Macronutrients, fatty acids, cholesterol and prostate cancer risk

Transcription

1 Original article Annals of Oncology 16: , 2005 doi: /annonc/mdi010 Macronutrients, fatty acids, cholesterol and prostate cancer risk E. Bidoli 1 *, R. Talamini 1, C. Bosetti 2, E. Negri 2, D. Maruzzi 3, M. Montella 4, S. Franceschi 5 & C. La Vecchia 2,6 1 Servizio di Epidemiologia e Biostatistica, Centro di Riferimento Oncologico, Aviano (PN); 2 Istituto di Ricerche Farmacologiche Mario Negri, Milan; 3 Unità Operativa di Urologia, Azienda Ospedaliera di Pordenone, Pordenone; 4 Servizio di Epidemiologia, Istituto Tumori Fondazione Pascale, Naples; 6 Istituto di Statistica Medica e Biometria, Università degli Studi di Milano, Milan, Italy; 5 International Agency for Research on Cancer, Lyon Cedex, France Received 6 May 2004; revised 4 August 2004; accepted 18 August 2004 Introduction Background: The role of selected macronutrients, fatty acids and cholesterol in the etiology of prostate cancer was analyzed using data from a case control study carried out in five Italian areas between 1991 and Patients and methods: Cases were 1294 men with incident, histologically confirmed prostate cancer, and admitted to the major teaching and general hospitals of study areas. Controls were 1451 men admitted for acute, non-neoplastic conditions to the same hospital network. Information on dietary habits was elicited using a validated food frequency questionnaire including 78 food groups and recipes. Odds ratios (OR) and 95% confidence intervals (CI) were estimated for increasing levels of nutrient intake. Results: A direct association with prostate cancer was found for starch intake (OR = 1.4 in the highest versus the lowest quintile of intake; 95% CI: ), whereas an inverse association emerged for polyunsaturated fatty acids (OR = 0.8; 95% CI: ). Among polyunsaturated fatty acids, linolenic acid (OR = 0.7; 95% CI: ) and linoleic acid (OR = 0.8; 95% CI: ) were inversely related to prostate cancer. When the six major macronutrients were included in the same model, the adverse effect of high intake of starch and monounsaturated fatty acids was statistically significant together with the protective effect of polyunsaturated fatty acids. Results were consistent in separate strata of age, body mass index and family history of prostate cancer. Conclusions: Starch and monounsaturated fatty acids were directly associated with prostate cancer risk and polyunsaturated fatty acids were inversely associated. Key words: case control study, diet, macronutrients, prostate cancer A possible ecological link between prostate cancer and diet was originally suggested based on international differences in mortality rates and national average intakes of fats [1]. The epidemiological evidence on the relation between prostate cancer and intake of fats remains largely unclear [2], although some prospective and case control studies have tended to display an adverse effect of elevated fat intake. Some case control studies found significantly direct associations between prostate cancer risk and various measures of fat intake that persisted after adjustment for total calories, most notably saturated fat [3, 4], monounsaturated fats [5, 6], and alpha-linolenic fatty acid [7, 8]. A prospective study in the USA showed an association of saturated fat, monounsaturated fat and alpha-linolenic acid *Correspondence to: Dr E. Bidoli, Unit of Epidemiology and Biostatistics, Centro di Riferimento Oncologico, via Pedemontana Occ., 12, Aviano (PN), Italy. Tel: ; Fax: ; epidemiology@cro.it with advanced prostate cancer [9] whereas prospective studies conducted in the Netherlands [10] and Norway [11] showed no association with total fat and saturated fat intakes, while a protection, although not statistically significant, was observed for linolenic fatty acid intake in the Netherlands only. Other studies found no consistent relation between fat and prostate cancer risk [12 14]. Moreover, less is known about other macronutrient intakes such as proteins, carbohydrates or dietary cholesterol and prostate cancer risk. The present paper provides further insight on the relation between prostate cancer and intakes of six macronutrients, including various types of fat, cholesterol and selected fatty acids using data from a large case control study conducted in Italy and based on a validated food frequency questionnaire. Materials and methods The data were derived from a case control study on prostate cancer conducted between 1991 and 2002 in five Italian areas: the provinces of Pordenone, Padua, and greater Milan in northern Italy, Latina in central Italy, and Naples in southern Italy [15, 16]. Cases were 1294 men q 2005 European Society for Medical Oncology

2 (median age 66 years, range 46 74) with incident, histologically confirmed prostate cancer admitted to the major teaching and general hospitals in the areas under surveillance. Controls were 1451 men (median age 63 years, range 46 74) from the same geographical areas and admitted to major hospitals of the study areas for a wide spectrum of acute illnesses unrelated to malignant neoplasms, digestive diseases, or any medical condition associated with long-term modification of diet. Controls were comparable to cases with reference to age (5-year age categories), year of interview, and study center (frequency matching). The main diagnostic categories were traumatic conditions, mostly sprains and fractures (21%); non-traumatic orthopedic disorders, such as lower back and disc disorders (33%); acute surgical conditions, mostly abdominal such as appendicitis or strangulated hernia (17%); and other illnesses, such as eye, ear, nose, skin and dental disorders (29%). The same interview-based structured questionnaire and coding manual were used in each study center. Centrally trained and supervised interviewers identified and questioned patients in hospitals. Interviewing nurses were introduced to patients by attending clinical staff. On average, less than 5% of cases and controls refused to be interviewed. The data were centrally checked for consistency. The questionnaire included information on age, education and other sociodemographic characteristics, anthropometric measures and history of selected diseases. A food frequency questionnaire (FFQ) was employed to assess the usual diet during the 2 years before diagnosis or hospital admission for the controls. The FFQ included 78 foods, food groups or recipes divided into seven sections: (i) bread, cereals and first courses; (ii) second courses (e.g. meat and other main dishes); (iii) side dishes (i.e. vegetables); (iv) fruits; (v) sweets, desserts and soft drinks; (vi) milk, hot beverages and sweeteners; (vii) alcoholic beverages. For vegetables and fruit subject to seasonal variation, consumption in season, and the corresponding duration, were elicited. At the end of each section, 1 or 2 open questions were used to report foods not included in the questionnaire but eaten at least once a week. For 40 food items, the serving size was defined in natural units (e.g. 1 teaspoon of sugar, 1 egg), whereas for the remaining ones, it was defined as small, average, or large with the help of pictures. Dietary supplements were not considered, given their low levels of consumption by this population. Macronutrients, cholesterol and fatty acids were computed using the Italian food composition database [17]. The reproducibility and validity of the FFQ were satisfactory: with reference to reproducibility, the correlation coefficients were 0.6 for proteins, 0.7 for sugars, 0.6 for starch, and between 0.5 and 0.6 for fatty acids and cholesterol [18]; corresponding values for validity were 0.6 for proteins, 0.6 for sugars, 0.7 for starch, and between 0.3 and 0.6 for fatty acids and cholesterol [19]. Statistical analysis Odds ratios (OR), and their corresponding 95% confidence intervals (CI), for increasing levels of nutrient intakes compared to the lowest one, were computed using unconditional multiple logistic regression models [20]. The models included terms for 5-year age categories, study center, education (<7, 7 11 or 12+ years), family history of prostate cancer, and total energy intake (energy from alcohol included). Adjustment for energy was made firstly using the residual model [21], and for comparative purposes, by means of a fully partitioned model, in order to allow for the mutual confounding effect of major macronutrients [22]. When nutrients were entered in the model as quintile of intake in the residual model, these were based on the distribution of cases and controls combined. The test for trend was based on the likelihood-ratio test between the models with and without a linear term for each nutrient s quintile. Results Table 1 shows the distribution of prostate cancer cases and control subjects according to age, center, education, and other selected variables. By design, cases and controls had similar age distribution. Cases tended to be significantly more Table 1. Distribution of 1294 cases of prostate cancer and 1451 controls according to age, residence and selected covariates, Italy, Prostate cancer Controls No. (%) No. (%) Age (years) (0.5) 12 (0.8) (4.4) 121 (8.3) (12.1) 298 (20.5) (24.0) 359 (24.7) (32.4) 364 (25.1) (26.7) 297 (20.5) Center Pordenone 709 (54.8) 753 (51.9) Gorizia 191 (14.8) 203 (14.0) Milan 166 (12.8) 196 (13.5) Naples 129 (10.0) 185 (12.7) Latina 99 (7.7) 114 (7.9) Body Mass Index (kg/m 2 ) b < (23.3) 368 (25.4) (26.8) 356 (24.6) (25.1) 365 (25.2) >_ (24.7) 358 (24.7) Education (years) a,b <7 636 (49.6) 844 (58.5) (29.9) 407 (28.2) >_ (20.5) 192 (13.3) Occupational physical activity at age 30 a,b Very active 518 (40.2) 684 (47.1) Moderately active 335 (26.0) 393 (27.1) Inactive 437 (33.9) 374 (25.8) Family history of prostate cancer a,c No 1204 (93.0) 1423 (98.1) Yes 90 (7.0) 28 (1.9) Total energy intake (kcal/day) a < (14.8) 290 (20.0) (19.3) 290 (20.0) (22.3) 290 (20.0) (21.7) 290 (20.0) >_ (21.8) 291 (20.0) a Significant difference between cases and controls (P < 0.01) (adjusted for age and study center). b The sum does not add up to the total because of some missing values. c First-degree relatives.

3 154 Table 2. Odds ratios (OR) a of prostate cancer and corresponding 95% confidence intervals (CI) according to intake of macronutrients, selected fatty acids, and cholesterol, Italy, Nutrient Mean b Standard deviation Quintile, OR (95% CI) 1 (lowest) c x 2 trend (P value) OR (continuous) d Macronutrients Proteins (g) ( ) 1.2 ( ) 1.2 ( ) 1.0 ( ) 0.46 (0.50) 1.10 ( ) Sugars (g) ( ) 1.1 ( ) 1.0 ( ) 1.1 ( ) 0.40 (0.53) 1.02 ( ) Starch (g) ( ) 1.3 ( ) 1.2 ( ) 1.4 ( ) 4.79 (0.03) 1.07 ( ) Total fat (g) ( ) 1.1 ( ) 1.0 ( ) 1.0 ( ) 0.06 (0.81) 1.04 ( ) Fat from animal sources (g) ( ) 1.0 ( ) 0.9 ( ) 0.9 ( ) 0.64 (0.42) 0.99 ( ) Fat from vegetable ( ) 1.0 ( ) 1.0 ( ) 1.0 ( ) 0.03 (0.86) 1.04 ( ) sources (g) Saturated fatty ( ) 1.1 ( ) 1.1 ( ) 0.9 ( ) 0.16 (0.69) 1.03 ( ) Unsaturated fatty ( ) 0.9 ( ) 1.0 ( ) 1.0 ( ) 0.10 (0.75) 1.04 ( ) Monounsaturated fatty ( ) 1.1 ( ) 1.1 ( ) 1.3 ( ) 1.02 (0.31) 1.03 ( ) Polyunsaturated fatty ( ) 1.0 ( ) 0.8 ( ) 0.8 ( ) 6.09 (0.01) 0.87 ( ) Fatty acids and cholesterol Oleic acid (g) ( ) 1.1 ( ) 1.1 ( ) 1.3 ( ) 1.24 (0.27) 1.11 ( ) Linoleic acid (g) ( ) 1.1 ( ) 0.9 ( ) 0.8 ( ) 5.68 (0.02) 0.87 ( ) Linolenic acid (g) ( ) 1.1 ( ) 1.0 ( ) 0.7 ( ) 8.92 (0.003) 0.84 ( ) Other polyunsaturated fatty ( ) 1.1 ( ) 1.3 ( ) 1.4 ( ) 4.88 (0.03) 1.07 ( ) Cholesterol (mg) ( ) 1.1 ( ) 1.0 ( ) 0.9 ( ) 1.31 (0.25) 0.93 ( ) a Estimates from multiple logistic regression models including terms for age (5-year age categories), study center, education, family history of prostate cancer and total energy intake, according to the residual model. b Among controls, per day. c Reference category. d OR relative to 1 SD among controls. educated than controls, to have a lower occupational physical activity at age 30 together with a more frequent family history of prostate cancer. Cases reported higher energy intake than controls and, therefore, subsequent analyses of various nutrients were adjusted for energy intake. Table 2 gives the mean daily intake among controls of six macronutrients, cholesterol and fatty acids, and the ORs of prostate cancer according to quintile of intake, and continuously for an increment of intake equal to 1 SD among controls. Mean daily intake of macronutrients among controls was 27 g for saturated fatty acids, 38 g for monounsaturated fatty acids, 14 g for polyunsaturated fatty acids, 90 g for proteins, 86 g for sugars, and 197 g for starch. Cholesterol intake was about 324 mg/day. Unsaturated fatty acids represented almost two-thirds of the average fat intake among controls. Proteins, sugars, total fat, fat from vegetable and animal sources, saturated fatty acids and monounsaturated fatty acids appeared unrelated to prostate cancer risk. High starch intake was significantly related to an increased risk of prostate cancer (OR = 1.4 in the highest versus the lowest quintile of intake; 95% CI: ), but the pattern of association was not perfectly linear, whereas polyunsaturated fatty acids were inversely related to prostate cancer with an OR of 0.8 (95% CI: ). Among specific fatty acids, oleic acid was unrelated to prostate cancer, linolenic (OR = 0.7; 95% CI: ) and linoleic (OR = 0.8; 95% CI: ) acids were inversely associated to prostate cancer risk. Cholesterol was not related to prostate cancer risk. The other polyunsaturated fatty acids group, which displayed a very low intake, presented an irregular protective effect. The relation between macronutrients and prostate cancer risk was further examined in separate strata of age (<60, and >_ 70 years), body mass index (BMI) (tertiles) and family history (No/Yes) of prostate cancer (Table 3). Although differences in the estimated coefficients were observed across strata, these were compatible with the effect of random variation, since heterogeneity tests were not significant. In particular, favorable influence of polyunsaturated fats was consistent across strata of age and BMI, and the direct association with starch intake was apparently stronger among the older age (>_ 70 years) group. Likewise, no consistent heterogeneity was observed in separate strata of energy intake, education, stage (1 2 and 3 4) or Gleason score (<7 and >_7) (data not shown). When a fully partitioned model was computed, the ORs relative to 100 kcal per day were statistically significant for

4 155 Table 3. Odds ratios (OR) a,b of prostate cancer and corresponding 95% confidence intervals (CI) according to intake of macronutrients, selected fatty acids, and cholesterol, Italy, Nutrient Age (years) BMI (kg/m 2 ) (tertiles) Family history < >_ 70 < >_ 27.7 No Yes Macronutrients Proteins (g) 1.5 ( ) 1.0 ( ) 0.9 ( ) 0.9 ( ) 1.2 ( ) 1.1 ( ) 1.1 ( ) 0.7 ( ) Sugars (g) 0.8 ( ) 1.4 ( ) 0.8 ( ) 0.9 ( ) 1.2 ( ) 1.3 ( ) 1.2 ( ) 0.8 ( ) Starch (g) 1.5 ( ) 1.3 ( ) 1.9 ( ) 1.3 ( ) 1.7 ( ) 1.4 ( ) 1.4 ( ) 1.6 ( ) Total fat (g) 1.2 ( ) 1.1 ( ) 0.8 ( ) 0.9 ( ) 1.4 ( ) 0.9 ( ) 1.0 ( ) 1.4 ( ) Fat from animal sources (g) 1.3 ( ) 0.9 ( ) 0.6 ( ) 0.8 ( ) 0.9 ( ) 1.2 ( ) 0.9 ( ) 0.9 ( ) Fat from vegetable sources (g) 1.1 ( ) 1.1 ( ) 0.9 ( ) 0.9 ( ) 1.6 ( ) 0.8 ( ) 1.1 ( ) 0.5 ( ) Saturated fatty 1.0 ( ) 0.9 ( ) 0.7 ( ) 0.7 ( ) 1.1 ( ) 1.1 ( ) 0.9 ( ) 1.6 ( ) Unsaturated fatty 1.0 ( ) 1.1 ( ) 0.9 ( ) 0.9 ( ) 1.4 ( ) 0.8 ( ) 1.0 ( ) 0.7 ( ) Monounsaturated fatty 1.9 ( ) 1.3 ( ) 0.8 ( ) 1.1 ( ) 1.9 ( ) 1.0 ( ) 1.3 ( ) 0.5 ( ) Polyunsaturated fatty 0.8 ( ) 0.9 ( ) 0.6 ( ) 0.6 ( ) 0.8 ( ) 0.9 ( ) 0.8 ( ) 1.1 ( ) Fatty acids and cholesterol Oleic acid (g) 1.6 ( ) 1.3 ( ) 0.9 ( ) 1.2 ( ) 2.0 ( ) 0.9 ( ) 1.3 ( ) 0.5 ( ) Linoleic acid (g) 0.7 ( ) 0.9 ( ) 0.6 ( ) 0.5 ( ) 0.8 ( ) 0.9 ( ) 0.8 ( ) 1.0 ( ) Linolenic acid (g) 0.8 ( ) 0.8 ( ) 0.5 ( ) 0.5 ( ) 0.7 ( ) 0.9 ( ) 0.7 ( ) 1.1 ( ) Other polyunsaturated fatty 1.6 ( ) 1.3 ( ) 1.4 ( ) 1.2 ( ) 1.3 ( ) 1.6 ( ) 1.3 ( ) 1.7 ( ) Cholesterol (mg) 1.4 ( ) 0.9 ( ) 0.8 ( ) 0.9 ( ) 0.9 ( ) 1.1 ( ) 0.9 ( ) 3.1 ( ) a Estimates from multiple logistic regression models including terms for age (5-year age categories), study center, education, family history of prostate cancer and total energy intake, according to the residual model. b ORs of highest quintile of intake versus lowest. BMI, body mass index. Table 4. Odds ratios (OR) a of prostate cancer and corresponding 95% confidence interval (CI) relative to 100 kcal/day for major macronutrients, Italy, OR 95% CI Proteins 1.08 ( ) Sugars 1.03 ( ) Starch 1.07 ( ) Saturated fatty acids 0.98 ( ) Monounsaturated fatty acids 1.11 ( ) Polyunsaturated fatty acids 0.84 ( ) a Estimated from multiple logistic regression models including terms for age (5-year age categories), study center, education, family history of prostate cancer and all the above nutrients. starch (OR = 1.07; 95% CI: ), monounsaturated fatty acids (OR = 1.11; 95% CI: ) and polyunsaturated fatty acids (OR = 0.84; 95% CI: ), whereas saturated fatty acids (OR = 0.98; 95% CI: ) were unrelated to prostate cancer (Table 4). Discussion This study, one of the largest case control investigations of diet and prostate cancer to date, showed that starch was directly associated with prostate cancer risk in this Italian population. Conversely, polyunsaturated fatty acids seemed to exert a favorable effect. Monounsaturated fats showed a direct association with risk of prostate cancer in the fully partitioned model. Intakes of proteins, sugars, saturated fat, and dietary cholesterol were unrelated to risk. The role of starch has not been described in previous studies. We did not confirm the possible effect of total fat but only of some of its components. Fat intake is the most studied dietary factor although its effect is still unclear among studies [2]. In particular, total or animal fat have been associated with prostate cancer risk in some studies [3 5, 9, 23 27] but not in others [11 14]. At least part of the effect of starch intake should, however, still be a real one. Not only energy was adjusted for by means of the residual method [21] and partition model [22], but dietary findings were similar among strata of energy intake. The direct association with starch is of specific interest, in consideration of the fact that the Italian population shows the highest intake of this macronutrient among western countries [28]. The main sources of starch were white bread, pasta, rice, crackers and cookies whereas sugars derived mainly from fruits and table sugar. The putative association of starch intake with prostate cancer is difficult to explain. Starch may be related to a reduced intake of beneficial substances inversely related to prostate cancer risk (e.g. fibers or selected micronutrients) [15, 16] or be responsible for a glycemic overload, compensated by

5 156 an increase in serum insulin and in insulin-like growth-1 factor (IGF-1) associated with prostate cancer [29]. This study lends further support to the existence of differences between various types of fat, rather than total fat itself, with respect to prostate cancer risk [2]. A significant protective association was displayed by polyunsaturated fatty acids, in particular linoleic and linolenic fatty acids which in this Italian population largely derive from olive oil [30]. Previous studies found an adverse effect of this type of fats [7 9, 27, 31] or a non-significant though protective association of linolenic fatty acid [10]. When the six macronutrients were entered simultaneously in the fully partitioned model, the significant effect of starch and polyunsaturated fats persisted whereas monounsaturated fats resulted at risk. Some studies [5, 6, 9], though not all [3, 13, 32], found a link between monounsaturated fats and cancer risk. A number of biological mechanisms have been postulated to explain fat carcinogenesis [2], however, discrepancies among studies suggest caution. It is possible that, in our study, these macronutrients are markers of a diet rich in olive oil and possibly in raw vegetables intake rather that having an active role in prostate cancer. Potential recall and selection biases are possible as in most case control studies. Awareness about any particular dietary hypothesis in prostate cancer etiology, however, was still limited in the Italian public at the time of the study, and the issue had not received great media attention. This study was not population-based, but the catchment areas were comparable for cases and controls. It is possible that dietary habits of hospital controls may differ from those of the general population; by study design, however, great attention was paid to exclude all diagnoses that might have been associated with or have determined special dietary habits of controls. Moreover, the comparability of recall between cases and controls is improved by interviewing all subjects in a hospital setting. Adjustment for total energy intake should have reduced potential bias due to differential over- or underreporting of food intakes. The major strength of this study is related to its uniquely large dataset. The consistency of findings when major categories of controls were separately used is also worth noting. For instance, OR in the highest versus the lowest quintile of starch intake was 1.4, 1.3 and 1.4 considering controls from surgery, orthopedic and other wards, respectively. Furthermore, our findings are strengthened by the nearly complete participation of identified cases and controls, the reliance on a validated food-frequency questionnaire [18, 19], the assessment of a broad range of nutrients, and the geographically heterogeneous dietary habits over Italy, which increases study power to detect any significant or meaningful associations. Allowance was also made for various potential confounding factors, including education, which was directly associated with prostate cancer risk, likely reflecting a higher prevalence of PSA testing among more educated men. Additional allowance for alcohol drinking (which did not appreciably differ in cases and controls), BMI and physical activity did not meaningfully change any of the results. In conclusion, the findings of this study support the hypothesis that, in the Italian population, intakes of some macronutrients are related to prostate cancer. This underlines the potential importance of diet and consequently of possible dietary changes in the risk of this cancer. Acknowledgements We wish to thank Mrs O. Volpato for study coordination, and Mrs L. Mei and Mrs I. Calderan for editorial assistance. The contributions of the Italian Association for Research on Cancer (year 2002), the Italian League Against Cancer, and the Italian Ministry of Education (COFIN 2003) are gratefully acknowledged. References 1. Armstrong B, Doll R. Environmental factors and cancer incidence and mortality in different countries, with special reference to dietary practices. Int J Cancer 1975; 15: Kolonel LN. Fat, meat and prostate cancer. Epidemiol Rev 2001; 23: Rohan TE, Howe GR, Burch JD, Jain M. Dietary factors and risk of prostate cancer: a case control study in Ontario, Canada. Cancer Causes Control 1995; 6: Lee MW, Wang RT, Hsing AW et al. Case control study of diet and prostate cancer in China. Cancer Causes Control 1998; 9: West DW, Slatery ML, Robinson LM et al. Adult dietary intake and prostate cancer risk in Utah: a case control study with special emphasis on aggressive tumors. Cancer Causes Control 1991; 2: Kristal AR, Cohen JH, Qu P, Stanford JL. Associations of energy, fat, calcium, and vitamin D with prostate cancer risk. Cancer Epidemiol Biomarkers Prev 2002; 11: Godley PA, Campbell MK, Gallagher P et al. Biomarkers of essential fatty acid consumption and risk of prostatic carcinoma. Cancer Epidemiol Biomarkers Prev 1996; 5: De Stefani E, Deneo-Pellegrini H, Boffetta P et al. Alpha-linolenic acid and risk of prostate cancer: a case control study in Uruguay. Cancer Epidemiol Biomarkers Prev 2000; 9: Giovannucci E, Rimm EB, Colditz GA et al. A prospective study of dietary fat and risk of prostate cancer. J Natl Cancer Inst 1993; 85: Schuurman AG, van den Brandt PA, Dorant E et al. Association of energy and fat intake with prostate carcinoma risk: results from the Netherlands Cohort Study. Cancer 1999; 86: Veierod MB, Laake P, Thelle DS. Dietary fat intake and risk of prostate cancer: a prospective study of 25,708 Norwegian men. Int J Cancer 1997; 73: Severson RK, Nomura AMY, Grove JS, Stemmermann GN. A prospective study of demographics, diet, and prostate cancer among men of Japanese ancestry in Hawaii. Cancer Res 1989; 49: Key TJ, Silcocks PB, Davey GK et al. A case control study of diet and prostate cancer. Br J Cancer 1997; 76: Villeneuve PJ, Johnson KC, Kreiger N, Mao Y. Risk factors for prostate cancer: results from the Canadian National Enhanced Cancer Surveillance System. The Canadian Cancer Registries Epidemiology Research Group. Cancer Causes Control 1999; 10: Pelucchi C, Talamini R, Galeone C et al. Fibre intake and prostate cancer risk. Int J Cancer 2004; 109:

6 Bosetti C, Talamini R, Montella M et al. Retinol, carotenoids and the risk of prostate cancer: A case control study from Italy. Int J Cancer 2004; 112: Salvini S, Gnagnarella P, Parpinel MT et al. The food composition database for an Italian food frequency questionnaire. J Food Compos Anal 1996; 9: Franceschi S, Barbone F, Negri E et al. Reproducibility of an Italian food frequency questionnaire for cancer studies. Results for specific nutrients. Ann Epidemiol 1995; 5: Decarli A, Franceschi S, Ferraroni M et al. Validation of a foodfrequency questionnaire to assess dietary intakes in cancer studies in Italy. Results for specific nutrients. Ann Epidemiol 1996; 6: Breslow NE, Day NE. Statistical methods in cancer research, vol. 1: The analysis of case control studies. IARC Scientific Publications No. 32. Lyon: International Agency for Research on Cancer Willett WC, Stampfer MJ. Total energy intake: implications for epidemiologic analysis. Am J Epidemiol 1996; 124: Decarli A, Favero A, La Vecchia C et al. Macronutrients, energy intake, and breast cancer risk: implications from different models. Epidemiology 1997; 8: Le Marchand L, Kolonel LN, Wilkens LR et al. Animal fat consumption and prostate cancer: a prospective study in Hawaii. Epidemiology 1994; 5: Whittemore AS, Kolonel LN, Wu AH et al. Prostate cancer in relation to diet, physical activity, and body size in blacks, whites, and Asians in the United States and Canada. J Natl Cancer Inst 1995; 87: Vlajinac HD, Marinkovic JM, Illic MD, Kocev NI. Diet and prostate cancer: a case control study. Eur J Cancer 1997; 33: Hayes RB, Ziegler RG, Gridley G et al. Dietary factors and risk for prostate cancer among blacks and whites in the United States. Cancer Epidemiol Biomarkers Prev 1999; 8: Ramon JM, Bou R, Romea S et al. Dietary fat intake and prostate cancer risk: a case control study in Spain. Cancer Causes Control 2000; 11: Serra-Majem L, La Vecchia C, Ribas-barba L et al. Changes in diet and mortality from selected cancers in southern Mediterranean countries, Eur J Clin Nutr 1993; 47: S25 S Chokkalingam AP, Pollak M, Fillmore CM et al. Insulin-like growth factors and prostate cancer: a population-based case control study in China. Cancer Epidemiol Biomarkers Prev 2001; 10: Lipworth L, Martinez ME, Angell J et al. Olive oil and human cancer: an assessment of the evidence. Prev Med 1997; 26: Newcomer LM, King IB, Wicklund KG, Stanford JL. The association of fatty acids with prostate cancer. Prostate 2001; 47: Tzonou A, Signorello LB, Lagiou P et al. Diet and cancer of the prostate: a case control study in Greece. Int J Cancer 1999; 80: