Nutrition and CancerResearch Communication

Serum Levels of Micronutrients, Antioxidants and Total Antioxidant Status Predict Risk of Breast Cancer in a Case Control Study1
(Manuscript received 22 August 2001. Initial review completed 20 September 2001. Revision accepted 16 October 2001) Simon Ching, David Ingram,* Roland Hahnel, John Beilby and Enrico Rossi2 Department of Clinical Biochemistry, PathCentre, Nedlands, Western Australia, 6009, *Breast Centre, Sir Charles Gairdner Hospital and Centre for Molecular Immunology and Instrumentation, University of Western Australia, QE II Medical Centre, Nedlands, Western Australia, 6009

ABSTRACT We performed a case control study to assess the association between serum micronutrient and antioxidant levels and the risk of breast cancer. Newly diagnosed breast cancer cases were recruited before any treatment and matched with controls randomly selected from the electoral roll. Blood samples were collected from 153 breast cancer cases and 151 controls. Serum samples were analyzed for retinol, -tocopherol, lycopene, - and -carotene by HPLC, and total antioxidant status by the Troloxequivalent antioxidant assay. Serum albumin, bilirubin and uric acid levels were also determined. After adjustment for age at menarche, parity, dietary fat and alcohol intake, we observed the following reductions in odds ratios for breast cancer risk comparing the highest with the lowest quartiles: 0.47 [95% condence interval (CI) 0.24, 0.91] for -carotene; 0.53 (CI 0.28, 1.01) for retinol; 0.50 (CI 0.26, 0.97) for bilirubin and 0.47 (CI 0.24, 0.94) for total antioxidant status. We conclude that increased serum levels of -carotene, retinol, bilirubin and total antioxidant status are associated with reductions in breast cancer risk. J. Nutr. 132: 303306, 2002.
KEY WORDS: antioxidants micronutrients case control study humans

breast cancer mainly on the basis of their ability to reduce DNA damage (1). It is possible that together with other compounds in the diet, the total antioxidant status and not individual antioxidants in vivo may be an important protective factor. Because of the difculty in measuring all known antioxidants separately and the interactions among different antioxidant species, several methods have been developed to assess the total antioxidant capacity of human serum (2). Retinoids and carotenoids derived from dietary micronutrients such as vitamin A and -carotene have several potentially anticarcinogenic properties in mammary cancer cell lines, including effects on proliferation, differentiation and apoptosis (3). Commonly recognized diet-derived micronutrients and antioxidants in serum include retinol, -tocopherol, lycopene, and and -carotene. In addition to micronutrients and antioxidants, serum also contains other species with recognized antioxidant activity, such as albumin, uric acid and bilirubin (4). Many previous studies examining the effect of dietary intake or serum levels of micronutrients and antioxidants such as -carotene on the development of breast cancer have had mixed results. For example, a prospective study of breast cancer found that serum levels of retinol, -tocopherol, or - and -carotene were not associated with reductions in breast cancer risk (5), whereas a meta-analysis of 11 studies showed that higher dietary consumption of -carotene was associated with a reduced relative risk (6). The aim of the present case control study was to compare serum levels of individual micronutrients and antioxidants and total antioxidant status in breast cancer patients and matched controls to help determine whether they might play a preventive role in the development of breast cancer. SUBJECTS AND METHODS
Study population. Women referred for management of conrmed breast cancer at either a private clinic or the outpatient clinic of Sir Charles Gairdner Hospital, Perth, Western Australia, were recruited for the study between December 1992 and November 1994. Eligible cases were aged between 30 and 84 y and were residents of the Perth area. Breast cancer cases ineligible for the study were as follows: pregnant women; women with a history of antibiotic use within the previous 6 wk; women with a previous history of breast cancer; women unable to speak or read sufcient English; women undergoing surgery within 3 d of diagnosis; and cases in which a denite diagnosis of breast cancer was not established until the time of surgery. A total of 341 women were diagnosed with breast cancer during the study period. A large number of women were excluded because they did not meet the study criteria. The study protocol stipulated collection of blood and urine samples after diagnosis of breast cancer but before any surgery or other treatment. Serum samples were available from 153 cases. The present data were collected as part of a larger dietary case control study of breast cancer in which 3-d urine specimens were analyzed for lignan and isoavonoid phytoestrogen content (7). Cases were individually matched according to 5-y age group to control women living in the same postal code area and selected randomly from the Perth electoral roll. Matched controls were invited by letter to participate in a dietary study, with no mention of breast cancer being made. Women who reported a personal history of breast cancer were considered ineligible as controls. All subjects were

Downloaded from jn.nutrition.org by guest on September 28, 2011

breast cancer

There is compelling evidence that diet plays a role in the development of breast cancer; despite extensive investigation, however, the substances responsible have yet to be identied. Free radicals generated in vivo are postulated to cause oxidative DNA damage that can result in malignancy. Diet-derived antioxidants have been proposed to be protective against

1 We thank Healthway (Health Promotion Fund of Western Australia) for their nancial support. 2 To whom correspondence should be addressed. E-mail: ric.rossi@health.wa.gov.au.

0022-3166/02 $3.00 2002 American Society for Nutritional Sciences.

303

304

CHING ET AL.

informed of the nature and requirements of the study and provided written consent. In most instances, the same researcher interviewed both the case and matched control and applied a standard questionnaire to obtain demographic, reproductive and life style information. Of the 441 women randomly chosen as controls for the study, 249 did not wish to participate and 45 could not be contacted. Serum samples were available from 151 control subjects. Serum analyses. A single blood sample was drawn by venipuncture and the separated serum stored under light tight conditions at 70C. The median storage time for sera from both cases and controls was 6.6 y. Serum samples were thawed and extracted under dim light conditions and analyzed for retinol, -tocopherol, lycopene and - and -carotene by isocratic reverse-phase HPLC with tocopherol acetate as internal standard (8). The between-run precision CV were 3.5, 3.0, 4.8, 7.6 and 3.5%, respectively. Albumin, bilirubin and uric acid in serum were assayed on an automatic random access biochemistry analyzer (Hitachi 917, Tokyo, Japan). We assessed total antioxidant status by the Trolox-equivalent antioxidant capacity assay (Randox Laboratories, Antrim, UK), which measures the degree of suppression of a stable radical cation by serum (9). Urine analyses. A sample of urine pooled over a 3-d collection period was assayed for urea and ammonia so that a measure of total nitrogen excretion could be obtained as an index of total food intake (10). Urinary urea was measured using an enzymatic rate method on an automatic random access biochemistry analyzer (Hitachi 747). Urinary ammonia was assayed by a glutamate dehydrogenase enzymatic method and was measured on a centrifugal analyzer (Cobas Bio, Roche, Switzerland). Statistical analysis. The following descriptive statistics were obtained: reproductive variables including age at menarche, age at rst full-term birth, parity, months of lactation, and age at menopause were categorized consistently with classications in the published literature (11). Hormone replacement therapy was categorized as current vs. not current use. Family history of breast cancer (rst and second degree), benign breast problems, menopausal status and abortions were classied as yes or no. For women whose menopausal status was unclear, the concentrations of serum follicle stimulating hormone and estradiol were determined by RIA to allow correct categorization. An initial analysis of risk factors in participants showed that age at menarche, parity and dietary fat intake are associated with breast cancer risk and the regression model was therefore adjusted for these variables. Because some studies show that alcohol consumption has an association with breast cancer risk (12) and because this dietary constituent had the potential to interfere with the metabolism of nutrient antioxidants, we included alcohol intake as a confounding variable; thus, the nal regression model adjusted for these four variables. Statistical analysis was by multivariate logistic regression with SPSS statistical software (Chicago, IL) with an odds ratio and its 95% condence interval (CI) used to represent the relative risk for breast cancer. Whether a variable had a signicant effect was judged by a P-value (two-sided) 0.05. Tests for linear trend were done by the tting of a continuous variable.

TABLE 1
Descriptive characteristics of breast cancer cases and controls1
Variable Age, y Age at menarche, y Parity, n Age rst term birth, y Lactation, mo Age at menopause, y Weight, kg Height, cm BMI2, kg/m2 Alcohol, g/d Energy intake, kJ/d Fat intake, g/d Total urine nitrogen, g/d Cases 54.0 (45.062.5) 13.0 (12.014.0) 2.0 (2.03.5) 24.0 (21.027.0) 4.5 (0.515.0) 50.0 (47.053.0) 65.0 (59.071.5) 162.0 (157.0165.0) 24.8 (22.627.8) 2.6 (0.014.1) 7988 (61499612) 69.9 (51.392.1) 14.8 (11.418.1) Controls 54.0 (45.563.0) 13.0 (12.014.0) 3.0 (2.04.0) 25.0 (21.027.0) 7.5 (1.819.0) 50.0 (45.053.0) 65.0 (58.074.0) 162.0 (157.0165.0) 25.0 (22.028.2) 2.9 (0.39.4) 8342 (68309810) 72.0 (60.088.2) 15.9 (12.219.8)

1 Values are medians (interquartile range). 2 BMI, body mass index.

Downloaded from jn.nutrition.org by guest on September 28, 2011

lows: 1.27 (CI 0.69, 2.35) for -tocopherol; 1.10 (CI 0.63, 1.93) for lycopene; 1.09 (CI 0.58, 2.00) for albumin; and 0.70 (CI 0.37, 1.31) for uric acid. Odds ratios could not be calculated for serum -carotene because the levels were too low for meaningful analysis. Table 2 shows the odds ratios for breast cancer risk, both as crude ratios and adjusted for age at menarche, parity, alcohol intake and total fat intake. Increasing serum concentrations of -carotene, retinol, bilirubin and total antioxidant status were associated with signicant reductions in breast cancer risk, with signicant trends through the quartiles. After adjustment for confounding variables, the observed reductions in odds ratios for breast cancer risk comparing the highest with the lowest quartiles were as follows: 0.47 (CI 0.24, 0.91) for -carotene; 0.53 (CI 0.28, 1.01) for retinol; 0.50 (0.26, 0.97) for bilirubin; and 0.47 (0.24, 0.94) for total antioxidant status. DISCUSSION We observed signicant reductions of breast cancer risk with increasing serum concentrations of -carotene, retinol, bilirubin and total antioxidant status. Our ndings may have been inuenced by a low participation rate or the collection of samples immediately after diagnosis of breast cancer and we summarize these sources of potential bias. Controls were recruited from the community with a 50% participation rate. The main reason for nonparticipation was the onerous task faced by the subjects of collecting three consecutive 24-h urine specimens required for another arm of this study while maintaining normal activities. As a result, it is possible that women with an interest in their health and diet would have been more likely to volunteer, although the effect of this self-selection process is unknown. Of the women diagnosed with breast cancer during the study period, a large number were excluded because they did not meet the study criteria. The main reason for exclusion of breast cancer cases from the study was a personal decision by some women to request immediate surgery, which did not allow sufcient time for the study protocol. The other common reason for exclusion was that the nal diagnosis would not be conrmed until the time of their surgery. This would have excluded breast cancer patients on a random rather than systematic basis.

RESULTS The characteristics of cases and controls were similar (Table 1). The nal statistical analysis showed no signicant differences between the case and control groups in age, age at menarche or menopause, parity, age at rst full-term birth, duration of lactation, anthropometric variables or the nutritional variables, i.e., intake of alcohol, total energy, total fat or the percentage of energy from fat. There was also no signicant difference between the cases and controls for the excretion of total nitrogen over a 3-d period. Of the micronutrients and antioxidants analyzed, several did not show signicant associations with breast cancer risk when measured in serum. The odds ratios for analytes with no signicant associations with breast cancer risk obtained by comparing the highest with the lowest quartiles were as fol-

SERUM MICRONUTRIENTS IN BREAST CANCER

305

TABLE 2
Crude and adjusted odds ratios (OR) and 95% condence intervals (95% CI) for risk of breast cancer associated with serum -carotene, retinol, bilirubin and total antioxidant status
Quartiles -Carotene, mol/L 0.4 0.40.7 0.71.1 1.1 Test for trend Retinol, mol/L 1.90 1.902.20 2.202.50 2.50 Test for trend Bilirubin, mol/L 4.1 4.15.5 5.57.5 7.5 Test for trend Total antioxidant status, mmol/L 1.17 1.171.24 1.241.31 1.31 Test for trend Cases, n Controls, n Crude OR (95% CI) Adjusted OR1 (95% CI)

48 48 28 27 51 44 26 30 49 35 34 35 45 36 36 33

35 40 35 41 39 34 36 42 28 41 42 39 24 32 40 47

1.00 0.90 (0.49, 1.64) 0.62 (0.32, 1.20) 0.49 (0.26, 0.93) 1.00 1.01 (0.55, 1.85) 0.57 (0.30, 1.09) 0.56 (0.30, 1.05) 1.00 0.51 (0.27, 0.96) 0.47 (0.25, 0.89) 0.55 (0.29, 1.04) 1.00 0.60 (0.30, 1.16) 0.48 (0.24, 0.93) 0.45 (0.23, 0.88)

1.00 0.90 (0.49, 1.67) 0.57 (0.29, 1.11) 0.47 (0.24, 0.91) P 0.016 1.00 0.99 (0.53, 1.84) 0.49 (0.25, 0.96) 0.53 (0.28, 1.01) P 0.040 1.00 0.45 (0.23, 0.87) 0.44 (0.22, 0.86) 0.50 (0.26, 0.97) P 0.050 1.00 0.52 (0.26, 1.03) 0.41 (0.20, 0.82) 0.47 (0.24, 0.94) P 0.030

Downloaded from jn.nutrition.org by guest on September 28, 2011

1 Adjusted for age at menarche, parity, alcohol intake and total fat intake.

Breast cancer cases who participated had their blood collection at a very stressful time, i.e., immediately after breast cancer diagnosis, which may represent a potential source of bias. These women may have generally eaten less at the time of their blood collection; however, their total urine nitrogen excretion over a 3-d period was not signicantly different from that of the controls (Table 1), suggesting that differences in serum micronutrient and antioxidant values reect differences in the types of foods consumed, not simply a general reduction in food intake after breast cancer diagnosis. Of the micronutrients and antioxidants analyzed, serum -tocopherol, lycopene, albumin and uric acid did not show any signicant associations with breast cancer risk. We cannot explain why some analytes were associated and others not, other than to comment that by testing multiple analytes, there was a possibility for a purely chance association for some of them. The observed reductions in odds ratios for breast cancer risk comparing the highest with the lowest quartiles were uniformly approximately one half that of the lowest quartiles, representing slightly less risk reduction than previously observed in these cases and controls (7) for the urine excretion of the phytoestrogens equol 0.27 (CI 0.10, 0.69) and enterolactone 0.36 (CI 0.15, 0.86). Although phytoestrogens have several potentially anticarcinogenic properties, one of their postulated modes of action is as antioxidants. Taken in conjunction with the present results on serum analytes, this raises the possibility that diets that result in greater urine excretions of equol and enterolactone also increase serum levels of -carotene and retinol. Many studies have examined the association between dietderived micronutrients and antioxidants and the risk of breast cancer, with a variety of results. Recent reports showing no association between dietary intake of -carotene and vitamins

A and E and breast cancer include large prospective studies of Swedish women (13) and postmenopausal women from Iowa (14). Other studies have reported a reduced risk of breast cancer; for example, a meta-analysis of 11 studies (6) comparing high vs. low dietary consumption of -carotene gave a relative risk estimate for breast cancer of 0.82 (CI 0.76, 0.91). In some studies in which protective effects were observed, no single dietary marker could explain the effect. For example, the U.S. Nurses Health Study showed that premenopausal women who consumed ve or more servings per day of fruits and vegetables had lower risks of breast cancer than those who had less than two servings per day (15). Similar ndings were reported in a case control study of premenopausal women from New York (16), which concluded that when the single components evaluated were found together in fruits and vegetables, they had a synergistic effect in reducing breast cancer risk. However, a more recent meta-analysis of eight prospective studies concluded that fruit and vegetable intake during adulthood is not signicantly associated with reduced breast cancer risk (17). An advantage of the present study over other studies of nutrition and breast cancer is that it did not rely solely on dietary recall or records but on direct measurement of serum antioxidants and total antioxidant status. We found that higher serum -carotene levels correlated with a lower risk of breast cancer; however, there is a possibility that the lower levels in the cases are the result of development of the cancer over many years. Previous epidemiologic studies reported mixed ndings for serum -carotene levels. For example, a weak protective effect for serum -carotene levels on risk of cancer in general (18) and breast cancer in particular (19) has been reported, whereas a prospective study of breast cancer risk found no association of serum levels of retinol, -tocopherol and - and -carotene with breast cancer risk (5). There

306

CHING ET AL.
6. Gandini, S., Merzenich, H., Robertson, C. & Boyle, P. (2000) Metaanalysis of studies on breast cancer risk and diet: the role of fruit and vegetable consumption and the intake of associated micronutrients. Eur. J. Cancer 36: 636 646. 7. Ingram, D., Sanders, K., Kolybaba, M. & Lopez, D. (1997) Case-control study of phyto-oestrogens and breast cancer. Lancet 350: 990 994. 8. Thurnham, D. I., Smith, E. & Flora, P. S. (1988) Concurrent liquidchromatographic assay of retinol, alpha-tocopherol, beta-carotene, alpha-carotene, lycopene, and beta-cryptoxanthin in plasma, with tocopherol acetate as internal standard. Clin. Chem. 34: 377381. 9. Miller, N. J., Rice-Evans, C., Davies, M. J., Gopinathan, V. & Milner, A. (1993) A novel method for measuring antioxidant capacity and its application to monitoring the antioxidant status in premature neonates. Clin. Sci. (Lond.) 84: 407 412. 10. Burge, J. C., Chorban, P., McKnight, T, Kyler, M. K. & Flancbaum, L. (1992) Urinary ammonia plus urea as an estimate of total urinary nitrogen in patients receiving parenteral nutrition support. J. Parenter. Enteral Nutr. 17: 529 531. 11. Kelsey, J. L. (1993) Breast cancer epidemiology: summary and future directions. Epidemiol. Rev. 15: 256 263. 12. Longnecker, M. P., Berlin, J. A., Orza, M. J. & Chalmers, T. C. (1988) A meta-analysis of alcohol consumption in relation to risk of breast cancer. J. Am. Med. Assoc. 260: 652 656. 13. Michels, K. B., Holmberg, L., Bergkvist, L., Ljung, H., Bruce, A. & Wolk, A. (2001) Dietary antioxidant vitamins, retinol, and breast cancer incidence in a cohort of Swedish women. Int. J. Cancer 91: 563567. 14. Kushi, L. H., Fee, R. M., Sellers, T A., Zheng, W. & Folsom, A. R. (1996) Intake of vitamins A, C, and E and postmenopausal breast cancer. The Iowa Womens Health Study. Am. J. Epidemiol. 144: 165174. 15. Zhang, S., Hunter, D. J., Forman, M. R., Rosner, B. A., Speizer, F. E., Colditz, G. A., Manson, J. E., Hankinson, S. E. & Willett, W. C. (1999) Dietary carotenoids and vitamins A, C, and E and risk of breast cancer. J. Natl. Cancer Inst. 91: 547556. 16. Freudenheim, J. L., Marshall, J. R., Vena, J. E., Laughlin, R., Brasure, J. R., Swanson, M. K., Nemoto, T. & Graham, S. (1996) Premenopausal breast cancer risk and intake of vegetables, fruits, and related nutrients. J. Natl. Cancer Inst. 88: 340 348. 17. Smith-Warner, S. A., Spiegelman, D., Yaun, S. S., Adami, H. O., Beeson, W. L., van den Brandt, P. A., Folsom, A. R., Fraser, G. E., Freudenheim, J. L., Goldbohm, R. A., Graham, S., Miller, A. B., Potter, J. D., Rohan, T. E., Speizer, F. E., Toniolo, P., Willett, W. C., Wolk, A., Zeleniuch-Jacquotte, A. & Hunter, D. J. (2001) Intake of fruits and vegetables and risk of breast cancer: a pooled analysis of cohort studies. J. Am. Med. Assoc. 285: 769 776. 18. Comstock, G. W., Bush, T. L. & Helzlsouer, K. (1992) Serum retinol, beta-carotene, vitamin E, and selenium as related to subsequent cancer of specic sites. Am. J. Epidemiol. 135: 115121. 19. Toniolo, P., Van Kappel, A. L., Akhmedkhanov, A., Ferrari, P., Kato, I., Shore, R. E. & Riboli, E. (2001) Serum carotenoids and breast cancer. Am. J. Epidemiol. 153: 11421147. 20. Mayer, M. (2000) Association of serum bilirubin concentration with risk of coronary artery disease. Clin. Chem. 46: 17231727. 21. Hunt, S. C., Kronenberg, F., Eckfeldt, J. H., Hopkins, P. N., Myers, R. H. & Heiss, G. (2001) Association of plasma bilirubin with coronary heart disease and segregation of bilirubin as a major gene trait: the NHLBI family heart study. Atherosclerosis 154: 747754.

is similar lack of agreement for a possible association with serum retinol. Recent work supports the concept that bilirubin functions as an antioxidant at concentrations within the serum reference interval and may provide protection against atherosclerosis, coronary artery disease and inammation (20,21). Because no prior hypothesis existed for serum bilirubin levels and cancer, our result requires conrmation by others before its possible importance is assessed. The total antioxidant status as assessed by the Troloxequivalent antioxidant activity assay was a signicant protective factor in this study. In a comparison study of three methods for the measurement of total antioxidant capacity of human serum, the authors commented that this assay suffered from some methodological problems (2) and the differences we observed between cases and controls might be explained as a chance nding, perhaps related to these methodological problems. In summary, our study nds that increased serum levels of -carotene, retinol, bilirubin and total antioxidant status are associated with reductions in breast cancer risk. Our observational associations are not necessarily causal in nature and may result from some unknown dietary characteristic or purely from chance. Nevertheless, there is a possibility that the same diets that gave rise to high levels of urine equol and enterolactone previously observed (7) could also increase serum levels of -carotene and retinol. Identication of the components of such diets may have a future role in dietary interventions designed to reduce the incidence of breast cancer. LITERATURE CITED
1. Ambrosone, C. B. (2000) Oxidants and antioxidants in breast cancer. Antioxid. Redox Signal. 2: 903917. 2. Cao, G. & Prior, R. L. (1998) Comparison of different analytical methods for assessing total antioxidant capacity of human serum. Clin. Chem. 44: 1309 1315. 3. Prakash, P., Krinsky, N. I. & Russell, R. M. (2000) Retinoids, carotenoids, and human breast cancer cell cultures: a review of differential effects. Nutr. Rev. 58: 170 176. 4. Yu, B. P. (1994) Cellular defenses against damage from reactive oxygen species. Physiol. Rev. 74: 139 162. 5. Dorgan, J. F., Sowell, A., Swanson C. A., Potischman, N., Miller, R., Schussler, N. & Stephenson, H. E., Jr. (1998) Relationships of serum carotenoids, retinol, alpha-tocopherol, and selenium with breast cancer risk: results from a prospective study in Columbia, Missouri (United States). Cancer Causes Control 9: 89 97.

Downloaded from jn.nutrition.org by guest on September 28, 2011