Technical paper

Quantitation of Carotenoids in Raw and Processed Fruits in Japan

Masamichi YANO+�, Masaya KATO

+, Yoshinori IKOMA+, Akemi KAWASAKI

+, Yoshino FUKAZAWA+,

Minoru SUGIURA+, Hikaru MATSUMOTO

+, Yumiko OOHARA,, Akihiko NAGAO

- and Kazunori OGAWA.

+ Department of Citrus Research Okitsu, National Institute of Fruit Tree Science, .2/�0 Okitsunakacho, Shimizu-ku, Shizuoka-shi,

Shizuoka .,.�*,3,, Japan, Saga Prefectural Fruit Tree Research Station, 3+ Teraura, Ogi-cho, Ogi-gun, Saga 2./�**.+, Japan- National Food Research Institute, ,�+�+, Kannondai, Tsukuba-shi, Ibaraki -*/�20.,, Japan. Okinawa Subtropical Station, Japan International Research Center for Agricultural Sciences, +*3+�+ Maezato, Kawarabaru,

Ishigaki-shi, Okinawa 3*1�***,, Japan. (Present address, Department of Citrus Research Okitsu, National Institute of Fruit Tree

Science, .2/�0 Okitsunakacho, Shimizu-ku, Shizuoka-shi, Shizuoka .,.�*,3,, Japan)

Received June ,+, ,**. ; Accepted October ,1, ,**.

To obtain the quantitative and qualitative data available for estimating the intake of carotenoids from

fruits in Japan, carotenoids were analyzed with reversed phase high-performance liquid chromatography

(HPLC). Ten carotenoids were examined in 1/ raw fruits and +/ processed fruits, all of which were

harvested or purchased in Japan. Phytoene was detected in /2 of 3* fruit samples ; z-carotene, in /* of 3* ;

lycopene, in +- of 3* ; a-carotene, in +2 of 3* ; lutein, in /0 of 3* ; b-carotene, in 2* of 3* ; b-cryptoxanthin, in

02 of 3* ; zeaxanthin, in /2 of 3* ; all-trans-violaxanthin, in // of 3* ; and 3-cis-violaxanthin, in .1 of 3*

samples. Citrus fruits of the mandarin type (Satsuma mandarin and its hybrids, such as tangor) were rich

in b-cryptoxanthin, b-carotene, all-trans-violaxanthin, and 3-cis-violaxanthin ; ‘Star ruby’ grapefruit in

lycopene ; loquat, Japanese persimmon, and peach in b-cryptoxanthin, b-carotene, all-trans-violaxanthin,

and 3-cis-violaxanthin ; mango in b-carotene, all-trans-violaxanthin, and 3-cis-violaxanthin ; acerolas in

phytoene ; passion fruits in z-carotene. Carotenoid levels in common fruits, such as apple, grape, lemon,

pear, strawberry, kiwifruit, cherry, pineapple, and banana, were low.

Keywords : carotenoid, fruit, phytoene, lycopene, b-carotene, b-cryptoxanthin, violaxanthin

Introduction

Carotenoids are derived from phytoene after steps of

dehydrogenation, cyclization, hydroxylation, oxidation,

and epoxidation (Fig. +). They have several health bene-

fits, such as provitamin A activity, scavenging of free

radicals, enhancement of gap junctions, immunomodula-

tion, and modification of enzyme activity involved in

health and disease (Faure et al., +333).

Therefore, a large number of epidemiological studies

have been conducted on the relationship between caro-

tenoid intake or serum-carotenoid levels and the risk of

various chronic diseases (Ford et al., +333 ; Yuan et al., ,**+

; Zeagers et al., ,**+). These studies require that the com-

position of the foods being tested for individual

carotenoids are known. The USDA-NCC Carotenoid Data-

base (Holden et al. +333) provides the most valuable re-

source for this purpose worldwide. We have been inter-

ested in the e#ects of carotenoids, particularly in domes-

tic fruits, on the risk of various chronic diseases in Japan

(Sugiura et al., ,**,a ; Sugiura et al., ,**,b ; Sugiura et al.,

,**.a ; Sugiura et al., ,**.b). The USDA-NCC Carotenoid

Database is a valuable resource for this research. How-

ever, there may be some di#erences in the carotenoid

composition of fruits and vegetables grown in Japan and

those used for the establishment of the USDA-NCC

Carotenoid Database. For example, the characteristics of

cultivars, climate, and conditions for growing fruit in

Japan may produce carotenoid profiles that are di#erent

from those in the USDA-NCC Carotenoid Database. Ac-

curate carotenoid data is important when analyzing the

relationship between carotenoid intake from fruit grown

in Japan and the risk of various chronic diseases. Thus,

the objective of this study was to characterize the

carotenoid content in various raw and processed fruits

consumed in Japan. A total of 1/ fresh fruit and +/

processed fruit samples were collected in Japan. A total

of +* carotenoids, such as phytoene, z-carotene, lycopene,

a-carotene, lutein, b-carotene, b-cryptoxanthin, zeaxanthin,

all-trans-violaxanthin, and 3-cis-violaxanthin were deter-

mined by means of reversed phase HPLC analysis.

Materials and Methods

Carotenoids Carotenoids used in this study are as

* To whom correspondence should be addressed.E-mail : ym0*2,@a#rc.go.jp

Food Sci. Technol. Res., ++ (+), +-�+2, ,**/

follows : a-carotene (a-car), lutein (lut), and all-trans-

violaxanthin (t-vio) were obtained from DHI Water and

Environment (Horshholm, Denmark), lycopene (lyc), b-

carotene (b-car), and zeaxanthin (zea) were obtained from

Extrasynthese (Genay, France), b-cryptoxanthin (b-cry)

was obtained from Sokenkagaku (Tokyo). Phytoene

(phy) and z-carotene (z-car) were isolated from phy-

producing Escherichia coli or z-car-producing E. coli (Kato

et al, ,**.). Nine-cis-violaxanthin (c-vio) was prepared

from the flavedo of Satsuma mandarin (Kato et al., ,**.).

Sampling and Sample Preparation Ripe samples of

raw fruits were harvested from trees at the National

Institute of Fruit Tree Science (Okitsu), Shimizu-Okitsu

Shizuoka-shi Shizuoka, the National Institute of Fruit

Tree Science, Tsukuba-shi Ibaraki, Saga Prefectural Fruit

Tree Research Station, Ogi-cho Saga, and the Okinawa

Subtropical Station, Japan International Research Center

for Agricultural Sciences, Ishigaki-shi, Okinawa, all of

which are in Japan. The rest of the fruit samples were

purchased at local Japanese markets. For each sample, -

to / fruits of average size were selected. Each fruit was

homogenized and / g samples were removed for analysis.

Carotenoid analysis was performed in triplicate.

Carotenoid quantification For carotenoid analysis,

samples were homogenized in .*� (v/v) methanol con-

taining +*� (w/v) magnesium carbonate basic. Pigments

were extracted from the residues using an acetone : meth-

anol (1 : - v/v ) solution containing *.+� (w/v) ,,0-di-tert-

butyl-.-methylphenol and partitioned into diethyl ether

by using NaCl-saturated water. The diethyl ether layer

was mixed with ,*� (w/v) methanolic KOH for saponific-

ation. The sample was shaken and wrapped in alumi-

num foil to protect it from light. Prior to capping, the

sample was gently blanketed with nitrogen, before being

closed and placed in the dark overnight at room tempera-

ture. Water-soluble extracts were removed from the ex-

tracts by adding NaCl-saturated water. The pigments

repartitioned into the diethyl-ether phase were recovered

and dried. Subsequently, the residue was re-dissolved in

/ mL of a methyl tert-butyl ether (MTBE) : methanol (+ : +

v/v ) solution. Percent recoveries of carotenoids under

the condition of this sample preparation procedure were

32.,� for b-car, +*/./� for b-cry and 23.2� for z-car.

An aliquot (,* mL) was separated by a reverse-phase

HPLC system (Jasco) fitted with a YMC Carotenoid S-/

column of ,/*�..0 mm (i.d) (Waters, Milford, MA) at a flow

rate of + mL min�+. The eluent was monitored by a

photodiode array detector (MD-3+*, Jasco). The sample

was analyzed with three di#erent gradient elution sched-

ules. To assay t-vio, c-vio, lut, b-cry, a-car, and lyc the

gradient elution schedule consisted of an initial -* min of

3/� (v/v) methanol, +� (v/v) MTBE, and .� (v/v) water

followed by a linear gradient of 0� (v/v) methanol, 3*�(v/v) MTBE, and .� (v/v) water for 0* min (method A).

To assay z-car, and b-car, the initial solvent composition

consisted of /*� (v/v) methanol, .0� (v/v) MTBE, and

.� (v/v) water followed by a linear gradient of 0� (v/v)

methanol, 3*� (v/v) MTBE, and .� (v/v) water for 0* min

(method B). Zea was assayed by the gradient elution

schedule of Rouse# and Raley (+330). The initial compo-

sition consisted of 3*� (v/v) methanol, /� (v/v) MTBE,

and /� (v/v) water followed by a linear gradient of 3/�(v/v) methanol, /� (v/v) MTBE for +, min, 20� (v/v)

methanol, and +.� MTBE for 2 min, 1/� (v/v) methanol

and ,/� (v/v) MTBE for +* min, and /*� (v/v) methanol

and /*� (v/v) MTBE for ,* min (method C).

The peaks were identified by comparing their specific

retention times and absorption spectra with the authentic

standards. Standard curves for the carotenoid quan-

tification were prepared with those of the authentic

standards at ,20 nm for phy, .** nm for z-car, ./, nm for

t-vio, c-vio, lut, b-cry, a-car, lyc, and zea, and ./- nm for

b-car. The carotenoid concentration was estimated by

the standard curves and expressed as milligrams per +**

g fresh weight.

Results

The carotenoid content of ninety fruit samples was

determined. Qualitative and quantitative data of these

fruit samples are presented in Table +. Phytoene was

detected in /3 of 3* fruit samples. The high phy levels

were found in acerolas (++..12 mg/+** g fresh wt.), dried

apricots (,.2+. mg/+** g fresh wt.) and tangor cv. ‘Youko’

(,./-1 mg/+** g fresh wt.). Zeta-carotene was detected in

/* of 3* fruit samples. High z-car levels were found in

raw passion fruit (,..13 mg/+** g fresh wt.) and passion

fruit juice (+.+*- mg/+** g fresh wt.). Lycopene was

detected in only +- samples. High lyc levels were found

in watermelon (0.+2. mg/+** g fresh wt.), guava (..-2- mg/

+** g fresh wt.), papaya cv. ‘Fruit tower’ (,..2+ mg/+** g

fresh wt.), and grapefruit cv. ‘Star ruby’ (+.203 mg/+** g

fresh wt.). Alfa-carotene was detected in only +2

samples, but no samples showed high levels (�+ mg/+** g

fresh wt.). Similarly, although lut was detected in /2 of

3* samples, no samples showed high levels of lutein (�+

Fig. +. Carotenoid biosynthetic pathway.

M. YANO et al.14

Ta

ble

+.C

arot

enoi

dco

nte

nts

ofra

wan

dp

roce

ssed

fru

its.

Quantitation of Carotenoids in Raw and Processed Fruits in Japan 15

Ta

ble

+.(C

onti

nu

ed)

M. YANO et al.16

mg/+** g fresh wt.). Beta-carotene was detected most

frequently (in 2* of 3* samples), with the highest b-car

levels found in dried apricots (3.0,2 mg/+** g fresh wt.),

red-fleshed melon (0.2*0 mg/+** g fresh wt.), guava (/.*,1

mg/+** g fresh wt.), loquat cv. ‘Fusahikari’ (-.+2+ mg/+** g

fresh wt.), Mexican mango (,.,3/ mg/+** g fresh wt.), and

papaya cv. ‘Sun rise’ (+.32+ mg/+** g fresh wt.). Beta-

cryptoxanthin was the second most frequently detected

carotenoid and occurred in 02 of 3* samples. High b-cry

levels were found in raw papaya cv. ‘Sun rise’ (-.+2, mg/

+** g fresh wt.), Satsuma mandarin grown in a greenhouse

(,.320 mg/+** g fresh wt.), and Satsuma mandarin cv.

‘Yamashita beni’ (+.3.1 mg/+** g fresh wt.). Many citrus

fruits of the mandarin family (Satsuma mandarin and its

hybrids, such as tangors) also showed high b-cry levels.

Although zea was detected in /2 of 3* samples, no samples

showed high levels of zeaxanthin (�+ mg/+** g fresh wt.).

All-trans-violaxanthin (t-vio) was detected in // of 3,

samples. The highest t-vio level was found in the raw

Mexican mango (+./1+ mg/+** g fresh wt.). Nine-cis-

violaxanthin (c-vio) was detected in .1 of 3* samples, but

no samples showed high levels (�+ mg /+** g fresh wt.).

Citrus fruits of the mandarin type had relatively high

b-cry, b-car, t-vio, and c-vio levels (�*.+ mg/+** g fresh

wt.). Sweet oranges (Navel and Valencia) were rich in

t-vio and c-vio, but these carotenoids were not detected in

Valencia orange juice. Other citrus fruits showed poor

carotenoid profiles, except for ‘Star ruby’ grapefruit that

were rich in lycopene. Loquat, Japanese persimmon, and

peach showed relatively high levels of b-cry, b-car, t-vio,

and c-vio (�*.+ mg/+** g fresh wt.). Two tropical fruits,

mango and papaya, showed diverse carotenoid profiles ;

mangoes were rich in b-car and violaxanthin and papayas

were rich in b-car and b-cry. The other tropical fruits,

such as acerola and passion fruit, were surprisingly rich

in phy or z-car. Fruits that were distributed worldwide,

such as apple, grape, lemon, pear, strawberry, kiwifruit,

cherry, pineapple, and banana, had very poor carotenoid

profiles.

In carotenoid-rich fruits, such as the mandarin-type

citrus, loquat, peach, mango, papaya, and guava, several

samples were analyzed to determine di#erences in caro-

tenoid profiles between cultivars and to determine the

e#ect of culture conditions and processing on the caro-

tenoid profiles. Satsuma mandarins that were grown in

a greenhouse and matured in mid summer, interestingly,

accumulated a larger quantity of b-cry than the common

field-produced Satsuma mandarin (,.320 mg versus +..0,

mg/+** g fresh wt.). Among some of the carotenoid-rich

fruits, there were relatively large di#erences in both con-

tent and composition between cultivars. Beta-crypto-

xanthin contents in Satsuma mandarin cultivars ranged

from +.-.2 to +.3.1 mg/+** g fresh wt. Total carotenoid

contents in loquat ranged from *.023 to .../1 mg/+** g

fresh wt. Those in peach ranged from *./,3 to ,./13 mg/

+** g fresh wt., and those in mango ranged from -.-33 to

/.-,0 mg/+** g fresh wt. As for carotenoid composition,

the papaya cultivars could be divided into b-cry-rich and

lyc-rich types. With respect to the e#ect of processing,

both t-vio and c-vio were undetectable in most processed

fruit (European plum, fig, Satsuma mandarin, Valencia

orange, mango, and passion fruit).

Discussion

The aim of this study was to obtain qualitative and

quantitative carotenoid data for fruit samples that could

be used to estimate individual carotenoid intake from

fruits. Since the Japanese consume many fruits that are

specific to Japan and uncommon in other countries, we

wanted to determine the carotenoid levels of the fruits

most frequently consumed in Japan. In addition, since

little is known about the e#ect of the second group of

important carotenoids - phy, z-car, t-vio, and c-vio - on

human health (Kotake-Nara et al., ,**+ ; Nara et al., ,**+),

which may be di#erent from the primary group - lyc,

a-car, b-car, b-cry, lut, and zea - carotenoid levels in these

fruit were investigated. We studied +* carotenoids, in-

cluding phy, z-car, t-vio, and c-vio in ninety fruit samples

that were harvested at four Japanese research institutes

or that were purchased at local Japanese market.

As a broad range of citrus fruits is consumed in Japan

(Kitagawa and Kawada +320) and new citrus cultivars

have been enthusiastically developed, we initially focused

on citrus samples. The carotenoid profiles of the citrus

samples obtained were rich in b-cry. According to the

USDA-NCC Carotenoid Database, the most current and

well-developed database, b-cry is found in significant

levels (�*.+ mg/+** g fresh wt.) in relatively few fruits

and vegetables, such as tangerine, loquat, Japanese per-

simmon, papaya and sweet pepper (red). However, in

this study, we found that +, citrus samples (mostly

Satsuma mandarin and its hybrids, such as tangor), which

are common in Japan, were rich in b-cry (�+ mg/+** g

fresh wt.). In the USDA-NCC Carotenoid Database, the

b-cry levels of tangerine (corresponding to the Japanese

Satsuma mandarin type) and its juice are apparently

lower (*..2/ mg and *.++/ mg/+** g fresh wt. respective-

ly) than the levels of the Japanese Satsuma mandarin

type (ranging from *.-++ to ,.320 mg/+** g fresh wt.) in

this report. Such apparent di#erences in the b-cry levels

of citrus samples in the current study and the USDA-NCC

Carotenoid Database are very interesting and the most

important finding of this study. The di#erence can prob-

ably be attributed to di#erences in the cultivars in-

vestigated. In Japan, the most common fruit is Satsuma

mandarin (Kitagawa and Kawada +320), which is rich in

b-cry and consists of various types (from very early

maturing to very late maturing, and from field-produced

to greenhouse-produced). In addition, new citrus culti-

vars in Japan are mostly the hybrids of Satsuma manda-

rin, which are also rich in b-cry. These reasons may

account for the fact that many b-cry-rich citrus fruits

were detected in this study.

The b-cry concentrations in serum and mother’s milk

are much higher among the Japanese population than in

other countries (Sugiura et al., ,**,a ; Sugiura et al., ,**.a ;

Canfield et al., ,**-). This phenomenon can probably be

attributed to the fact that the Japanese consume larger

Quantitation of Carotenoids in Raw and Processed Fruits in Japan 17

quantities of b-cry from citrus fruits than other popula-

tions do.

The levels of b-car, lyc, b-cry, lut, and zea in some fruit

samples obtained in this study are slightly di#erent from

those listed in the USDA-NCC Carotenoid Database (for

example, lyc : +.203 versus +..0, mg/+** g fresh wt. in red-

fleshed grapefruit, 0.+2. versus ..202 mg/+** g fresh wt. in

watermelon, 0.2*0 versus +./3/ mg/+** g fresh wt. in red-

fleshed melon, and *./,* mg versus +...1 mg/+** g fresh

wt. in Japanese persimmon). These results are likely due

to the di#erence in the cultivars analyzed, and the rela-

tively marked di#erences in the cultivars observed were

among citrus fruits, loquat, peach, mango, papaya, and

guava. The variation in the levels of carotenoids, at least

among the most common cultivars (Lee et al., ,**+),

should be determined so as to obtain more precise esti-

mates of consumption levels.

Quantitative data for phy, z-car, t-vio, and c-vio were

obtained. To the best of our knowledge, this is the

second systematic analysis of the levels of these four

carotenoids. Studies of various carotenoids in Latin

American foods have been reported (Rodriguiz-Amaya

+333). Although the contents of these carotenoids in

fruit samples were not high (mostly�+ mg/+** g fresh

wt.), all four of the carotenoids were found in over /*� of

samples investigated. Therefore, fruits are a good source

of the four carotenoids, and eating large quantities of

fruit certainly results in the intake of large quantities of

phy, z-car, t-vio, and c-vio. Since little is known about

the health benefits of these four carotenoids, they should

be thoroughly investigated. With respect to the bio-

availability of epoxy-carotenoids, such as violaxanthin, a

recently published study ascertained that epoxy-

carotenoids are not absorbed by humans (Barua and

Olson, ,**+).

Fruit maturation is one of the important factors in-

fluencing carotenoids levels. Kato et al. (,**.) and Lee and

Castle (,**+) reported massive accumulation of

carotenoids in some citrus cultivars at the last stage of

maturation. Therefore carotenoid levels must be deter-

mined at the ripening stage for eating. In this study, we

obtained quantitative and qualitative data from such

fruits.

In conclusion, +* carotenoids were assayed by reversed

phase HPLC analysis in 1/ raw fruits and +/ processed

fruits harvested and purchased in Japan. The quantita-

tive and qualitative data obtained will be helpful for

estimating the intake of carotenoids from raw fruit and

fruit products in the Japanese population.

Acknowledgment This study was partially supported by Special

Coordination Funds of the Ministry of Education, Culture,

Sports, Science and Technology, Japan, and the Research and

Development Program for New Bio-industry Initiative of the

Bio-oriented Technology Research Advancement Institution.

ReferencesBarua, A.B. and Olson, J.A. (,**+). Xanthophyll epoxides, unlike

beta-carotene monoepoxides, are not detectibly absorbed by

humans. J. Nutr., +-+, -,+,�-,+/.

Canfield, L.M., Cladinin, M.T., Davies, D.P., Fernadez, M.C.,

Jackson, J., Hawkes, J., Goldman, W. J., Pramuk, K., Reyes, H.,

Sablan, B., Sonobe, T. and Bo, X. (,**-). Multinational study of

major breast milk carotenoids of healthy mothers. Eur. J.Nutr., .,, +--�+.+.

Faure, H., Fayol, V., Galabert, C., Glolier, P., Le Moel, G., Steghens,

J.P., Van Kappel, A. and Nabet, F. (+333). Carotenoids : +. Metab-

olism and physiology. Ann. Biol. Clin., /1, +03�+2-.

Ford, E.S., Will, J.C., Bowman, B.A. and Narayan, K.M. (+333).

Diabetes mellitus and serum carotenoids : findings from the

Third National Health and Nutrition Examination Survey.

Am. J. Epidemiol., +/ ; +.3, +02�+10.

Holden, J.M., Eldridge, A.L., Beecher, G.R., Buzzard, I.M., Bhagwat,

S., Davis, C.S., Douglass, L.W., Gebhardt, S., Haytwitz, D. and

Schakel, S. (+333). Carotenoids content of U.S. food : an update

of the database. J. Food Composition and Analysis., +,, +03�+30.

Kato, M., Ikoma, Y., Matsumoto, H., Sugiura, M., Hyodo, H. and

Yano, M. (,**.). Accumulation of carotenoids and expression of

carotenoid biosynthetic genes during maturation in citrus

fruit. Plant Physiol., +-., 2,.�2-1.

Kitagawa, H. and Kawada K. (+320). Production trends around the

world, Part -. Japan. In “Fresh Citrus Fruits” ed. by W.F.

Wardowski, S. Nagy and W. Grierson. AVI Publishing, Con-

necticut, pp. +.-�+/+.

Kotake-Nara, E, Kushiro, M., Zhang, H., Sugawara, T., Miyashita,

K. and Nagao, A. (,**+). Carotenoids a#ect proliferation of

human prostate cancer cell. J. Nutr., +-+, --*-�--*0.

Lee, S.H. and Castle, W.S. (,**+). Seasonal changes of carotenoid

pigments and color in Hamlin, Earlygold, and Budd Blood

oranges juices. J. Agric. Food Chem., .3, 212�22,.

Nara, E., Hayashi, H., Kotake M., Miyashita, K. and Nagao, A.

(,**+). Acyclic carotenoids and their oxidation mixtures inhib-

it the growth of HL-0* human promyelocytic leukemia cells.

Nutr. Cancer, -3,, ,1-�,2-.

Rodriguiz-Amaya D.B. (+333). Latin American food sources of

carotenoids. Arch Latinoam Nutr., .3 (- suppl +), 1.s-2.s.

Rouse#, R. and Ralay, L. (+330). Application of diode array detec-

tion with a C--* reversed phase column for the separation and

identification of saponified orange juice carotenoids. J. Agric.Food Chem., .., ,+10�,+2+.

Sugiura, M., Kato, M., Matsumoto, H., Nagao, A. and Yano, M.

(,**,a). Serum concentration of b-cryptoxanthin in Japan re-

flects the frequency of Satsuma mandarin (Citrus unshiu Marc.)

consumption. J. Health Sci., .2, -/*�-/-

Sugiura, M., Matsumoto, H. and Yano, M. (,**,b). Cross-sectional

analysis of Satsuma mandarin (Citrus unshiu Marc.) consump-

tion and health status based on a self-administered

quetionaires. J. Health Sci., .2, -00�-03.

Sugiura, M., Matsumoto, H., Kato, M., Ikoma, Y., Yano, M. and

Nagao, A. (,**.a). Multiple linear regression analysis of the

seasonal changes in the serum concentration of b-

cryptoxanthin. J. Nutr. Sci. Vitaminol., /*, +30�,*,.

Sugiura, M., Matsumoto, H., Kato, M., Ikoma, Y., Yano, M. and

Nagao, A. (,**.b). Seasonal changes in relationship between

concentration of b-cryptoxanthin and serum lipid levels. J.Nutr. Sci. Vitaminol., /*, .+*�.+/.

Yuan, J.M., Ross, R.K., Chu, X.D., Gao, Y.T. and Yu, M.C. (,**+).

Prediagnostic levels of serum beta-cryptoxanthin and retinol

predicts smoking-related lung cancer risk in Shanghai, China.

Cancer Epidemiol. Biomarkers Prev., +*, 101�11-.

Zeagers, M.P., Goldbohm R.A. and van den Brandt, P.A. (,**+). Are

retinol, Vitamin C, Vitamin E, folate and carotenoids intake

associated with bladder cancer risk ? Results from the Nether-

lands Cohort Study. Br. J. Cancer, ,2 ; 2/, 312�32-.

M. YANO et al.18