A Study on Serum Carotenoid Levels in Breast Cancer Patients of Indian Women in Chennai (Madras), India

Two-hundred and six breast cancer cases were histologically confirmed breast cancer diagnoses at the Cancer Institute in Chennai (Madras), India. One-hundred and fifty hospital controls were patients who had cancer at any site other than breast and gynecological organs, and 61 healthy controls were persons accompanying patients in the Cancer Institute. Serum levels of carotenoids such as,* -carotene, lycopene, cryptoxanthin, and zeaxanthin & lutein were determined by HPLC. Serum levels of total carotenes and total carotenoids including ,* carotene, which reflects food intake of colored vegetables and fruits and has a protective role for certain sites of cancer, were significantly lower among breast cancer cases and hospital controls compared to healthy controls, especially in post-menopausal women. Serum carotenoid levels appeared to change with menopausal status. Serum *-carotene levels tended to be lower among breast cancer cases than among hospital controls in premenopausal women. Serum xanthophyll levels were significantly lower among breast cancer cases than among healthy controls in post-menopausal women, but not in premenopausal women. Serum levels of retinol and a -tocopherol among breast cancer cases were not significantly different from those in postmenopausal healthy controls, but were higher than those in hospital controls. Serum estrone levels were significantly higher among breast cancer cases than among healthy controls, but serum levels of estradiol and estriol were not. In conclusion, Indian women with cancer of breast or of other sites might have low intake of green-yellow vegetables rich in fiber and carotenoids such as 9 -carotene and zeaxanthin & lutein. J Epidemiol, 1999 ; 9 : 306-314


A Study on Serum Carotenoid Levels in Breast Cancer Patients of Indian Women in Chennai (Madras), India
Yoshinori Ito 1, Kesavan C. Gajalakshmi 2, Ryuichiro Sasaki 3, Koji Suzuki 1, and Viswanathan Shanta 2 Two-hundred and six breast cancer cases were histologically confirmed breast cancer diagnoses at the Cancer Institute in Chennai (Madras), India. One-hundred and fifty hospital controls were patients who had cancer at any site other than breast and gynecological organs, and 61 healthy controls were persons accompanying patients in the Cancer Institute. Serum levels of carotenoids such as,* -carotene, lycopene, cryptoxanthin, and zeaxanthin & lutein were determined by HPLC. Serum levels of total carotenes and total carotenoids including ,*carotene, which reflects food intake of colored vegetables and fruits and has a protective role for certain sites of cancer, were significantly lower among breast cancer cases and hospital controls compared to healthy controls, especially in post-menopausal women. Serum carotenoid levels appeared to change with menopausal status. Serum *-carotene levels tended to be lower among breast cancer cases than among hospital controls in premenopausal women. Serum xanthophyll levels were significantly lower among breast cancer cases than among healthy controls in post-menopausal women, but not in premenopausal women. Serum levels of retinol and a -tocopherol among breast cancer cases were not significantly different from those in postmenopausal healthy controls, but were higher than those in hospital controls.
Serum estrone levels were significantly higher among breast cancer cases than among healthy controls, but serum levels of estradiol and estriol were not.
In conclusion, Indian women with cancer of breast or of other sites might have low intake of green-yellow vegetables rich in fiber and carotenoids such as 9 -carotene and zeaxanthin & lutein.
J Epidemiol, 1999 ; 9 : 306-314 carotenoid, retinol, a -tocopherol, breast cancer, estrogens Several observational studies have shown that green-yellow vegetables and some colored fruits which are rich in *carotene play a protective role against certain cancers such as those of lung, stomach, and colon 14). The concentration of ,3carotene in the serum is a well known indicator for consumption of carotenoids since * -carotene is the principal component of carotenoids found in green yellow vegetables and some colored fruits 5-7). In almost all epidemiological studies, serum levels of * -carotene and other carotenoids were lower in patients with lung or stomach cancer8-10) In contrast, there were no reports on significant inverse associations between j3 -carotene intake or serum * -carotene levels and the risk of breast cancer among Americans 11-45) and others [16][17][18][19][20]

Subjects
Breast cancer cases in this study were recruited at the Cancer Institute (W.I.A.) in Chennai (Madras), which has a central cancer clinic hospital and cancer registry covering the area. Cases were histologically confirmed breast cancer patients diagnosed at the Institute from July 1992 to February 1993. Controls selected during the same period were of two types. Hospital controls recruited in the same hospital were patients who had cancer at any site (mainly stomach and esophagus) other than breast, thyroid and gynecological organs. Healthy controls were persons accompanying patients with cancer other than in breast, thyroid and gynecological organs. Cases and controls were matched by age to within 5 years. Two-hundred and six breast cancer patients, 150 hospital controls and 61 healthy controls were enrolled in this study. Healthy controls were fewer than one-third of the number of breast cancer cases. Several subjects, including hospital controls, were excluded due to inability to measure serum levels of carotenoids and estrogens in them.

Methods
A fasting serum sample was collected from all study subjects at a luteal menstrual phase. The collected serum was immediately separated from blood cells in Chennai (Madras) and packed on dry ice for analysis of carotenoids and estrogens at a laboratory in Japan.
Biochemical tests were carried out using a biochemical analyzer (SMAC, Technicon Co. Ltd.). Serum estrogen levels were determined by radioimmunoassay 24). Serum levels of estrogens were separated into estrone (E1), estradiol (E2) and estriol (E3). Serum levels of carotenoids, retinol and *-tocopherol were stimulaneously determined using the HPLC method previously reported M. Serum levels of zeaxanthin & lutein and provitamin A were represented as the sum of zeaxanthin and lutein levels, a -carotene, g -carotene and cryptoxanthin levels, respectively. Those of total carotenes, total xanthophylls, and total carotenoids represented the sum of a -and 3 -carotenes and lycopene levels, cryptoxanthin, canthaxanthin and zeaxanthin & lutein levels, and total carotenes and total xanthophylls levels, respectively. Serum levels of lipid peroxides (TBARS) and superoxide dismutase (SOD) activity were estimated using the thiobarbituric acid reaction method 26) and the NADPH cytochrome c reductase-NTB method 27) . Almost subjects measured serum TBARS levels were excluded because serum samples were in small quantities. The information about alcohol consumption and smoking status was collected individually by trained interviewers.
Statistical analysis for the differences between means were tested by the ANOVA statistical method after adjusting for age. The odds ratio on healthy controls against breast cancer cases was calculated by the logistic regression model in controlling of age, which was conducted using the independent variable divided into four equal categories of the analyzed subjects (breast cancer cases and healthy controls) by serum component levels. Statistical analyses for the ANOVA and the logistic regression model were conducted using a StatView statistical package28).

RESULTS
Comparison of serum levels of carotenoids and estrogens between breast cancer cases and controls in Indians.
The age range of cases and controls was 22 yrs to 80yrs. None of healthy controls smoked and drank ( Table 1). The percentages of current smokers and of regular drinkers were less among breast cancer cases than among hospital controls. Serum levels of albumin, uric acid, total cholesterol and breast cancer cases (Table 2). Serum levels of total protein, triglyceride were lower among hospital controls than among uric acid, total cholesterol and triglyceride were not significant TBARS: Thiobarbituric acid-reactive substances; SOD: Superoxide dismutase. #: Data represented as the mean value and standard deviation in parenthesis. Probability was calculated by the ANOVA analyses after adjusting for age. #: Data represented as mean value and standard devianon in parenthesis. Probability was calculated by the ANOVA analyses after adjusting for age. different between breast cancer cases and healthy controls, whereas serum levels of calcium and albumin levels were significantly lower among breast cancer cases and hospital controls compared to healthy controls. Serum TBARS levels tented to be highest among hospital controls, followed by breast cancer cases and healthy controls, while serum SOD activities appeared to be lowest among hospital controls, followed by breast cancer cases. Table 3 shows that mean age of cases and controls were comparable. The mean serum estrone (El) levels were significantly higher among breast cancer cases than among hospital controls and healthy controls. The mean serum estradiol (E2) and estriol (E3) levels among hospital controls were lower than that seen among healthy controls and breast cancer cases.
Serum levels of * -carotene and lycopene among breast cancer cases were not significantly different from those among hospital controls, whereas they were significantly lower among breast cancer cases than among healthy controls. Serum levels of carotenoids such as a-carotene, cryptoxanthin, canthaxanthin and zeaxanthin & lutein were lowest among hospital controls, followed by breast cancer cases and then healthy controls. Those of provitamin A, total carotenes, total xanthophylls and total carotenoids were also lowest among hospital controls, followed by breast cancer cases and healthy controls. Serum retinol levels were not significantly different between breast cancer cases and healthy controls, while those were significantly lower among hospital controls than among breast cancer cases. Serum a -tocopherol levels were lowest among hospital controls, followed by breast cancer cases and healthy controls. Mean age of cases and controls was also comparable in both premenopausal (breast cancer:36.2y, hospital controls:36.1y) and post-menopausal (breast cancer:57.0y, hospital controls:57.0y). The trends of carotenoid levels appeared to differ by menopausal status ( Table 4,5 ). Serum levels of a -tocopherol and carotenoids, except for xanthophylls were significantly lower among breast cancer cases than among healthy controls in premenopausal, but not significantly among hospital controls. In contrast, serum levels of carotenoids, including xanthophylls were significantly lower among breast cancer cases and hospital controls than among healthy controls in post-menopausal. Serum retinol levels were not significantly different between breast cancer cases and healthy controls. Serum levels of E1 and E2 among breast cancer cases were significantly lower in post-menopausal than in premenopausal, as same as those among hospital controls and healthy controls. Moreover, serum El levels in both pre- Table 4. Comparison of serum levels of carotenoids, retinol, a -tocopherol, and estrogens between breast cancer cases and the controls in premenopausal#.
#: Data represented as mean value and standard deviation in parenthesis. Probability was calculated by the ANOVA analyses after adjusting for age. menopausal and post-menopausal were apparently higher among breast cancer cases than among the controls, while serum E2 and E3 levels in premenopausal were not higher among breast cancer cases than among healthy controls.
Serum levels of total cholesterol and triglyceride were not significantly different between breast cancer cases and healthy controls in both premenopausal and postmenopausal, whereas they were lower among hospital controls in postmenopausal.
The odds ratios for a category of healthy controls by each serum carotenoid were obtained significantly low value, as well as by serum levels of albumin, calcium and SOD activity (Table 6). In contrast, those by serum El levels appeared apparently high values, but those by retinol, * -tocopherol, total cholesterol, triglyceride and TBARS were not significantly.

DISCUSSION
The population of the study area remains steady about 4 million in recent years. The most frequent cancer among women in south India is cancer of cervix uteri (44/100,000) followed by breast cancer (21.7/100,000) according to the population based cancer registry in Madras 29). In western India, it is cancer of the breast followed by cancer of the uterine cervix 29, 30). The racial and international differences in breast cancer incidence rates and changes in their rates with migration and time, suggest that diet may influence breast cancer risk. The majority of Indians belong to the Hindu religion and do not eat beef. Eating raw vegetables is uncommon in south India. Cooking practices vary from region to region.
Although reports of the relation between high fat intake and the risk for breast cancer do not agree 14, 17, 2 31-37) there was a report that serum uric acid levels tended to be higher in breast cancer 38). In this study, serum levels of total cholesterol and triglyceride tended to be higher among breast cancer cases than hospital controls, but they were not always higher than among healthy controls. In addition, serum albumin levels were significantly and inversely associated with breast cancer cases by the logistic regression analyses. These results suggest that high intake of animal foods rich in protein might play a protective role for breast cancer in Indian women.
It has been reported that high intake and high serum levels of 9 -carotene diminishes the risk for developing cancer of the lung, stomach and colon 8-10) However, few reports fail to Odds ratio was calculated by the logistic regression model conducted using the independent variable divided into quater parts of serum component levels in healthy controls against breast cancer cases after adjusting for age. demonstrate a significant association between high serum levels of * -carotene and low risk for breast cancer11,12,15-17) Betacarotene intake has been reported to play a preventive role for breast cancer in certain western population studies [21][22][23]. In this study, serum levels of carotenoids, including xanthophllys, were significantly lower among breast cancer cases and inversely associated with breast cancer cases by the logistic regression analyses. Seasonal variation in serum levels of carotenoids were reported to be small; errors related to determining serum levels of carotenoid were found to be small, as well as 40). The variation in serum levels of carotenoids such * -carotene, lycopene, zeaxanthin & lutein and a -tocopherol was less than 6.5% in repeatability and less than 10% of this variation was ascribable to serum storage 25). Moreover, variations throughout the menstrual cycle is less than 9%-12% for ?*-carotene, zeaxanthin & lutein and lycopene 41) Serum levels of carotenoids, especially, *-carotene and zeaxanthin & lutein, were less than one-half of those among healthy controls in Indian women, comparing our results among Japanese female residents 25,42).
High intakes of green yellow vegetables rich in ? -carotene, fiber and minerals such as calcium are well known to elevate serum levels of *-carotene and zeaxanthin & lutein 5-7). The results obtained in this study suggest that low consumption of these foods relate to lower serum levels of * -carotene and calcium among breast cancer cases, especially in postmenopausal women, compared to the healthy controls. The carotenoids such as * -carotene and zeaxanthin play a role as antioxidant 43,44). Serum lipid peroxides, thiobarbituric acid-reactive substances (TBARS), are a biomarker of lipid peroxidation 45). In a case-control study, breast adipose tissue concentrations of some carotenoids such as * -carotene were reported to be significantly lower among breast cancer cases than among controls 46), whereas serum TBARS levels increased a proportion to tumor sizes in young breast cancer patients 47). In this study, serum TBARS levels were the highest among hospital controls, followed by the breast cancer cases and the healthy controls. They inversely corresponded to serum levels of carotenoids and serum SOD activities, substances with protective roles against lipid peroxidation by activated oxygen 48). It seems that high rates of oxidative activities , such as lipid peroxidation in cancerous organs , may prevail among breast cancer cases as compared to healthy controls, as well as to hospital controls.
We obtained similar results in this study as in the literature 49, 50) that serum levels of E1 and E2 were higher among breast cancer patients, especially in premenopausal women , while a study on high dietary fiber intake 51) found decreased serum levels of El and E2 in premenopausal women .
Moreover, the significant inverse relationship between serum zeaxanthin & lutein levels and serum El levels in postmenopausal closely agreed with a report52) of the strong and independent relationships between carotenoid intake or plasma lutein concentration and estrogen receptor status. Dietary fiber intake was inversely associated with the risk of breast cancer 53.54) and with improved survival among breast cancer patients 55). Some components found in vegetables might have a synergistic effect on breast cancer risk since a low risk for breast cancer appeared in studies on dietary intake of vegetables rich in carotenoids and fiber 21-23,53, 54).
In conclusion, our study demonstrated that serum levels of carotenoids such as * -carotene were significantly lower among breast cancer cases, especially in post-menopausal women, it appears that low intake of green yellow vegetables might relate to the risk for breast cancer among Indian women.