumption. Pub. Hlth. Rep. 110: 68 –79. Jones, D. Y., Koonsvitsky, B. P., Ebert, M. L., Jones, M. B., Lin, P.T.Y, Will, B. H. & Suttie, J. W. (1991a) Vitamin K status of free-living subjects consuming olestra. Am. J. Clin. Nutr. 53: 943–946. Jones, D. Y., Miller, K. W., Koonsvitsky, B. P., Ebert, M. L., Lin, P.Y.T, Jones, M. B. & DELuca, H. F. (1991b) Serum 25-hydroxyvitamin D concentrations of free-living subjects consuming olestra. Am. J. Clin. Nutr. 53: 1281–1287. Koivu, T. J., Henttonen, S. K., Mattila, P. H. & Piironen, V. I. (1997) Determination of phylloquinone in vegetables, fruits, and berries by high-performance liquid chromatography with electrochemical detection. J. Agric. Food Chem. 45: 4644 – 4649. Koonsvitsky, B. P., Berry, D. A., Jones, M. B., Lin, P.Y.T, Cooper, D. A., Jones, D. Y. & Jackson, J. E. (1997) Olestra affects serum concentrations of a-tocopherol and carotenoids but not vitamin D or vitamin K status in free-living subjects. J. Nutr. 127: 1636S–1645S. Krebs-Smith, S. M., Heimendinger, J., Subar, A. F., Patterson, B. H. & Pivonka, E. (1995) Using food frequency questionnaires to estimate fruit and vegetable intake: Association between the number of questions and total intakes. J. Nutr. Educ. 27: 80 – 85. Kristal, A. R., Patterson, R. E., Neuhouser, M. L., Thornquist, M., NeumarkSztainer, D., Rock, C. L., Berlin, M. C., Cheskin, L. & Schreiner, P. (1998) The olestra post-marketing surveillance study: Design and baseline results from the sentinel site. J. Am. Diet. Assoc. 98: 1290 –1296. Kristal, A. R., White, E., Davis, J. R., Coryell, G., Raghunathan, T., Kinne, S. & Lin, T. K. (1993) Effects of enhanced calling efforts on response rates, estimates of health behavior, and costs in a telephone survey using random-digit dialing. Pub. Hlth. Rep. 108: 372–379. Kuczmarski, R. J., Flegal, K. M., Campbell, S. M. & Johnson, C. L. (1994) Increasing prevalence of overweight among US adults. The National Health and Nutrition Examination Surveys, 1960 to 1991. J. Am. Med. Assoc. 272: 205–211. Lamon-Fava, S., Sadowski, J. A., Davidson, K. W., O’Brien, M., MCNamara, J. R. & Schaefer, E. J. (1998) Plasma lipoproteins as carriers of phylloquinone (vitamin K1) in humans. Am. J. Clin. Nutr. 67: 1226 –1231. Machlin, L. J. (1991) Vitamin A. In: Handbook of Vitamins (Machlin, L. J., ed.), 2nd ed., pp. 99 –144. Marcel Dekker, New York, NY. Mangels, A. R., Holden, J. M., Beecher, G. R., Forman, M. R. & Lanza, E. (1993) Carotenoid content of fruits and vegetables: An evaluation of analytic data. J. Am. Diet. Assoc. 93: 284 –296. Marangon, K., Herbeth, B., Lecomte, E., Paul-Dauphin, A., Grolier, P., Chancerelle, Y., Artur, Y. & Siest, G. (1998) Diet, antioxidant status, and smoking habits in French men. Am. J. Clin. Nutr. 67: 231–239. McDowell, M. A., Briefel, R. R., Alaimo, K., Bischoff, A. M., Caughman, C. R., Carroll, M. D., Loria, C. M. & Johnson, C. L. (1994) Energy and macronutrient intakes of persons ages 2 months and over in the United States: Third National Health and Nutrition Examination Survey, Phase 1, 1988 –91. Adv. Data Vital Health Stat., No. 255. National Center for Health Statistics, Hyattsville, MD. Olson, J. A. (1991) Vitamin A. In: Handbook of Vitamins (Machlin, L. J., ed.), 2nd ed., pp. 1–57. Marcel Dekker, New York, NY. Olson, J. A. (1996) Biochemistry of vitamin A and carotenoids. In: Vitamin A Deficiency: Health, Survival, and Vision (Sommer, A. & West, K. P., eds.), pp. 221–250. Oxford University Press, New York, NY. Patterson, R. E., Kristal, A. R., Coates, R., Tylavsky, F., Ritenbaugh, C., Caggiula, A., Van Horn, L. & Snetselaar, L. (1996) Low-fat diet practices of older women: Prevalence and implications for dietary assessment. J. Am. Diet. Assoc. 96: 670 – 679. Patterson, R. E., Kristal, A. R., Tinker, L. F., Carter, R. A., Bolton, M. P. & Agurs-Collins, T. 1999 Measurement characteristics of the women’s health initiative food frequency questionnaire. Ann. Epidemiol. (in press). Prince, D. M. & Welschenbach M. A. (1998) Olestra: A new food additive. J. Am. Diet. Assoc. 98: 565–569.

Downloaded from by guest on June 12, 2013

a fair amount of the variability in the serum concentration of that vitamin. Overall, the serum concentrations of carotenoids and fatsoluble vitamins that were observed in this study population are similar to those reported in previous descriptive studies and surveys of healthy subjects in the U.S. (Ascherio et al. 1992, Brady et al. 1996, Briefel et al. 1996, Harris and DawsonHughes 1998a, Sadowski et al. 1989). The dietary patterns and lifestyle characteristics are also similar to those of the general U.S. population (Crespo et al. 1996, MCDowell et al. 1994), although the prevalence of overweight is somewhat higher than the national average of 33% (Kuczmarski et al. 1994), which probably reflects the region in which the study sentinel site is based (the midwestern U.S.). In conclusion, several demographic, dietary, and lifestyle factors were identified as determinants of serum carotenoid and fat-soluble vitamin concentrations in the first clinical cross section at the sentinel site of the OPMSS. Differing physiologic and metabolic characteristics of these compounds contributed to differences in the ability to explain the variability in serum concentrations. These baseline findings enable assessment of an independent effect of olestra exposure on these serum concentrations as one aspect of nutritional monitoring.




Rock, C. L. (1997) Carotenoids: Biology and treatment. Pharmacol. Therapeutics 75: 185–197. Rock, C. L., Flatt, S. W., Wright, F. A., Faerber, S., Newman, V., Kealey, S. & Pierce, J. P. (1997) Responsiveness of carotenoids to a high-vegetable diet intervention designed to prevent breast cancer recurrence. Cancer Epidemiol. Biomarkers Prev. 6: 617– 623. Rock, C. L., Lovalvo, J. L., Emenhiser, C., Ruffin, M. T., Flatt, S. W. & Schwartz, S. J. (1998) Bioavailability of b-carotene is lower in raw than in processed carrots and spinach in women. J. Nutr. 128: 913–916. Rock, C. L., Swendseid, M. E., Jacob, R. A. & McKee, R. W. (1992) Plasma carotenoid levels in human subjects fed a low-carotenoid diet. J. Nutr. 122: 96 –100. Romanchik, J. E., Morel, D. W. & Harrison, E. H. (1995) Distribution of carotenoids and a-tocopherol among lipoproteins do not change when human plasma is incubated in vitro. J. Nutr. 125: 2610 –2617. Sadowski, J. A., Hood, S. J., Dallal, G. E. & Garry, P. J. (1989) Phylloquinone in plasma from elderly and young adults: Factors influencing its concentration. Am. J. Clin. Nutr. 50: 100 –108. Salamone, L. M., Dallal, G. E., Zantos, D., Makrauer, F. & Dawson-Hughes, B. (1993) Contributions of vitamin D intake and seasonal sunlight exposure to plasma 25-hydroxyvitamin D concentration in elderly women. Am. J. Clin. Nutr. 58: 80 – 86. Schlagheck, T. G., Kesler, J. M., Jones, M. B., Zorich, N. L., Dugan, L. D., Davidson, M. H. & Peters, J. (1997) Olestra’s effect on vitamins D and E in humans can be offset by increasing dietary levels of these vitamins. J. Nutr. 127: 1666S–1685S. Sokoll, L. J. & Sadowski, J. A. (1996) Comparison of biochemical indexes for

assessing vitamin K nutritional status in a healthy population. Am. J. Clin. Nutr. 63: 566 –573. Swanson, C. A., Gridley, G., Greenberg, R. S., Schoenberg, J. B., Swanson, G. M., Brown, L. M., Hayes, R., Silberman, D. & Pottern, L. (1993) A comparison of blacks and whites in three areas of the United States. Nutr. Cancer 20: 153–165. Thurnham, D. I., Davies, J. A., Crump, B. J., Situnayake, R. D. & Davis, M. (1986) The use of different lipids to express serum tocopherol:lipid ratios for the measurement of vitamin E status. Ann. Clin. Biochem. 23: 514 –520. Traber, M. G. & Kayden, H. J. (1989) Preferential incorporation of alpha-tocopherol vs. gamma-tocopherol in human lipoproteins. Am. J. Clin. Nutr. 49: 517–526. Traber, M. G., Burton, G. W., Ingold, K. U. & Kayden, H. J. (1990) RRR- and SRR-alpha-tocopherols are secreted without discrimination in human chylomicrons, but RRR-alpha-tocopherol is preferentially secreted in very low density lipoproteins. J. Lipid Res. 31: 675– 685. Underwood, B. A. (1990) Methods for assessment of vitamin A status. J. Nutr. 120 (suppl.): 1459 –1463. Webb, A. R., Pilbeam, C., Hanafin, N. & Holick, M. F. (1990) An evaluation of the relative contributions of exposure to sunlight and of diet to the circulating concentrations of 25-hydroxyvitamin D in an elderly nursing home population. Am. J. Clin. Nutr. 51: 1075–1081. Weststrate, J. A. & van het Hoff, K. H. (1995) Sucrose polyester and plasma carotenoid concentrations in healthy subjects. Am. J. Clin. Nutr. 62: 591– 597.

Downloaded from by guest on June 12, 2013