Human Nutrition and Metabolism

Human Nutrition and Metabolism

Plasma Lycopene Concentrations in Humans Are Determined by Lycopene Intake, Plasma Cholesterol Concentrations and Selected Demographic Factors1,2 Susan T. Mayne,3 Brenda Cartmel, Fabricio Silva, Chi S. Kim, Barbara G. Fallon, Kenneth Briskin, Tongzhang Zheng, Marianna Baum,* Gail Shor-Posner* and W. Jarrard Goodwin Jr.* Department of Epidemiology and Department of Otolaryngology, Yale University School of Medicine, New Haven, CT, 06520 and the *Department of Otolaryngology and Department of Psychiatry and Behavioral Science, University of Miami, Miami, FL, 33136.

KEY WORDS:

●

lycopene

●

humans

●

plasma

●

determinants

Lycopene is one of the major carotenoids in western diets, accounting for ;50% of carotenoids in human serum (Gerster 1997). Lycopene is a nonprovitamin A carotenoid, concentrated in tomatoes and tomato products. Interest in the compound lycopene has been generated based upon recent observational epidemiologic studies indicating that persons who ingest more lycopene, or who have higher concentrations of lycopene in plasma or in adipose tissue, are at reduced risk of certain chronic diseases, including cancer and coronary heart disease, as reviewed elsewhere (Gerster 1997, Hoffman and Weisburger 1997). For example, Kohlmeier et al. (1997) conducted a multi-center case-control study of antioxidant nutrients in adipose tissue and risk of myocardial infarction in Europe, and found that cases had lower levels of three carotenoids, a-carotene, b-carotene, and lycopene, compared to their matched controls. In multivariate models containing all

●

carotenoids

Downloaded from jn.nutrition.org by guest on May 30, 2013

ABSTRACT Higher plasma lycopene concentrations have been associated with a reduced risk of several chronic diseases. Determinants of lycopene concentrations in humans have received limited attention. We had blood lycopene concentrations and lycopene consumption data available from 111 participants in a two-center cancer prevention trial involving b-carotene and examined determinants of plasma lycopene levels cross-sectionally. The median plasma lycopene level was 0.59 mmol/L (range 0.07–1.79). Low plasma concentrations of lycopene were associated with the following variables in univariate analyses: study site (Florida lower than Connecticut, P 5 0.001), being nonmarried (P 5 0.02), having lower income (P 5 0.003), being nonwhite race/ethnicity (P 5 0.03), having lower dietary lycopene intake (r 5 0.29, P 5 0.002), having lower plasma cholesterol (r 5 0.43, P 5 0.0001) and triglyceride levels (r 5 0.26, P 5 0.005), and consuming less vitamin C (r 5 0.20, P 5 0.03). Women had slightly higher plasma lycopene levels than men (0.65 vs. 0.58 mmol/L; P 5 0.31), despite lower dietary intake of lycopene (1,040 vs. 1,320 mg/d; P 5 0.50). Plasma lycopene levels did not differ in smokers and nonsmokers. In stepwise regression analyses, the determinants of plasma lycopene were plasma cholesterol, dietary lycopene, and marital status; these three variables explained 26% of the variance in plasma lycopene. Relatively few lifestyle and demographic factors were important determinants of plasma lycopene levels, with plasma cholesterol, marital status, and lycopene intake being of greatest importance. J. Nutr. 129: 849 – 854, 1999.

of the carotenoids, only adipose tissue lycopene remained significantly associated with a lower risk of myocardial infarction (odds ratio 5 0.52, 95% CI 0.33– 0.82, comparing the 90th to the 10th percentile). As another example, Giovannucci et al. (1995) reported that male health professionals who consumed higher levels of lycopene-rich foods (tomatoes, tomato sauce, tomato juice, pizza) were at significantly lower risk for the subsequent development of prostate cancer (relative risk 5 0.65, 95% CI 5 0.44 – 0.95 comparing .10 vs. ,1.5 servings/wk). Whereas much is known about the determinants of the carotenoid b-carotene in human plasma, far less is known about the determinants of plasma lycopene. The limited data that do exist suggest that determinants of blood lycopene levels differ from those of blood b-carotene levels. For example, numerous studies have found that smokers have significantly lower blood levels of b-carotene than nonsmokers (Brady et al. 1996, Fukao et al. 1996, Margetts and Jackson 1996, Pamuk et al. 1994, Stryker et al. 1988); this does not seem to be the case for lycopene (Brady et al. 1996, Peng et al. 1995, Ross et al. 1995, Tsubono et al. 1996). Also, consumption of alcoholic beverages was inversely associated with blood

1 Results based on interim analyses of these data presented at Experimental Biology 1996, Washington D.C. [S. T. Mayne, B. Cartmel, B. G. Fallon, C. Friedman, K. Briskin, F. Silva, M. Baum, G. Shor-Posner & W. J. Goodwin Jr. (1996) Determinants of lycopene levels in human plasma. FASEB J. 10: A240.(abs.)] 2 Supported by grants #R01 CA 42101 and CA 64567. 3 To whom correspondence and reprint requests should be addressed.

0022-3166/99 $3.00 © 1999 American Society for Nutritional Sciences. Manuscript received 5 September 1998. Initial review completed 21 October 1998. Revision accepted 21 December 1998. 849

MAYNE ET AL.

850

SUBJECTS AND METHODS Subjects. The study population for this cross-sectional analysis included participants who were part of a randomized, double-blind, placebo-controlled trial. The goal of the trial was to determine whether supplemental b-carotene reduces the incidence of second primary tumors and local recurrences in patients curatively treated for early stage cancers of the oral cavity, pharynx, or larynx. Patients were recruited from two recruitment sites, one based at Yale University and recruiting from the entire state of Connecticut, and the second based at the University of Miami and recruiting from south Florida. Participants in the clinical trial were recruited from 35 hospitals in Connecticut and from 14 hospitals in south Florida. Institutional Review Board approval was obtained from all hospitals from which patients were recruited (49 total hospitals). To be eligible, patients had to have a recently diagnosed Stage I or Stage II squamous cell carcinoma of one of the following sites: tongue, gum or mouth, oropharynx, hypopharynx, pharynx, or larynx. Patients with carcinoma in situ at the above sites were also eligible. Patients had to be between 20 and 75 y of age, have completed their treatments for the first cancer, be considered free of cancer at any site at entry into the trial, have no significant co-morbidities, and not have taken supplements of retinol, b-carotene, vitamin E, or selenium other than multivitamins within the past year. Physician consent was obtained prior to contacting potential participants. Participants were approached for participation by letter and then by phone; those who agreed were subsequently visited in-person by a trained nurse- or physician-interviewer/phlebotomist (usually in the participant’s home), who obtained consent prior to proceeding. Participants were randomized to receive either supplemental b-carotene (50 mg/d; Lurotin, BASF, Parsippany, NJ) or a corresponding placebo. Dietary data collection. One year after randomization, trained interviewers assisted subjects in completing the Block Health Habits/ History Questionnaire v.2.1 (long form), which is a food frequency questionnaire consisting of a list of 98 food items, plus additional questions regarding dietary behaviors (Block et al. 1986). Participants were asked to report on their usual dietary patterns over the past 12 mo. Nutrient intake was computed from the questionnaires using the HHHQ software package (version 3.4, 1995), provided with the

questionnaire by the National Cancer Institute. Nutrient calculations are based upon USDA food composition databases. Subjects were also asked about smoking and drinking habits. Questionnaires were excluded if 10 or more food items were missing, if energy intake was implausible (.20.9 MJ/d, excluding alcohol) or if the questionnaire was missing data on key lycopene-containing foods, such as tomatoes and tomato juice. Phlebotomy and biochemical analyses. The interviewers, who were also trained in phlebotomy, obtained blood samples at the 12 mo interview. Blood was collected into two 10 mL heparinized evacuated tubes. Bloods were kept cold in the dark until the plasma could be separated. Plasma was aliquotted and stored at 270°C pending analysis. Samples from the Miami recruitment site were stored temporarily at 270°C, then shipped frozen to the clinical trial laboratory at Yale, where all samples were analyzed. Plasma lycopene was analyzed by reverse-phase high pressure liquid chromatography as described previously (Mayne et al. 1998). The laboratory participated in the National Institute of Standards and Technology micronutrient measurement proficiency testing program. The coefficients of variation for the lycopene assay averaged ,10%. Plasma cholesterol and plasma triglycerides were analyzed in duplicate by enzymatic assays (Sigma diagnostics, methods #352 and 339, respectively, Sigma, St. Louis, MO). Data analysis. Data were analyzed using PC-SAS software (SAS/STAT version 6; SAS Institute, Cary, NC). For descriptive statistics, median plasma and dietary lycopene were calculated, stratified by several variables. Medians were used because the distribution of plasma lycopene was skewed, and the sample size was relatively small. Wilcoxon rank sum tests and a median test were used to test statistical significance. Pearson’s correlation coefficients were calculated for continuous variables. As the plasma lycopene distribution was not normally distributed, concentrations were also log transformed. Forward stepwise regression analysis, and multiple regression analysis, was used to determine the predictors of plasma lycopene concentrations. Variables evaluated in these models included several dietary variables (i.e., consumption of individual carotenoids; fat; cholesterol; carbohydrate; energy; vitamins A, C, E), serologic variables (i.e., plasma cholesterol, triglycerides, plasma carotenoids, season when blood was drawn), and demographic variables (i.e. sex, site, education, smoking status, income, drinking, race/ethnicity, age, body mass index, religious affiliation, previous cancer site). A P value # 0.05 was considered significant.

RESULTS


concentrations of b-carotene (Brady et al. 1996, Fukao et al. 1996, Stryker et al. 1988); an inverse association with alcohol was observed in some studies of lycopene (Buiatti et al. 1996, Forman et al., 1995), but not others (Ascherio et al. 1992, Brady et al. 1996, Tsubono et al. 1996). Women generally have higher concentrations of many carotenoids, including aand b-carotene (National Center for Health Statistics 1996), but this relationship was also not observed consistently with lycopene (Ascherio et al. 1992, Brady et al. 1996, Michaud et al. 1998). Several studies suggest that lycopene levels are inversely associated with age (Ascherio et al. 1992, Brady et al. 1996, Campbell et al. 1994, Michaud et al. 1998, Peng et al. 1995, Vogel et al. 1997), and positively associated with plasma cholesterol (Ascherio et al. 1992, Brady et al. 1996, Campbell et al. 1994, Michaud et al. 1998, Vogel et al. 1997). The observation that smokers tend to have lower b-carotene levels but not lower lycopene levels than nonsmokers may reflect a specific effect of smoking on b-carotene, may be a consequence of dietary patterns of smokers and nonsmokers with regard to b-carotene versus lycopene, or may reflect both factors. Therefore, studies of determinants of plasma lycopene levels must also consider dietary intake of lycopene. The purpose of this study was to perform a cross-sectional evaluation of a large number of potentially important determinants of blood lycopene levels, including age, gender, smoking, drinking, dietary intake, plasma cholesterol, body mass index, race/ethnicity, seasonality, and marital status, in participants in a cancer prevention intervention trial.

A total of 138 persons in the clinical trial were interviewed at the 12-mo time point and blood successfully obtained by venipuncture. One blood sample was not successfully analyzed for lycopene because of an insufficient sample. Of the 137 persons with plasma lycopene data, a total of 111 (81%) had adequate dietary questionnaires available for analysis. Table 1 shows the demographic characteristics of the 111 participants who had complete dietary questionnaires and blood samples available at year 1 and also the median plasma lycopene level and median dietary lycopene intake in this population, stratified by key demographic factors. The study population was primarily made up of male Caucasians, although the Miami recruitment site provided some ethnic diversity to the study population. Despite a prior diagnosis of oral, pharynx, or larynx cancer (all tobacco-related cancers), one fourth of the participants were still smoking at the oneyear time point. The median age of the study population was 65 y (range 40 –76 y); the median body mass index was 25 kg/m2 (range 15– 41 kg/m2); and 68% of the participants had a prior laryngeal cancer, 23% a prior oral cancer, and 9% a prior pharyngeal cancer. Fifty-eight percent of patients had been treated previously with radiation therapy, 34% with surgery, and 8% with the combination of radiation plus surgery.

PLASMA LYCOPENE DETERMINANTS

851

TABLE 1 Median, 25th and 75th percentiles of plasma and dietary lycopene levels in study participants, by sex, site, marital status, smoking, income, intervention, drinking, and race/ethnicity1 Plasma lycopene (mmol/L)

111

0.59 (0.30, 0.86)

91 20

0.58 (0.28, 0.88) 0.65 (0.45, 0.85)

0.31

1320 (765, 2355) 1040 (805, 2110)

0.50

90 21

0.64 (0.39, 0.89) 0.28 (0.21, 0.52)

0.001***

1470 (945, 2490) 750 (220, 1090)

0.0002***

74 37

0.64 (0.39, 0.88) 0.37 (0.20, 0.79)

0.02*

1305 (835, 2490) 1250 (595, 2055)

0.19

29 82

0.52 (0.25, 1.00) 0.59 (0.32, 0.79)

0.75

1410 (945, 2355) 1210 (775, 2235)

0.41

40 71

0.36 (0.24, 0.73) 0.69 (0.39, 0.91)

0.003**

985 (590, 2305) 1380 (945, 2300)

0.08

62 49

0.59 (0.32, 0.88) 0.54 (0.30, 0.85)

0.81

1270 (685, 2330) 1320 (865, 2135)

0.92

70 41

0.63 (0.29, 0.89) 0.58 (0.32, 0.73)

0.43

1270 (825, 2255) 1320 (690, 2300)

0.96

100 11

0.60 (0.34, 0.88) 0.28 (0.21, 0.70)

0.03*

1390 (870, 2345) 690 (185, 950)

0.0002***

P-value

Dietary lycopene (mg/d)

P-value

1290 (775, 2300)

1 Values are medians with 25th and 75th percentiles in parentheses.

* P # 0.05, ** P # 0.01, *** P # 0.001

The median plasma lycopene level and median dietary lycopene intake in this population were 0.59 mmol/L and 1,290 mg/d, respectively (Table 1). In this stratified, univariate analysis, plasma lycopene concentrations were significantly lower in Florida participants compared to Connecticut participants (P 5 0.001), in those not married (P 5 0.02), in those with lower incomes (P 5 0.003), and in nonwhites compared to whites (P 5 0.03). The subgroups with significantly lower plasma lycopene levels also had significantly lower lycopene intake, with the exception of the nonmarried group (P 5 0.19) and the lower-income group (P 5 0.08). Consistent with our own work (Mayne et al. 1998) and that of others (Nierenberg et al. 1997, Ribaya-Mercado et al. 1995), b-carotene supplementation had no effect on plasma lycopene concentrations (P 5 0.81). Females tended to have slightly higher plasma lycopene levels than males (P 5 0.31). Plasma lycopene levels did not vary significantly by season when blood was drawn (data restricted to Connecticut site only; November–February, March–June, July–October) or education (above versus below median; data not shown). The following variables were significantly correlated with plasma lycopene concentrations, with and without log transformation, in univariate analyses: plasma cholesterol, dietary lycopene, plasma triglyceride, and daily vitamin C intake (Table 2). None of the other dietary, serologic, or demographic variables were significantly correlated with plasma lycopene concentrations. Determinants of plasma lycopene levels based upon stepwise regression analyses are shown in Table 3. Three variables, log plasma cholesterol, marital status, and log dietary lycopene intake, were selected in a forward stepwise regression model, and these three variables explained 26% of the variance in plasma lycopene levels. Two additional variables, study site


Overall Sex Male Female Study Site Connecticut Florida Marital Status Married Not married Smoking Status Smoker Former/Never Income Lower Higher Intervention b-carotene Placebo Drinking Current Ex/Never Race/Ethnicity White Non-Hispanic Other

n

and vitamin C intake, were identified as potential determinants using backward elimination; however, incorporating each into the model increased the R-squared only slightly from 0.26 to 0.269 for site and from 0.269 to 0.274 for site plus vitamin C intake. Therefore, the most parsimonious model includes only the three variables shown in Table 3. A model incorporating 10 variables (log plasma cholesterol, marital status, log dietary lycopene, age, energy intake, study site, race/ethnicity, sex, income, and body mass index) explained only 28% of the variance in plasma lycopene levels.

TABLE 2 Correlation between selected continuous variables and plasma lycopene levels1,2 Plasma lycopene

Plasma cholesterol Log plasma cholesterol Dietary lycopene Log dietary lycopene Plasma triglyceride Dietary vitamin C Age

Log plasma lycopene

r

P-value

r

P-value

0.43 0.42 0.29 0.32 0.26 0.20 20.10

0.0001 0.0001 0.002 0.0007 0.005 0.03 0.28

0.37 0.37 0.25 0.32 0.25 0.21 20.07

0.0001 0.0001 0.007 0.0007 0.01 0.03 0.46

1 n 5 111 2 Body mass index, energy intake, dietary b-carotene, dietary cho-

lesterol, energy from alcohol, dietary retinol, and dietary a-carotene were not correlated with plasma lycopene levels (20.10 , r , 0.10; P $ 0.30).

MAYNE ET AL.

852

TABLE 3 Determinants of plasma lycopene levels: Stepwise regression analyses1 Variable

Forward Stepwise Log plasma cholesterol Marital status Log dietary lycopene

b

SE(b)

Partial r2

Model r2

P-value

0.92 20.33 0.16

0.24 0.12 0.06

0.14 0.08 0.05

0.14 0.22 0.26

0.0001 0.002 0.01

1 Dependent variable 5 log lycopene levels

of the fact that lycopene is predominantly transported in low density lipoproteins; these lipoproteins carry the bulk of cholesterol in the plasma. We also found that subjects who were married had significantly higher lycopene levels than nonmarried subjects, even after adjusting for dietary intake of lycopene and plasma cholesterol concentrations. Married subjects had slightly higher proportions of several of the variables that were associated with higher lycopene levels in the univariate analysis, for example, those with higher incomes were more likely to be married (72% of married had high incomes versus 49% of nonmarried subjects), 82% of Connecticut residents were married compared to 78% of Florida residents, and 77% of married subjects were former/never smokers versus 68% of nonmarried subjects. Thus, the marital status variable seems to be capturing many other variables that are associated with higher plasma lycopene levels. The rather striking observation that the participants from Florida consumed half as much lycopene as participants from Connecticut may be a consequence of regional differences in food preferences or might also reflect racial and ethnic differences in food preferences, as the black and Hispanic participants in the study were mostly from Florida. The difference in intakes is not likely to be due to measurement error of the dietary instrument because blood levels of lycopene were also significantly lower in the study subjects from Florida, despite the fact that all bloods were analyzed in one laboratory in Connecticut. Others have reported that lycopene intakes are lower in African-American men than in Caucasian men, and


We analyzed the foods that contributed the most to lycopene intake in this population. Relatively few foods were important contributors: spaghetti/lasagna/other pasta (grouped together in the Block food frequency questionnaire) were the top food source for 54% of the population, and tomatoes/ tomato juice (also grouped in the questionnaire) were the top food source for 42% of the population. Other food items contributed only minor amounts to the overall dietary lycopene intake (e.g., salsa, red chili sauce were the top food source for 2% of the population, the remaining 2% came from various other foods). Lycopene bioavailability has been noted to be somewhat greater for heat-processed versus unprocessed tomato products (Stahl and Sies 1992); therefore, we ran additional regression models to look at the influence of different food sources of lycopene on plasma lycopene levels. In a regression model containing variables for lycopene from spaghetti/lasagna/other pasta and lycopene from tomatoes/tomato juice, the b-coefficient for the former was 1.7-fold the magnitude of the coefficient for the latter (P values 5 0.06 and 0.1, respectively). Thus, lycopene from heat-processed tomato products was apparently more bioavailable than lycopene from unprocessed tomato products. The relationship between dietary lycopene intake versus plasma lycopene concentration, by study site, is shown in Figure. 1. Study subjects from Florida had notably lower lycopene intakes (P 5 0.0002) and plasma concentrations (P 5 0.001) compared to participants from Connecticut. There was considerable variability in this relationship, with some subjects having relatively high plasma lycopene levels (.0.9 mmol/L) despite relatively low intake (,1,500 mg/d), and others having relatively low plasma lycopene levels (,0.2 mmol/L) despite high intake levels (.2,000 mg/d). The b-coefficient for residents of Florida was somewhat greater than that for residents of Connecticut, suggesting a greater efficiency of carotenoid absorption in Florida residents, possibly due to lower overall intake levels of lycopene (Fig. 1). DISCUSSION These results indicate that relatively few variables, plasma cholesterol, marital status, and dietary lycopene intake, were important determinants of plasma lycopene levels in our study population. The finding that dietary lycopene intake was significantly associated with plasma lycopene concentrations was expected, although some studies have not found a significant association between dietary lycopene intake and plasma concentrations (Coates et al. 1991). The strong association with plasma cholesterol was reported in other studies (Ascherio et al. 1992, Brady et al. 1996, Campbell et al. 1994, Michaud et al. 1998, Vogel et al. 1997), and most likely is a consequence

FIGURE 1 Lycopene intake (mg/d) versus plasma lycopene concentrations (mmol/L) in 111 persons from Connecticut and south Florida. b-coefficient, P value by site are as follows: Connecticut only: b 5 0.00006, P 5 0.11; Florida only: b 5 0.0002, P 5 0.006.

PLASMA LYCOPENE DETERMINANTS

plasma cholesterol and plasma triglyceride levels, and consuming less vitamin C. However, in multivariate analyses, only three of these variables were found to be important determinants of plasma lycopene concentrations: plasma cholesterol, marital status, and dietary lycopene intake. ACKNOWLEDGMENTS We thank Joan Buenconsejo for assistance with the data analysis; Judith Fine, Director of the Rapid Case Ascertainment Shared Resource of the Yale Cancer Center, for patient identification; the nurse and physician interviewers who collected the plasma samples; the 49 collaborating hospitals in Connecticut and Florida; the physicians of the participants; and most of all, the participants for their dedication to this research study.

LITERATURE CITED 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. Block, G., Hartman, A. M., Dresser, C. M., Carroll, M. D., Gannon, J. & Gardner, L. (1986) A data-based approach to diet questionnaire design and testing. Am. J. Epidemiol. 124: 453– 469. 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. Buiatti, E., Munoz, N., Kato, I., Vivas, J., Muggli, R., Plummer, M., Benz, M., Franceschi, S. & Oliver, W. (1996) Determinants of plasma anti-oxidant vitamin levels in a population at high risk for stomach cancer. Int. J. Cancer. 65: 317–322. Campbell, D. R., Gross, M. D., Martini, M. C., Grandits, G. A., Slavin, J. L. & Potter, J. D. (1994) Plasma carotenoids as biomarkers of vegetable and fruit intake. Cancer Epidemiol. Biomarkers Prev. 3: 493–500. Coates, R. J., Eley, J. W., Block, G., Gunter, E. W., Sowell, A. L., Grossman, C. & Greenberg, R. S. (1991) An evaluation of a food frequency questionnaire for assessing dietary intake of specific carotenoids and vitamin E among lowincome black women. Am. J. Epidemiol. 134: 658 – 671. Forman, M. R., Beecher, G. R., Lanza, E., Reichman, M. E., Graubard, B. I., Campbell, W. S., Marr, T., Yong, L. C., Judd, J. T. & Taylor, P. R. (1995) Effect of alcohol consumption on plasma carotenoid concentrations in premenopausal women: A controlled dietary study. Am. J. Clin. Nutr. 62: 131–135. Forman, M. R., Lanza, E., Yong, L. C., Holden, J. M., Graubard, B. I., Beecher, G. R., Melitz, M., Brown, E. D. & Smith, J. C. (1993) The correlation between two dietary assessments of carotenoid intake and plasma carotenoid concentrations: Application of a carotenoid food-composition database. Am. J. Clin. Nutr. 58: 519 –524. Fukao, A., Tsubono, Y., Kawamura, M., Ido, T., Akazawa, N., Tsuji, I., Komatsu, S., Minami, Y. & Hisamichi, S. (1996) The independent association of smoking and drinking with serum b-carotene levels among males in Miyagi, Japan. Int. J. Epidemiol. 25: 300 –306. Gerster, H. (1997) The potential role of lycopene for human health. J. Am. Coll. Nutr. 16: 109 –126. Giovannucci, E., Ascherio, A., Rimm, E. B., Stampfer, M. J., Colditz, G. A. & Willett, W. C. (1995) Intake of carotenoids and retinol in relation to risk of prostate cancer. J. Natl. Cancer Inst. 87: 1767–1776. Hoffmann, I. & Weisburger, J. H. (1997) International symposium on the role of lycopene and tomato products in disease prevention. Cancer Epidemiol. Biomarkers Prev. 6: 643– 645. Kohlmeier, L., Kark, J. D., Gomez-Garcia, E., Martin, B. C., Steck, S. E., Kardinaal, A.F.M., Ringstad, J., Thamm, M., Masaev, V., Riemersma, R., Martin-Moreno, J. M., Huttunen, J. K. & Kok, F. J. (1997) Lycopene and myocardial infarction risk in the EURAMIC study. Am. J. Epidemiol. 146: 618 – 626. Margetts, B. M. & Jackson, A. A. (1996) The determinants of plasma b-carotene: interaction between smoking and other lifestyle factors. Eur. J. Clin. Nutr. 50: 236 –238. Mayne, S. T., Cartmel, B., Silva, F., Kim, C. S., Fallon, B. G., Briskin, K., Zheng, T., Baum, M., Shor-Posner, G. & Goodwin, W. J. Jr. (1998) Effect of supplemental b-carotene on plasma concentrations of carotenoids, retinol and alpha-tocopherol in humans. Am. J. Clin. Nutr. 68: 642– 647. Michaud, D. S., Giovannucci, E. L., Ascherio, A., Rimm, E. B., Forman, M. R., Sampson, L. & Willett, W. C. (1998) Associations of plasma carotenoid concentrations and dietary intake of specific carotenoids in samples of two prospective cohort studies using a new carotenoid database. Cancer Epidemiol. Biomarkers Prev. 7: 283–290. National Center for Health Statistics. (1996) National Health and Nutrition Examination III Biochemical Component. Presented at Experimental Biology. Washington, D.C. Nierenberg, D. W., Dain, B. J., Mott, L. A., Baron, J. A. & Greenberg, E. R. (1997) Effects of 4 y of oral supplementation with b-carotene on serum


some have suggested that this differential intake might contribute to the black-white differential in prostate cancer incidence and mortality in the US (Giovannucci et al. 1995). Subjects with lower incomes also had significantly lower plasma lycopene levels compared to subjects with higher incomes. The variables race/ethnicity, study site, and income could all be operating via a common mechanism, although income was stratified separately for each site because Connecticut has one of the highest per capita incomes in the US. Consistent with reports of others (Ascherio et al. 1992, Brady et al. 1996, Campbell et al. 1994, Michaud et al. 1998, Peng et al. 1995, Vogel et al. 1997), we observed an inverse correlation between plasma lycopene concentrations and age. The correlation we observed between dietary lycopene intake and plasma lycopene concentrations (r 5 0.29) is very similar to what was reported by other investigators (Brady et al. 1996, Forman et al. 1993, Yong et al. 1994). Forman et al. (1993) measured lycopene intake by both food records and food frequency questionnaires in one population and reported that higher correlations with plasma lycopene could be achieved via the use of food records (r 5 0.45– 0.53) as compared to food frequency questionnaires (r 5 0.24 – 0.26). This implies that food frequency questionnaires may have more measurement error than food records; however, the ease of administration, coding, and analysis of food frequency questionnaires accounts for their widespread use in epidemiologic studies. The nutrient database used as well as the questionnaire can impact nutrient estimation and corresponding measurement error. VandenLangenberg et al. (1996) examined the impact of using two different carotenoid databases on estimates of carotenoid intake. Lycopene intake estimates were affected by the choice of the nutrient database, with HHHQ producing lower estimated lycopene intakes than the USDA-NCI database; however, the two databases ranked individuals similarly, and the diet-serum correlations were nearly identical (lycopene r 5 0.29 for HHHQ versus 0.25 for USDA-NCI database). From this, it is possible that lycopene intake in our population is underestimated, but the correlation with plasma values is probably unaffected by the choice of database. The multivariate model we developed, which includes terms for plasma cholesterol, marital status, and dietary lycopene intake, explained 26% of the variance in plasma lycopene concentrations. In comparison, Campbell et al. (1994) reported that a model including the terms age, alcohol, plasma cholesterol, body mass index, energy intake, supplement use, vegetable/fruit intake, and vitamin E intake explained 26% of the variance in plasma lycopene in a population of young, healthy nonsmokers, and Ascherio et al. (1992) reported that a model containing the variables age, plasma cholesterol, plasma triglyceride, alcohol intake, energy intake, and carotene intake explained 38% of the variance in plasma lycopene levels in men but only 9% in women. Better model prediction of plasma lycopene in men versus women has also been reported by others (model R2 of 0.50 in nonsmoking men versus 0.16 in nonsmoking women; Michaud et al. 1998). Thus, a large proportion of the variance in plasma lycopene levels remains unexplained in all of the studies completed to date. This is typical for studies of carotenoids and may reflect the fact that some individuals seem to absorb carotenoids relatively well, whereas others are relatively poor absorbers of carotenoids. The biochemical basis for this is unknown. In conclusion, we found that low plasma concentrations of lycopene were associated with the following variables in univariate analyses: study site (Florida lower than Connecticut), being nonmarried, having a lower income, being nonwhite race/ethnicity, having low dietary lycopene intake, having low

853

854

MAYNE ET AL.

concentrations of retinol, tocopherol, and five carotenoids. Am. J. Clin. Nutr. 66: 315–319. Pamuk, E. R., Byers, T., Coates, R. J., Vann, J. W., Sowell, A. L., Gunter, E. W. & Glass, D. (1994) Effect of smoking on serum nutrient concentrations in African-American women. Am. J. Clin. Nutr. 59: 891– 895. Peng, Y. M., Peng, Y. S., Lin, Y., Moon, T., Roe, D. J. & Ritenbaugh, C. (1995) Concentrations and plasma-tissue-diet relationships of carotenoids, retinoids, and tocopherols in humans. Nutr. Cancer 23: 233–246. Ribaya-Mercado, J. D., Ordovas, J. M. & Russell, R. M. (1995) Effect of b-carotene supplementation on the concentrations and distribution of carotenoids, vitamin E, vitamin A, and cholesterol in plasma lipoprotein and non-lipoprotein fractions in healthy older women. J. Am. Coll. Nutr. 14: 614 – 620. Ross, M. A., Crosley, L. K., Brown, K. M., Duthie, S. J., Collins, A. C., Arthur, J. R. & Duthie, G. G. (1995) Plasma concentrations of carotenoids and antioxidant vitamins in Scottish males: Influences of smoking. Eur. J. Clin. Nutr. 49: 861– 865. Stahl, W. & Sies, H. (1992) Uptake of lycopene and its geometrical isomers is greater from heat-processed than from unprocessed tomato juice in humans. J. Nutr. 122: 2161–2166. Stryker, W. S., Kaplan, L. A., Stein, E. A., Stampfer, M. J., Sober, A. & Willett,

W. C. (1988) The relation of diet, cigarette smoking, and alcohol consumption to plasma b-carotene and alpha-tocopherol levels. Am. J. Epidemiol. 127: 283–296. Tsubono, Y., Tsugane, S. & Gey, K. F. (1996) Differential effects of cigarette smoking and alcohol consumption on the plasma levels of carotenoids in middle-aged Japanese men. Jpn. J. Cancer Res. 87: 563–569. VandenLangenberg, G. M., Brady, W. E., Nebeling, L. C., Block, G., Forman, M., Bowen, P. E., Stacewicz-Sapuntzakis, M. & Mares-Perlman, J. A. (1996) Influence of using different sources of carotenoid data in epidemiologic studies. J. Am. Diet. Assoc. 96: 1271–1275. Vogel, S., Contois, J. H., Tucker, K. L., Wilson, P.W.F., Schaefer, E. J. & LammiKeefe, C. J. (1997) Plasma retinol and plasma and lipoprotein tocopherol and carotenoid concentrations in healthy elderly participants of the Framingham Heart Study. Am. J. Clin. Nutr. 66: 950 –958. Yong, L. C., Forman, M. R., Beecher, G. R., Graubard, B. I., Campbell, W. S., Reichman, M. E., Taylor, P. R., Lanza, E., Holden, J. M. & Judd, J. T. (1994) Relationship between dietary intake and plasma concentrations of carotenoids in premenopausal women: Application of the USDA-NCI carotenoid food-composition database. Am. J. Clin. Nutr. 60: 223–230.
