Major Dietary Patterns and Risk of Renal Cell Carcinoma in a Prospective Cohort of Swedish Women1

ABSTRACT

Links between specific foods and the risk of renal cell carcinoma (RCC) are not well established. Dietary patterns may be a better predictor of RCC risk. Our aim was to identify and examine major dietary patterns and their relation to the risk of RCC in a large prospective cohort study of Swedish women. Complete dietary information was available from a FFQ from 46,572 women aged 40-76 y at baseline. We conducted factor analysis to identify dietary patterns. Cox proportional hazard models were used to estimate rate ratios (RRs) and 95% CIs. During a mean of 14.3 y of follow-up, we identified 93 cases of RCC. We observed 3 major dietary patterns in the cohort: Healthy (vegetables, tomato, fish, fruits, poultry, whole grains), Western (sweets, processed meat, refined grains, margarine/butter, high-fat dairy products, fried potato, soft drinks, meat) and Drinker (wine, hard liquor, beer, snacks) pattern. Higher Healthy pattern scores were not significantly associated with decreased risk of RCC (highest vs. lowest fertile RR = 0.81; 95% CI 0.45-1.48 and RR = 0.54; 95% CI 0.27-1.10 among women ≤ 65 y). There was a suggestion of an inverse association between the Drinker pattern and RCC risk (RR comparing the 2nd and 3rd with the first tertile, 0.56; 95% CI, 0.34-0.95; and 0.72; 95% CI, 0.42-1.22, respectively, P = 0.08 by WaId test); the association was clearer among women ≤ 65 y (P = 0.02 by Wald test). Our data suggest an inverse association between Drinker pattern and the risk of RCC. J. Nutr. 135: 1757-1762, 2005.

KEY WORDS: * renal cell carcinoma * diet * dietary patterns * factor analysis * cohort study

The worldwide estimated incidence of kidney cancer is ~1.9% of all malignancies (1). In the year 2000, there were 189,000 new cases, 91,000 deaths, and 480,000 persons living with kidney cancer worldwide (1). Renal cell carcinoma (RCC)3 represents >80% of all tumors of the kidney (2) and accounts for 1.5% of all cancer cases in Sweden. The 5-y survival rate is 30-60% (3). Considered risk factors are obesity, smoking, and certain kidney diseases. However, the variation in RCC incidence and mortality found geographically (1), among migrants (4), and by socioeconomic status (5,6), suggests that diet may play a major etiological role. There is suggestive evidence that a Western diet high in meat (7-9) and low in vegetables and fruits (10,11) may promote RCG development, but contradictory results have so far hindered definitive conclusions. Previous studies assessed diet as specific foods or nutrients. Yet individuals consume combinations of foods at a time, and therefore experience effects that are a consequence of the interactions among nutrients from these foods. Associations of diet with the disease may not be captured entirely by examining individual constituents because of these interactions. Recently, a few studies used the broader eating habits or patterns that reflect many dietary exposures working together in evaluation of the association between diet and chronic diseases (12-16). Dietary pattern analysis as suggested by Jacobson and Stanton (17) basically employs “factor analysis” for defining dietary patterns that aggregate interrelated variables (foods) into composite “factors.” These factors represent eating patterns in the study population. Then, cumulative effects of multiple exposures included in a dietary pattern can be sufficiently large to be detectable (18). Furthermore, the utility of analyzing diet as a set of patterns provides an alternative to the use of highly correlated food groups and nutrients simultaneously. The purpose of our study was to examine the relation between major dietary patterns and RCC in a large prospective population-based cohort study of Swedish women.

MATERIALS AND METHODS

Subjects. The Swedish Mammography Cohort (SMC) is a population- based cohort study of 66,651 women living in Vstmanland and Uppsala counties in central Sweden. At baseline (1987-1990), these women filled in a mailed 6-page questionnaire, which they received together with an invitation to mammography screening. The questionnaire included items about diet, weight, height, education, family history of breast cancer, parity, and age at first birth. We excluded women who did not adequately complete the questionnaire (n = 1990) and those with extreme energy intake estimates (n = 793), which probably reflected careless completion of the dietary questionnaire (below or above the mean 3 SD for loge-transformed energy). Through linkage to the Swedish Cancer Registry, we identified and excluded all women with a previous diagnosis of cancer other than nonmelanoma skin cancer (n = 2437). The SMC at baseline included 61,431 women. Ten years later in 1997, a more comprehensive questionnaire was sent to all cohort members who were alive. Information on lifestyle factors that were not asked for at baseline, such as cigarette smoking history, and information about diseases such as hypertension and diabetes was obtained. The response rate for the follow-up questionnaire was 70%. The study was approved by the Ethics Committee at Uppsala University Hospital and by the Regional Ethics Committee of the Karolinska Institute, Sweden.

Dietary assessment. The diet of the participating women was assessed with the use of a self-administered FFQ, which included 67 food items commonly eaten in Sweden. There were 7 open questions about daily consumption of 4 different types of bread, glasses of milk, and 5 types of fat used on sandwiches and in cooking. For other food items, the women were asked how often, on average, per week or per day during the last 6 mo, they had consumed these foods. There were 8 predefined frequency categories of consumption that ranged from “never/seldom” to “4 times/d.” The selected frequency category for each food item was converted to a monthly frequency. For example, a response of “4-6 servings/wk” was converted to 21.4 servings/mo (5 servings/wk). Subjects with >40 missing food items were excluded (n = 463). For energy calculation, we used age- specific portion sizes (ages < 53, 53-65, and >65 y) based on mean values from 5922 d of weighed food records kept by 213 women randomly selected from the study population. Nutrient composition values obtained from the Swedish National Food Administration food data base were used for these calculations (19).

Food groupings. We first classified the FFQ food items into 26 food groups to minimize within-person variations in consumption of individual foods (Table 1). The food grouping scheme was based on the similarity of nutrient profiles or culinary usage of the foods. Some individual food items were preserved either because it was inappropriate to incorporate them into a certain food group (e.g., eggs, margarine, tea) or because they were assumed to represent distinct dietary patterns (e.g., wine, liquor, and 3 types of beer). For the alcoholic beverages (wine, liquor, beer), missing frequency responses were considered as “never/ seldom” consumption. This assumption is based on results from a methodological study showing that 74.1% of the missing answers for alcoholic beverages in the Swedish population are “true-zero” (20). However, to be conservative, we excluded subjects with all 5 missing answers for alcoholic beverages (n = 1731). Subjects with any missing single frequency for food groups other than alcohol were excluded (n = 12,665) because the same methodological study shows that missing answers for commonly consumed foods mean “true-zero” only in a small percentage of responders (e.g., 13.9% for bread, 16.7% for dairy, 18.4% for fruit and 39.2% for vegetables). Thus the study cohort included 46,572 subjects.

TABLE 1

Factor-loading matrix for 3 major dietary patterns derived from the FFQ at baseline (1987-1990) among 46,572 women in the Swedish Mammography Cohort1

Nondietary exposures. Weight and height were self-reported in the questionnaire. BMI was calculated as weight in kilograms divided by the square of height in meters (kg/m^sup 2^). The missing information on BMI (1389 subjects) was replaced with the median BMI (24.09 kg/m^sup 2^). Among the women who completed the 1997 questionnaire, we obtained information on diabetes, hypertension, and smoking status (30,737 women). Women were asked whether they were never, former, or current smokers and how many cigarettes they smoked on average during different decades in life. We calculated pack-years of smoking for “ever smokers” by multiplying a mean lifetime number of packs of cigarettes smoked per day by years of smoking.

Identification of RCC cases and follow-up of the cohort. We identified 93 incident cases of RCC in the study cohort of 46,572 women. These cases were diagnosed in the SMC between the return of the questionnaire (1987-1990) and June 30, 2004 by matching with the Regional Cancer Register that recorded all kidney cancers including RCC (ICD-9 diagnosis code 189.0) in the 2 counties. A methodological study showed that 98% of all cancer cases are recorded in the Swedish Cancer Register (21). Dates of deaths in the cohort were ascertained through the Swedish Death Register, and information about the date of movi\ng out from the study area was obtained by matching the cohort with the continuously updated Swedish Population Register. The completeness of follow-up of the cohort is close to 100% due to the high quality of the Swedish registers.

Statistical methods. After all exclusions due to missing data, the study cohort included 46,572 women at baseline; during follow- up, 93 women were diagnosed with RCC. To identify behavioral food patterns in our study population, we used factor analysis (22) of 26 food groups (expressed as frequency of consumption per month) as previously described in detail (23). We conducted the analysis using the FACTOR procedure in Statistical Analysis System (SAS; release 8; SAS Institute). We used eigenvalues > 1.8 to limit the factors and at the same time better identify meaningful factors (13). The factors were rotated by an orthogonal transformation (Varimax rotation function in SAS). The proportion of variance explained by each factor was calculated by dividing the sum of the squares of the respective factor loadings by the number of variables (i.e., food groups). The factor scores for all 3 patterns for each individual were determined by summing the frequency of each food group weighted by the factor loadings within each pattern (22). The validity of this method, estimated in a subsample of 129 women from the cohort by comparing dietary patterns based on the FFQ with those based on four 1-wk food records, was r = 0.47, 0.41, and 0.73 for Healthy, Western and Drinker pattern, respectively (Spearman correlation) (23); the reproducibility of this method after 1 y, estimated in another subsample of women (n = 212), was higher among younger women ≤ 65 y old (r = 0.66, 0.73, and 0.71, respectively) than among women > 65 y old (r = 0.28, 0.16, and 0.69, respectively).

To characterize the patterns in terms of nutrients, Spearman correlation coefficients were used (because the distributions of nutrient intake were usually skewed) to assess association between intakes of total fat, protein, carbohydrates, dietary fiber, vitamin C, β-carotene, and alcohol and the major dietary patterns.

To quantify the association between dietary patterns and RCC risk, Cox proportional hazards models were used to estimate rate ratios (RR) with 95% CI relating the factors to the occurrence of RCC. Follow-up time for study participants accumulated from the date of mammography (date of entry into the cohort). The follow-up was censored at date of death, date of migration out of the study area, date of diagnosis of RCC, or at the end of the follow-up period (June 30, 2004), whichever occurred first. Cox models were developed that controlled for age, BMI, and energy intake as continuous variables. Energy intake was included in the model because over- or underreporting of dietary items in general may lead to increased extraneous variation.

TABLE 2

Characteristics of dietary patterns by consumption of chosen food groups among 46,572 women in the Swedish Mammography Cohort1

In a subcohort analysis of 30,737 subjects (41 cases), we further controlled for hypertension (yes/no), diabetes (yes/no), and smoking status (quartiles of pack-y ≤ 5.40, 5.41-13.00, 13.01-22, >22), defining “never smokers” as having 0 pack-y.

As a basis for linear trend tests, we used median values for each tertile of factor scores and used them as continuous variables. A Wald test of heterogeneity across strata was used to test the overall effect of each variable.

RESULTS

During 665,981 person-y of follow-up of the SMC, we ascertained 93 cases of RCC. Among the 46,572 participants in the study, we identified 3 major dietary patterns that we named Healthy, Western, and Drinker patterns (Table 1). The Healthy dietary pattern, reflecting the consumption of foods commonly considered to be healthy, contained large amounts of vegetables, tomato, fish, fruits, poultry, and whole grains. This pattern accounted for the most variance in the FFQ (9.1%). The Western pattern reflected mainly consumption of sweets, processed meat, refined grains, margarine/butter, high-fat dairy products, fried potatoes, soft drinks, and meat. Alcoholic beverages (wine, liquor, beer) and snacks contributed heavily to the Drinker pattern. These 3 major patterns accounted for 25% of total variance in the FFQ. Each of the remaining factors explained <6% of the variance (data not shown).

Characteristics of the 3 major dietary patterns in terms of frequency of foods consumed are presented in Table 2. In terms of nutrients, the Healthy dietary pattern was strongly correlated with intakes of vitamin C (Spearman correlation = 0.64, P < 0.0001), β-carotene (r = 0.62, P < 0.0001), and dietary fiber (r = 0.57, P < 0.0001), whereas the Western pattern was highly correlated with intakes of total fat (r = 0.80, P < 0.0001), protein (r = 0.57, P < 0.0001), and carbohydrate (r = 0.57). The Drinker pattern was positively correlated with intake of alcohol (r = 0.68, P < 0.0001), and negatively correlated with intakes of dietary fiber (r = -0.39, P < 0.0001), carbohydrate (r = -0.26, P < 0.0001), and β- carotene (r = -0.16, P < 0.0001). Energy intake was highly correlated with the Western pattern (r = 0.72, P < 0.0001), less with the Healthy pattern (r = 0.41, P < 0.0001), and weakly inversely correlated with the Drinker pattern (r = -0.16, P < 0.0001).

TABLE 3

Major characteristics of the 46,572 women in the Swedish Mammography Cohort1

The baseline characteristics of the study cohort according to categories of dietary pattern scores are shown in Table 3. Women with higher Drinker-pattern scores were younger and less obese than those with the lower Drinker score. We compared characteristics of those included in the current analyses with the original cohort (n = 59,237) to ensure comparability. Characteristics of those included in the current analyses did not differ from the original cohort for age (mean = 50 y for the current analyses vs. 50 y), BMI (mean = 24.09 vs. 24.09 kg/m^sup 2^), and university education (5.2 vs. 4.8%).

TABLE 4

RR with 95% CI for RCC according to the 3 major dietary patterns among all women and those ≤65 y old in the Swedish Mammography Cohort

In age-adjusted and multivariate analysis, there was no association with the Healthy and Western patterns and risk of RCC (Table 4)- The Drinker pattern was associated with a significant 44% and nonsignificant 28% lower risk for RCC in the 2nd and 3rd fertiles compared with the 1st. The Wald test for the overall effect of the Drinker pattern was P = 0.08. No linear trends were observed across tertiles of the 3 major dietary patterns. To explore whether the association of RCC with the Drinker pattern was independent of alcohol intake, we adjusted additionally for wine, liquor, and beer. The resulting association between the Drinker pattern score and RCC risk was attenuated (RR comparing the 2nd and 3rd with the first tertile, 0.60; 95% CI, 0.35-1.02; and 0.90; 95% CI, 0.46-1.77, respectively).

In subcohort analyses, there was no indication of confounding of the observed risk estimates for any of the 3 patterns by smoking, diabetes, and hypertension. For example, in this subcohort, age, BMI, and energy adjusted RRs for the 2nd and 3rd tertiles of the Drinker pattern were 0.50 (95% CI 0.23-1.09) and 0.55 (95% CI 0.25- 1.21), respectively. Further adjustment for cigarette smoking, hypertension, and diabetes did not change these estimates RR = 0.50 (95% CI 0.23-1.09) and RR = 0.54 (95% CI 0.24-1.21), respectively.

We performed analyses in the subgroup of younger women (≤65 y of age) for whom the reproducibility of dietary patterns was higher than for the older women. We observed 27 and 46% nonsignificantly lower risk in the 2nd and highest tertiles of the Healthy dietary pattern compared with the 1st tertile (Table 4). In this younger subgroup, the inverse association with the Drinker pattern was also more pronounced (P-value for Wald test was 0.02).

DISCUSSION

In this study population of middle-aged and elderly Swedish women, we identified 3 major dietary patterns that we named Healthy, Western, and Drinker patterns. There was a suggestion of an inverse association between the Drinker pattern and risk of RCC. We found no significant association between the Healthy and Western dietary patterns and RCC risk in our study.

The first 2 major dietary patterns identified in our data (Healthy and Western) are similar to those identified among American women in the Nurses’ Health Study using the same method (factor analysis) (15). Although dietary pattern analyses should be interpreted with caution because patterns depend on the geographical and cultural (different dietary preferences) as well as methodological variations [sampling, food grouping, number of variables used in factor analysis (15), deciding on the number of factors, the rotations employed], 2 major patterns (Healthy and Western) were common in the American and Swedish women.

High consumption of fruits and vegetables characterized the Healthy pattern in our study. The majority of case-control studies (7,9-11,24-26), although not all (27-29), suggested an inverse association between fruit and vegetable consumption and RCC risk. Two cohort studies (30,31), one of which had only 14 cases (30), did not show a significant inverse association between fruits and vegetables consumption and risk of RCC. In traditional analysis of our cohort (32), we observed that high consumption of some fruits and some vegetables was associated with a significantly decreased risk of RCC. However, total fruit and vegetable consumption only tended (P = 0.46, Wald test) to be associated with decreased risk of RCC (RR = 0.59, 95% CI 0.26-1.34) (32), which is in agreement with the results observed for the Healthy pattern.

Our finding suggesting an inverse association of the Drinker pattern (highly loaded on wine, liquor, and beer) with ris\k of RCC is in agreement with accumulating evidence that alcohol consumption may be associated with decreased risk of RCC among women (11,31,33). In traditional analyses of alcohol intake and specific alcoholic beverages in our cohort, we also observed a significant inverse association showing that drinkers (≥1 servings/wk) had RR of 0.62 (95% CI, 0.41-0.94) compared with those drinking <1 serving/wk (34). The nature of the association between alcohol consumption and RCC is not well understood. Obesity and diabetes were shown to be associated with an increased RCC risk (35). Alcohol consumption improves insulin resistance and might be protective against the development of diabetes in postmenopausal women as was observed in a randomized, controlled trial (36). Furthermore, alcohol, through an increase in HDL and a decrease in LDL (37), may affect the risk of RCC similarly to the cholesterol-lowering statin drugs (38), which have been associated with a 20% decrease in RCC risk (39).

Although some (7-9,28) but not all (40) previous studies that examined the relation between meat intake and RCC reported an increased risk with increasing meat consumption, we did not observe any association with the Western pattern characterized by a high consumption of processed meat. Handa et al. (16) investigated the association between dietary patterns identified by factor analysis and the risk of RCC in a population-based, case-control study including 210 female and 251 male cases. After adjusting for potential confounders (age, smoking, and BMI), a “Desserts” dietary pattern including cakes, cookies, pies, ice cream, chocolate, and chips (a part of our Western pattern) tended to be associated with an increased risk of RCC.

The strengths of our study are its prospective design, the population-based cohort, the large size of the cohort, the long follow-up time, and the completeness of follow-up of the cohort. As we reported previously (23), the validity and reproducibility of our FFQ in terms of identifying the major dietary patterns by using factor analysis are reasonable in general. However, the observed lower reproducibility of the Healthy and Western patterns among older compared with younger women could lead to attenuation of the observed risk estimates. Indeed, in the subgroup of younger women, the inverse association with the Healthy pattern, which tended (P = 0.23, Wald test) to be significant, seemed to be more suggestive.

Although our prospective cohort is large, the major limitation of the present study is low power due to the relatively small number of RCC cases, which is reflected in the wide CI of our estimates. Large numbers of the original cohort members were excluded in the final analyses due to missing data for any single food item. However, a comparison of women included in the present analyses with those in the original cohort indicated that their characteristics were similar.

To our knowledge, this is the first prospective study to examine dietary patterns in relation to RCC risk. Our findings show that dietary patterns reflecting a combination of consumed foods may be important in the etiology of RCC, and the Drinker pattern may be associated with decreased risk. Additional studies or combining data from several cohorts is warranted to confirm our findings.

0022-3166/05 $8.00 2005 American Society for Nutritional Sciences.

Manuscript received 21 December 2004. Initial review completed 20 January 2005. Revision accepted 30 March 2005.

1 Supported by World Cancer Research Fund International, The Swedish Cancer Foundation, and the Swedish Research Council/ Longitudinal Studies.

3 Abbreviations used: RCC, renal cell carcinoma; RR, rate ratio; SMC, Swedish Mammography Cohort.

LITERATURE CITED

1. Ferlay, J., Bray, F., Pisani, P. & Parkin, D. M. (2001) GLOBOCAN 2000: Cancer Incidence, Mortality and Prevalence Worldwide, Version 1.0 CancerBase No. 5. Lyon, France.

2. Mathew, A., Devesa, S. S., Fraumeni, J. F., Jr. & Chow, W. H. (2002) Global increases in kidney cancer incidence, 1973-1992. Eur. J. Cancer Prev. 11: 171-178.

3. Lindblad, P. (2004) Epidemiology of renal cell carcinoma. Scand. J. Surg. 93: 88-96.

4. Hemminki, K. & Li, X. (2002) Cancer risks in second- generation immigrants to Sweden. Int. J. Cancer 99: 229-237.

5. Mellemgaard, A., Engholm, G., McLaughlin, J. K. & Olsen, J. H. (1994) Risk factors for renal cell carcinoma in Denmark. I. Role of socioeconomic status, tobacco use, beverages, and family history. Cancer Causes Control 5: 105-113.

6. Asal, N. R., Risser, D. R., Kadamani, S., Geyer, J. R., Lee, E. T. & Cherng, N. (1988) Risk factors in renal cell carcinoma: I. Methodology, demographics, tobacco, beverage use, and obesity. Cancer Detect. Prev. 11: 359-377.

7. Maclure, M. & Willett, W. (1990) A case-control study of diet and risk of renal adenocarcinoma. Epidemiology 1: 430-440.

8. De Stefani, E., Fierro, L., Mendilaharsu, M., Ronco, A., Larrinaga, M. T., Balbi, J. C., Alonso, S. & Deneo-Pellegrini, H. (1998) Meat intake, ‘mate’ drinking and renal cell cancer in Uruguay: a case-control study. Br. J. Cancer 78: 1239-1243.

9. McLaughlin, J. K., Gao, Y. T., Gao, R. N., Zheng, W., Ji, B. T., Blot, W. J. & Fraumeni, J. F., Jr. (1992) Risk factors for renal- cell cancer in Shanghai, China. Int. J. Cancer 52: 562-565.

10. Yuan, J. M., Gago-Dominguez, M., Castelao, J. E., Hankin, J. H., Ross, R. K. & Yu, M. C. (1998) Cruciferous vegetables in relation to renal cell carcinoma. Int. J. Cancer 77: 211-216.

11. Wolk, A., Gridley, G., Niwa, S., Lindblad, P., McCredie, M., Mellemgaard, A., Mandel, J. S., Wahrendorf, J., McLaughlin, J. K. & Adami, H. O. (1996) International renal cell cancer study. VII. Role of diet. Int. J. Cancer 65: 67-73.

12. Terry, P., Hu, F. B., Hansen, H. & Wolk, A. (2001) Prospective study of major dietary patterns and colorectal cancer risk in women. Am. J. Epidemiol. 154: 1143-1149.

13. Slattery, M. L., Boucher, K. M., Caan, B. J., Potter, J. D. & Ma, K. N. (1998) Eating patterns and risk of colon cancer. Am. J. Epidemiol. 148: 4-16.

14. Hu, F. B. (2002) Dietary pattern analysis: a new direction in nutritional epidemiology. Curr. Opin. Lipidol. 13: 3-9.

15. Fung, T., Hu, F. B., Fuchs, C., Giovannucci, E., Hunter, D. J., Stampfer, M. J., Colditz, G. A. & Willett, W. C. (2003) Major dietary patterns and the risk of colorectal cancer in women. Arch. Intern. Med. 163: 309-314.

16. Handa, K. & Kreiger, N. (2002) Diet patterns and the risk of renal cell carcinoma. Public Health Nutr. 5: 757-767.

17. Jacobson H. N. & Stanton, J. L. (1986) Pattern analysis in nutrition. Clin. Nutr. 5: 249-253.

18. Sacks, F. M., Obarzanek, E., Windhauser, M. M., Svetkey, L. P., Vollmer, W. M., McCullough, M., Karanja, N., Lin, P. H., Steele, P., Proschan, M. A. et al. (1995) Rationale and design of the Dietary Approaches to Stop Hypertension trial (DASH). A multicenter controlled-feeding study of dietary patterns to lower blood pressure. Ann. Epidemiol. 5: 108-118.

19. Bergstrom, L., K. E., Hagman, U., Erikson, H. & Bruce, A (1991) The food composition database KOST: the National Food Administration’s information system for nutritive values of food. Vr Fda 43: 439-447.

20. Hansson, L. M. & Galanti, M. R. (2000) Diet-associated risks of disease and self-reported food consumption: how shall we treat partial nonresponse in a food frequency questionnaire? Nutr. Cancer 36: 1-6.

21. Mattsson, B. & Wallgren, A. (1984) Completeness of the Swedish Cancer Register. Non-notified cancer cases recorded on death certificates in 1978. Acta Radiol. Oncol. 23: 305-313.

22. Kim, J.-O. & Mueller, C. W. (1978) Factor Analysis: Statistical Methods and Practical Issues. Sage Publications, Thousand Oaks, CA.

23. Khani, B. R., Ye, W., Terry, P. & Wolk, A. (2004) Reproducibility and validity of major dietary patterns among Swedish women assessed with a food-frequency questionnaire. J. Nutr. 134: 1541-1545.

24. Godley, P. A. & Escobar, M. A. (1998) Renal cell carcinoma. Curr. Opin. Oncol. 10: 261-265.

25. Negri, E., La Vecchia, C., Franceschi, S., D’Avanzo, B. & Parazzini, F. (1991) Vegetable and fruit consumption and cancer risk. Int. J. Cancer 48: 350-354.

26. Boeing, H., Schlehofer, B. & Wahrendorf, J. (1997) Diet, obesity and risk for renal cell carcinoma: results from a case control-study in Germany. Z. Ernaehrwiss. 36: 3-11.

27. Mellemgaard, A., McLaughlin, J. K., Overvad, K. & Olsen, J. H. (1996) Dietary risk factors for renal cell carcinoma in Denmark. Eur. J. Cancer 32A: 673-682.

28. Chow, W. H., Gridley, G., McLaughlin, J. K., Mandel, J. S., Wacholder, S., Blot, W. J., Niwa, S. & Fraumeni, J. F., Jr. (1994) Protein intake and risk of renal cell cancer. J. Natl. Cancer Inst. 86: 1131-1139.

29. McLaughlin, J. K., Mandel, J. S., Blot, W. J., Schuman, L. M., Mehl, E. S. & Fraumeni, J. F., Jr. (1984) A population-based case-control study of renal cell carcinoma. J. Natl. Cancer Inst. 72: 275-284.

30. Fraser, G. E., Phillips, R. L. & Beeson, W. L. (1990) Hypertension, antihypertensive medication and risk of renal carcinoma in California Seventh-Day Adventists. Int. J. Epidemiol. 19: 832-838.

31. Nicodemus, K. K., Sweeney, C. & Folsom, A. R. (2004) Evaluation of dietary, medical and lifestyle risk factors for incident kidney cancer in postmenopausal women. Int. J. Cancer 108: 115-121.

32. Rashidkhani, B., Lindblad, P. & Wolk, A. (2005) Fruits, vegetables and risk of renal cell carcinoma: a prospective study of Swedish women. Int. J. Cancer 113: 451-455.

33. Parker, A. S., Cerhan, J. R., Lynch, C. F., Ershow, A. G. & Cantor, K. P. (2002) Gender, alcohol consumption, and renal cell carcinoma. Am. J. Epidemiol. 155: 455-462.

34. Rashidkhani, B., Lindblad, P. & Wolk, A. (2004) Alcohol consumption and risk of renal cell carcinoma-a prospective study of Swedish women, Stockholm.

35. Lindblad, P., Chow, W. H., Chan, J., Bergstrom, A., Wolk, A., Gridley, G., McLaughlin, J. K., Nyren, O. & A\dami, H. O. (1999) The role of diabetes mellitus in the aetiology of renal cell cancer. Diabetologia 42: 107-112.

36. Davies, M. J., Baer, D. J., Judd, J. T., Brown, E. D., Campbell, W. S. & Taylor, P. R. (2002) Effects of moderate alcohol intake on fasting insulin and glucose concentrations and insulin sensitivity in postmenopausal women: a randomized controlled trial. J. Am. Med. Assoc. 287: 2559-2562.

37. Criqui, M. H. (1987) The roles of alcohol in the epidemiology of cardiovascular diseases. Acta Med. Scand. 717 (suppl.): 73-85.

38. Expert Panel on Detection, Evaluation, And Treatment of High Blood Cholesterol In Adults (2001) Executive Summary of The Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). J. Am. Med. Assoc. 285: 2486-2497.

39. Graaf, M. R., Beiderbeck, A. B., Egberts, A. C., Richel, D. J. & Guchelaar, H. J. (2004) The risk of cancer in users of statins. J. Clin. Oncol. 22: 2388-2394.

40. Lindblad, P., Wolk, A., Bergstrom, R. & Adami, H. O. (1997) Diet and risk of renal cell cancer: a population-based case-control study. Cancer Epidemiol. Biomark. Prev. 6: 215-223.

Bahrain Rashidkhani,* Agneta kesson,* Per Lindblad,[dagger] and Alicja Wolk*2

* Division of Nutritional Epidemiology, Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden and [dagger] Department of Urology, Sundsvall Hospital, Sundsvall, Sweden

2 To whom correspondence should be addressed.

E-mail: Alicja.Wolk@imm.ki.se.

Copyright American Institute of Nutrition Jul 2005