Document jyMdKLLKMXrQpmbye6pQvkOkN

Cancer Causes and Control 1996, 7, pp. 5--18 Nutrition and renal cell cancer Alicja Wolk, Per Lindblad, and Hans-Olov Adami (Received 10April 1995; accepted in revisedform 31July 1995) Epidemiologic evidence on the relation between nutrition and renal cell cancer is reviewed. Kidney cancer, compris ing 1.7 percent of all malignant diseases diagnosed worldwide, shows about a 20-fold international variation in the incidence in men and 10-fold in women. This substantial variation indicates an important causal role of environmental factors. Renal cell (parenchymal) cancer (RCC) accounts for about 80 percent of all kidney cancers.While the etiology of RCC is incompletely understood, analytic epidemiologic studies provide consistent support for a positive associa tion of obesity with risk of RCC; the dose-response observed supports a causal relationship. Only a few prospective studies, all of them limited in size, have been published, while ecologic and case-control studies suggest that diet may be important in the etiology of RCC. However, contradictory results and methodologic limitations in some casecontrol studies prevent definite conclusions concerning diet and RCC. A positive association of protein and fat intake, as well as their main food sources (meat, milk, fats),with risk of RCC - as suggested by ecologic studies - has no clear support in analytic epidemiologic studies. A protective effect of vegetables and fruits has been observed in most casecontrol studies, while the majority do not show an association between alcohol, coffee, and risk of RCC. Recent reports indicated an increased risk of RCC associated with consumption of fried/sauteed meat and low intakes of magnesium or vitamin E. An apparent positive associationwith total energy intake, perhaps due to bias, needs further investigation. Cancer Causes and Control 1996,7,5-18 Key words: Diet, epidemiology, kidney cancer, obesity, renal cell cancer, review. Introduction It has been estimated that kidney cancer afflicted 127,000 individuals worldwide in 1985, thereby accounting for 1.7 percent of all malignant diseases.1 In the same year, about 80,000 patients died from kidney cancer.2 Incidence and mortality rates have been increasing slowly both in the United States3 and in several,4'5 but not all,6 European countries. There have been no major advances in treat ment of kidney cancer over the last decades, although the survival rates are increasing gradually, presumably due to earlier diagnosis.4 Prevention, therefore, emerges as an increasingly important strategy in the combat of kidney cancer. In adults, cancer of the kidney encompasses two major histopathologic entities, namely renal cell (parenchymal) cancer (RCC) and renal pelvic cancer. The latter arises in the transitional cell epithelium -- in the same way as can cer of the ureter (usually grouped with cancer of the kidney) and urinary bladder. Renal pelvic cancer is associated strongly with tobacco smoking and with analgesics containing phenacetin, while data on dietary associations are sparse. This overview, therefore, will be confined--unless otherwise stated--to renal cell cancer, which accounts for about 80 percent of all kidney cancers in adults. In the literature, the distinction between renal cell cancer and renal pelvic cancer is not always explicit. This ambiguity pertains, in particular, to ecologic stud ies. The international variation in the incidence of kidney cancer is about 20-fold in men and at least 10-fold in women. The highest rates are found in European popula tions -- notably Scandinavian -- and in North America.7 Although part of the geographic variation could be due Drs Wolk, Lindblad, and Adami are with the Department ofCancer Epidemiology, University Hospital, Uppsala, Sweden. DrAdami is also af filiated with theDepartment ofEpidemiology, Harvard School ofPublicHealth, Boston MA, USA. Address correspondence toDr Wolk, Department ofCancerEpidemiology, University Hospital, S-7S18S Uppsala, Sweden. This project wasfundedpartially by the Swedish Cancer Society. 1996 Rapid Science Publishers 5Cancer Causes and Control. Vol 7.1996 A. Wolk et al to differences in diagnostic intensity and autopsy rates, the importance of environmental factors in renal cell car cinogenesis is likely to be substantial. For many years, however, RCC attracted limited epidemiologic interest, with smoking and possibly obesity being the only reason ably well-documented risk factors. During the last few years, research on the etiology of RCC has been expanded substantially. A recently pub lished, large, international, collaborative case-control study813 provided several new and important leads for future research. For instance, while analgesic use appeared to be unrelated to the risk of RCC, clear asso ciations were found with certain occupational exposures9 as well as with endocrine and reproductive factors in women.10 There is also substantial evidence suggesting an association between hypertension and/or its treatment and RCC risk.11 Most importantly, however, strong and complex associations were observed between obesity and risk of RCC, notably in women,12 and further evi dence was provided for a role of diet in the etiology of RCC. In the following, these recent findings are inter preted and discussed in the context of previous epide miologic results from ecologic and analytic studies, as well as in the light of evidence from animal experiments. Ecologic dietary studies Several of the prevailing nutritional hypotheses concern ing the etiology of renal cancer have been generated from ecologic studies.14'17 A comparison has been made be tween incidence and mortality data for renal cancer and per capita food consumption data of the Food and Agri culture Organization (FAO) from different countries or, less commonly, within-country data. As with any correla tional studies, the possibility of confounding by other risk factors exists, in particular, obesity and smoking, which are important risk factors for RCC. Armstrong and Doll14 found a strong positive correlation between age-adjusted incidence of renal cancer in 23 countries and per capita consumption of animal protein (r = 0.82), total fat (r = 0.76), total protein (r = 0.63), and calories (r = 0.60). At the food level, the correlations were highest for milk (r = 0.74) and meat (r = 0.72); other correlations were below 0.6. In a worldwide comparison of 18 countries, Wynder et al15 found a positive correlation between mortality from kidney cancer and per capita consumption of milk (r = 0.76), fats and oils (r = 0.67), and sugar (r = 0.67). At the nutrient level, calories (r = 0.85), fat (r = 0.83), and protein from animal sources (r = 0.80) were correlated strongly positively with kidney cancer; in contrast, pro tein (r = -- 0.79) and calories (r = -- 0.72) from plant products were negatively associated. The plant foods most strongly correlated with kidney cancer mortality 6 Cancer Causes and Control.Vol 7.1996 were cereals (r=-- 0.67, r = -- 0.59),14,15 pulses (r = -- 0.56)14 and vegetables (r = -- 0.47).14 Shennan16 drew attention to a possible association be tween coffee consumption and kidney cancer by demon strating a strong correlation (r = 0.79) between ageadjusted mortality rates and per capita consumption of coffee in 16 countries. The correlation of coffee with kidney cancer mortality, based on an analysis of 32 coun tries, was 0.62, while the correlation with incidence rates in 23 countries was lower (r = 0.5l).14 Correlations between per capita intake of different alcoholic beverages (spirits, wine, beer) and age-adjusted mortality rates for renal cancer in 41 states of the United States were signifi cant only for beer, in both men (0.75) and women (0.42).17 In another ecologic study among five ethnic groups in Hawaii (USA), statistically significant positive associa tions with age-adjusted incidence rates of kidney cancer were observed for beer but not for wine or liquor.18 Case-control dietary studies Nutrients Results from exploratory ecologic studies during the 1970s were tested subsequently in several case-control studies. In most of them, analyses were performed at the food level and only in two of the studies19'20 were specific nutrients also investigated. Recently, a large mul ticenter study 13 on diet and RCC was performed with analyses of nutrients. Briefly, coordinated populationbased case-control studies using the same protocol were carried out in five countries: Australia, Denmark, Swe den, USA, and Germany. Dietary data were collected by self-administered food-frequency questionnaires, with ex ception of in-person interviews in Sweden. A total of 1,185 incident, histopathologically confirmed cases (698 men, 487 women) and 1,526 controls (915 men, 611 women), frequency-matched to cases by gender and age, were included in the analyses. (Germany was excluded due to differences in the food questionnaire.) Results from pooled analyses13 as well as from individual centers were published,21'22 or submitted for publication.23'24 Case-control studies of macro- and micronutrients and risk of renal cell cancer are summarized in Table 1 and Table 2. The interpretation of the associations of any specific nutrient with the risk of RCC, especially energy and energy-deriving macronutrients, is hampered by high collinearity among the major nutrients. However, methods to separate the individual effects of protein, fat, carbohydrates, and total energy have been proposed,25'31 which are satisfactory, if not perfect. In two studies,13'20 the association of energy intake with RCC was investigated and both reported a positive relation among incident cases, although this was signifi- Nutrition and renal cell cancer o co O CD S o_ Go 1c11 c CO G G G c 2 o G c .O (0 E CO G *o JnokC_ CO G c3 Sa CO o 8c G - O OL E> G C LU sG > D P .<0 55 I I I: I I I I IIII Q m cd o cm ooooo *D is "sD 3 2_ Q> -o 2 S "D CWO 3CO oh; c 1 fils. SS ?O c3 wa) Cccm>BCwfoloc^cii(J(O/)?QrtQd.>i>'oo"-- in-C*)0>CMCMi-r- Ot- 73 a II > CO CO o o o oo I I I I 05 CM (M O 00 N 00r;0)0) " i" d d ;o coo Vp 0s co IoT) w w 5i i i i 0> o, .i IO 00 CM 05 O O CO r r O^ cr s ii C<O 3Tt0^0f0- 0CM CoO LO 00 CM LO COIf) CM O CO O^ N CO - CMCO CO CM c wI 3- il I Ss s m 1 Tc3 E3 oO) O > 'r" c T33 Ec < 5g w CO * .^ c c c Ills ! 2 < CO to d% t SO *TISo 1c*8 < = 3 5 CO <<Q(0D Q. X n oco w >* Hi o3 E to 3 C O c T3 3 "8 E E 2 2 Hi o ics l3 g t B) -C a) Q 1c5E5c. u|<a w ' -a V ok-S Q._r ;oS (0 to .2, ^ i 2 >. 3 S' i hgi -iSr so re J3 Q. Cancer Causes and Control.Vbl 7.1996 7 A. Walk et al Table 2. Vitamins and renal cell cancer Study (ref) Year Location Population McLaughlin, efa/19 1984 Maclure and Willett20 1990 Wolk, efa/13 1995 Minnesota, USA Men and women Boston, MA, USA Men and women Multicenter study: Australia, Denmark, Sweden, USA Men and women a Energy-adjusted risk estimates. b Preformed vitamin A. Number of cases Relative risks for high- cf low-intake categories 313 men 182 women 203 Vitamin C Carotene 0.7 (men) 0.8 (men) 1.2 (women) 1.3 (women) -- 0.6a Retinol 1.2 (men) 1.3 (women) 0.9a,b Vitamin E -- -- 1,185 0.9a 0.9a 1,2a 0.9a cant only in one 13 of these studies. We cannot exclude a potential risk for differential misclassification in dietary self-reports in these studies, i.e., cases might under- or overreport their remote food intake. Indeed, differential misclassification was suspected in the earlier study,20 where energy intake was associated with risk of RCC among incident but not among prevalent' cases inter viewed more than one year after diagnosis.20 Analysis of subset data from the US in the multicenter study, how ever, did not reveal any difference in risk estimates for energy between short-term (dietary self-reports < 1 year after diagnosis) and long-term reports > 1 year); the rela tive risk (RR) per 100 kcal increase in total energy intake was RR = 1.02 in both case series.13 Protein intake was associated positively with risk of RCC in the case-control study by Maclure and Willett20 for incident and for prevalent cases, and in the study from Minnesota21 based on 415 cases (260 incident and 155 prevalent). However, results from the multicenter study (including 260 incident cases from the Minnesota [USA] study) did not clearly support this hypothesis.13 There was an intricate problem of separating the effect of protein per se from protein-calories. The situation was similar for fat intake. Although Maclure and Willett20 suggested a weak positive association of saturated fat and animal fat with RCC, results from the multicenter study neither confirmed nor refuted this hypothesis. In a model including, simultaneously, all macronutrients ex pressed in calories (Table 1: the multicenter study --all centers and each center separately), the RR (all centers) for 100 protein-kcal was slightly higher than for 100 fatkcal and for 100 carbohydrate-kcal or 100 kcal per se. However, due to the wide confidence interval for protein-kcal, the hypothesis about a positive relation of protein (independently of calories) to RCC was not con firmed. Although the US center revealed a higher risk 8 Cancer Causes and Control.'Vol 7.1996 estimate for protein-kcal than other centers, in a random effects model using data from all centers, the risk esti mate for the highest protein quartile was RR = 1.0 (95 percent confidence interval [Cl] = 0.7-1.6). The hypothesis about positive association of RCC with pro tein, although biologically plausible,21 has no clear sup port from the limited number of analytical epidemiologic studies. Wynder et alK hypothesized a role for cholesterol in the development RCC. However, neither of the studies13,20 corroborated this hypothesis. The hypothesis that /^carotene has a protective effect was tested in three studies;13'19'20 however, none of them reported a significant negative association with RCC (Table 2). Vitamin A and retinol were not associated with a risk of RCC in any of the studies.13'19'20 No effect of vitamin C was observed in the study by McLaughlin et al}9 In the multicenter study, the protective effect of vita min C was significant only in the subgroup of nonsmo kers, RR = 0.6 (Cl = 0.4-0.9).13 In the multicenter study, low levels of vitamin E (below the lowest 10 percent of intake in the studied populations) increased the risk significantly, by about 40 percent. In the same study, a significant positive association with RCC was observed for the lowest 10 percent of magnesium intake, with an approximately 60 percent increase in risk. Although var ious models of action have been proposed for vitamin E, its role as an antioxidant or modulator of immune func tions has been emphasized most.32 The epidemiologic finding for magnesium has support in animal studies.33 It has been postulated that the higher incidence of renal cancer in experimental animals deficient in magnesium is due to an impairment of humoral and cellular immu nity. Also, magnesium deficiency has been reported to alter the fidelity of DNA replication, which could trigger a carcinogenic process.34 Calcium, potassium, and iron were not associated with renal cell cancer risk.13 Nutrition and renal cell cancer Foods All case-control studies on diet and renal cell can cer,13'19,20,35'41 had been analyzed at the food level (Table 3). It was concluded in the multicenter study that food results could be confounded by energy. After adjustment for energy, positive significant associations with some food (milk, eggs, foods rich in animal protein, cereal/ bread) were no longer significant and the protective ef fect of vegetables and fruit became more pronounced.13 In the context of these new results, previously reported positive associations of meat and milk with RRC may have been confounded by energy intake. In the Minnesota study,21 no association was observed for degree of `doneness' and method of cooking of red meat (baked/roasted/boiled/stewed/broiled/grilled or fried/sauteed). In the multicenter study (including 260 in cident cases from the Minnesota study comprising 415 cases), fried meats were associated with a significantly increased risk of RCC (RR = 1.5,Cl = 1.2-1.8); the degree of `doneness' of meat showed a significant dose-response ip trend < 0.05).13 Pan frying/sauteing of meat induces the formation of heterocyclic amines, which appear to be absorbed rapidly and distributed to the kidneys. In feeding experiments in primates, DNA adducts have been observed in different organs, including the kid neys.43 It is probable, therefore, that human populations who consume large amounts of meat also ingest large quantities of carcinogens derived from fried meats. This provides an alternative explanation for the international correlations between per capita meat consumption and rates of kidney cancer, a relation widely ascribed to ani mal protein and fat. Despite experimental evidence that nitrosamines can induce renal cell carcinoma,44 no asso ciation with preserved meat was found in the multicenter study13 and in the Minnesota study,21 suggesting that in take of nitrites and nitrates may not be an important risk factor for RCC. The majority of case-control studies reported a nega tive association of vegetables and fruits with RCC (Table 3). In detailed analyses of vegetables and fruits in the mul ticenter study,13 orange/dark-green vegetables seemed to have the strongest inverse association, with RR = 0.7 (Cl = 0.5-0.9) for the highest quartile ( > 3.4 times/wk); P trend = 0.04. Significant negative associations with RCC were even more clear in nonsmokers: more fre quent consumption of cruciferous (> 1.7 times/wk) or orange/dark-green vegetables (> 3.4 times/wk) or total fruits (> 16 times/wk) was associated with RR = 0.6 (Cl = 0.4-0.9) for all these three foods. Orange/darkgreen vegetables -- a rich source of jS-carotene -- also con tain plant phenols, with an antipromotive effect in vivo and in vitro.** Vegetables from the Brassicacae family (cruciferous) -- a rich source of vitamin C--contain organ ic isothiocyanates, substances with antitumorigenic ef fects in experimental studies, notably in the initiation stage of the tumorigenesis.46 Allium vegetables, which showed a slightly weaker association in nonsmokers for consumption > 1.9 times/wk (RR = 0.7,Cl = 0.4-1.1) than cruciferous and orange/dark-green vegetables, may lower risk for several forms of cancer in humans,47 per haps due to free radical scavenging activity, immune sys tem modulation, or a direct cytotoxic effect on cancer cells.48 Coffee and tea The findings in case-control studies concerning coffee and tea consumption are conflicting. Most studies of cof fee showed no association,15,37,38'40,49'50 or a weak non significant increase,19,20,35,51 or decrease41,52 in the risk of RCC. The study in France53 based on 196 cases, showed a nonsignificant twofold increased risk in men and a four fold increased risk in women, when comparing ever- cf never-drinkers of coffee. In the study of 160 cases from Los Angeles (California, USA),36 drinking five or more cups of coffee per day gave a nonsignificant, twofold risk in women but not in men. The results from the large mul ticenter study-where adjustments were made for age, study center, smoking, and body mass index (BMl)(wt/ ht2) -- indicated a twofold significant increase in the risk in women (but not in men) drinking more than six cups of coffee per day; however, no dose-response was observed.13 Two studies of tea and RCC36,53 reported decreased risk in both genders with the strongest negative association (70 percent reduction in risk) ob served in women drinking tea daily in comparison with never-drinkers.36 However, most case-control studies, 20,37,38,40,49 incju(jing the large multicenter study,13 reported no association, or increased risk only in women,19,52 or only in men41 Despite the numerous stu dies, there has been no convincing evidence linking RCC and consumption of coffee or tea. Alcohol A hypothesized positive association between alcohol in take and risk of RCC has not been corroborated in casecontrol studies; only one54 of 16 studies reported a non significantly increased risk among ever- cfnever-drinkers and a majority15'19'20'35'36'40'41'50'53'55 reported no associa tion. Indeed, there is even some suggestion of a negative association which seems to be more pronounced in women than in men and possibly associated with wine consumption.13,38,52,56,57 In the multicenter study,13 drink ing of three or more glasses of wine per week was asso ciated with 80 percent decreased risk in women (RR = 0.2,0 = 0.1-0.4); consumption of wine in men was generally lower and the RR for men in the highest Cancer Causes and Control. Vol 7.1996 9 A. Wolk et al quartile (drinking 1.3 or more glasses per week) was 0.8 (Cl = 0.5-1.3). These findings need further investigation. Artificial sweeteners The question of artificial sweeteners has been addressed in five case-control studies. In four of them,19'35'36'52 the results were negative and did not indicate any association between RCC risk and consumption of saccharine or other artificial sweeteners. In the analyses by McLaugh lin et al,19 adjustments were made for age, smoking, and weight. The most recent study, by Asal et al,49 showed a significantly increased risk in men (odds ratio [OR] = 2.1, Cl = 1.3-3.5) after adjustments for age, smoking, and weight; the association in women was not significant (OR = 1.3). Body weight The most consistent findings in case-control studies of renal cell cancer have been those concerning overweight and obesity12'15,19,20'36'38'40'41'49,51'54'57'59 (Table 4). Only one hospital-based study38 failed to show a positive associa tion between BMI and RCC. Irrespective of the measure used, i.e., relative body weight (RBW) or BMI (weight [kg] divided by height squared [m2], or by height raised to a power of 1.5 [in women]), increasing body weight was associated with an increasing risk of RCC. Most stu dies were consistent, with a linear relationship between body weight and risk of RCC. The collective evidence from the case-control studies presented in Table 4 and especially the results from the large multicenter study based on 1,732 cases12 seem to indicate that body weight might have stronger impact on RCC in women than in men. In this multicenter study, rate of weight change (esti mates as kg of weight change per annum) appeared to be an independent risk factor among women but not among men. Weight changes (fluctuations) which were asso ciated positively with RCC risk in univariate analyses were not significant in the multivariate model including simultaneously BMI level, rate of weight change, and number of changes.12 The Swedish center (one of five centers included in the multicenter analyses of relative weight and RCC) analyzed data on BMI and weight fluc tuations in detail.60 In contrast to the overall multicenter analyses, this substudy indicated that women who have decreased their weight > 5 kg two or more times during their lifetime are at increased risk of RCC, independently of BMI (OR = 3.9, Cl = 1.2-12.5; adjusted for age, education, smoking, BMI, amphetamine use). In men, the number of weight-loss periods (criterion for weight loss in men was > 10 kg) was not associated with an increased risk of RCC. However, few male cases and controls reported such weight losses, and the statistical power was thus low. 10 Cancer Causes and Control. Vol 7.1996 Lack of agreement between the Swedish substudy and the multicenter results for weight changes might indicate a random effect observed in the substudy. Therefore, the hypothesis about increased RCC risk associated with weight changes/fluctuations should be investigated further. Hormonal influences are one plausible mechanism by which BMI and weight fluctuations might be involved in renal carcinogenesis. Estrogen and progesterone recep tors have been demonstrated in normal and malignant renal cells.61 In male Syrian hamsters, potent estrogens given in high doses induce renal tumors, with a cumula tive incidence approaching 100 percent.62'64 Sex steroids may affect renal cell proliferation and growth and there by carcinogenesis, by direct endocrine receptor-mediated effects,61 by regulation of receptor concentrations,65 or through paracrine growth factors (e.g., epidermal growth factor).66 Another possibility is that obesity and weight fluctua tions are associated with a metabolic syndrome in a sub set of persons with abdominal or upper body obesity. As abdominal fat tissue is mobilized more easily,67 voluntary weight reduction may be expected to be more successful in persons with upper body obesity than in those with obesity in the lower part of the body. The metabolic syn drome entails a multitude of features, e.g., hypertension, decreased insulin sensivity, increased levels of biologi cally active insulin-like growth factor, and, in women, anovulation and excessive androgen production.67 Further renal damage may occur as a result of hyperten sion and metabolic complications of obesity, with or without association with a metabolic syndrome. Such dam age may predispose the kidney to other carcinogens. There is some logical consistency between apparent risk reported for total energy intake and risk associated with BMI. Cohort studies Vitamin A In a prospective cohort of about 16,000 men,68 serum samples were collected and stored. During follow-up of one to four years, eight men developed kidney cancer. They did not have a significantly lower, mean-standar dized retinol concentration in the serum than 172 con trols in whom no cancer in any form developed (218 cf 229 IU/dl).68 In a community cohort of 3,012 individuals from Evans County, Georgia (USA),69 followed up for 12 to 14 years, kidney cancer developed in two persons; their mean serum retinol was 11.4 /ig/dl lower than in the control group. Both these studies are based on num bers too small to allow any conclusion to be drawn. Nutrition and renal cell cancer Fruits Vegetables Fats --------------------------------------------------------------------------------------------------------------------------------------------------- Relative risks for high- c f low-intake categories Dairy Meat B '5 o in < Z n 9 > I I 5o 3B bB CO CO < ddZ 1 1 cc 2 .ooC. ^c^a ID CM co o T-- CO CO o _ CD 0$) 11 o S' rn O |ol 03 <cDo CoO. "8O<O5 O J2O i to i .Q LO C> CO GO 03 03 O o oo o o o -*-1 03 w03 ss E^ CO 1- Bo 03 CO do 0) .o B O) CO dd w_ J2 ^ E oM I Is- 03 CO E, I oo CM N oo II c" A "fcfl Eo =, g =, 'jr 75 1 1 9So CO CO < < d-ZZ II ra o. E S rao A03A03A03 odd , .O S 1 e 1> Ie&e Sills B o (0 o o> A A A XI 00 00 00 T**. dodo ~ O) i_ I I I I I I I I I I E .o _ o CM o c 03 CO o -- -CcJ iT(0J < < 00 ZZ < z =* E <o cc E co cfc -oc -5E o> .oc LO < N CO h*. 03 Z t-1 d d o =1 s e -5 E Pk- r- "22 E t<od =-i -ag aQ. E 15 JO 3? o-_i 15 Q. JO 5o2o|5S>8c---gIc - j=Q"2. S2oJ3 aoE2"ra' < < LO < < < * IP CO CO i-- O O) N CO zz z Z Z co ^ od C ^ 1= Eo e, J, fc'i* fOOo>es) C8_O II T3 "a I I Eo AA CO 03 T* o II c" ,,c _ c" --, E E c oE_ --cE2_ oE_ E2 sP _,E_r_lau-_a E55 ooEEJi,^'rT;'rJsp9p c > S. E JZ -D |! oEco3c__^_-;_cc_*q_*wf"_"^_t3c_c_p_3c_c_3\Q_jc__oQ_r_.*^*_":__.o'03'0320C.v0i-j'.0S^3 O. A CM 03 ^^d^^^rNdt-' ddr^o n 3OQ. CD ID o ICo-O I0D3 ToC--O CoCOO CCoMO < </) =>* c oE rn< S3 Ess ?m?- Eps c < c E^E i D. i -cu -S- -oc -ca O O) cO <2. Tc> C K C j= C w cd) o S v; 2 c 2 2 5 82 3 ii 5 < co c c Eo ES o CM r-- 013- CCOO CCTMO~ CO*"M Eo -2S T53 c _ i>l.i|.0i.^Eo$ c = c c c o> c C to: 2^ ^2 z co BE r_o 3? . H o *1 Ii i- D. c . Eo 5 *co |B< 5 12 <"w3 2 B CO c > m S' =7 E a o t= hco. = 5 *co CD ? 3 .2 <0 IS CO 2o 03 _j 03 03Q0) yO) 222 Q3 .E O C 03 2 . 16 <5 <o <5 PJ>C, M t%l3S .11 CD t- 3w W3 ^ CD uCancer Causes and Control.'Vol 7.1996 men women aNA = not associated with risk of renal cell carcinoma. b Energy-adjusted risk estimates. Number of cases Location Population Study (ref.) Year A. Wolk et al Table 4. Body weight and renal cell cancer Study (ref.) Year Location Population Number of cases Wynder, etal15 1974 USA 129 men 73 women McLaughlin, etal19 1984 Goodman, et al52 1986 Yu, etal36 1986 Minnesota, USA USA Los Angeles, CA, USA 313 men 182 women 173 men 71 women 107 men 51 women Asal, et al49 1988 Oklahoma, USA 209 men 105 women Kadamani, et al57 Oklahoma, USA 1989 Maclure and Willett20 Boston, MA, USA 1990 Talamini, etal36 Northern Italy 1990 Partanen, et al51 Finland 1991 McCredie and Stewart58 Australia 1992 McLaughlin, et alA0 Shanghai, China 1992 21 men 40 women 135 men 68 women 150 men 90 women 338 men and women 310 men 179 women 82 men 59 women Benhamou, etal53 1993 France 134 men 57 women Finkle, etal59 1993 Kreiger, etal41 1993 USA Kaiser Program Ontario, Canada 161 women 282 men 181 women Hiatt, ef a/54 1994 Mellemgaard, et al12 1995 USA Kaiser Program Multicenter study: Australia, Denmark, Sweden, USA 163 men 88 women 1,050 men 682 women Variable Time period studied Relative risks (95% confidence interval) for high cf low weight Men Women RBW3 Recent weight RBW >125 RBW >125 NA 29% cases cf 10% controls (P= 0.05) BMIb Recent weight >28 cf <24 >26 cf < 22 BMIb Recent weight 1.5 (1.0-2.4) 2.1 (1.2-3.9) >28 cf <24 >28 cf <24 2.7 (1.5-5.9) 2.4 (1.2-6.9) BMIb -- Q4 c/Q1d Q4 c/Q1d 1 year ago 1.8 (0.8-4.0) 2.7 (0.8-9.3) 10 years ago 2.5 (1.0-5.9) 3.3 (1.0-11.5) Age 20 2.2 (1.0-4.9) 3.1 (0.9-10.5) BMIb -- > 140 c/<120 > 140 c/<120 Age 20 2.5 (1.4-4.6) 2.0 (0.9-4.2) Highest weight 2.8 (1.6-5.0) 1.8 (0.8-4.0) Recent weight 3.3 (1.8-6.1) 1.2 (0.6-2.6) RBW3 Recent weight > 140 c/<120 > 140 c/<120 3.8 3.0(P< 0.05) BMIb Recent weight >28 Cf <28 >28 cf <28 1.7 (1.1-2.8) 1.7 (0.9-3.2) BMP Recent weight > 27c/< 24 both genders 0.7 (0.5-1.1) Overweight Recent weight Overweight cf normal: Obesity both genders 1.2 (0.9-1.7) BMP Usual weight >25 c/< 23 >31 cf<27 1.6 (1.1-2.5) 1.3 (0.8-2.1) BMP Age 40 >22 c/< 19 >29 c/< 24 3.6 (1.1-12.6) 3.6 (0.9-14.7) Age 50 >23 cf<20 > 29 cf< 24 1.7(0.5-57) 3.3 (0.7-15.1) BMP Age 20 >24 c/< 18 > 24 c/< 18 1.1 (0.5-2.4) 4.6 (1.3-16.4) Prior to diagnosis > 27 c/< 20 >27 cf<20 2.4 (1.0-5.9) 3.5 (1.0-11.8) BMP Recent weight -- > 75 c/< 25 2.6 (1.4-4.8) BMP 5 years ago >27 c/< 23 >25 c/<21 Age 25 1.3 (0.8-2.2) > 30 cf< 26 2.5 (1.4-4.6) >28 cf<23 1.9 (1.2-3.1) 1.0 (0.5-1.8) Two periods combined 2.2 (1.5-3.4) 2.2 (1.4-3.5) BMP Recent weight >28 c/< 25 >28 Cf< 22 0.9 (0.5-1.6) 1.2 (0.5-2.9) BMP Usual weight > 27 c/< 23 >33 c/<27 1.6 (1.3-2.1) 2.0 (1.5-2.7) Highest weight >30 cf<25 > 37 cf< 29 1.4 (0.9-1.8) 2.5 (1.8- 3.5) Q4 cfQid Q4C/Q,d Age 20 1.4 (1.0-1.9) 1.5 (1.1-2.1) Age 30 1.8 (1.3-2.3) 2.2 (1.16-3.1) Age 40 1.6 (1.2-2.0) 2.1 (1.5-2.8) Age 50 1.4 (1.1-1.9) 2.3 (1.6-3.3) Age 60 1.4 (1.0-1.9) 1.9 (1.3-2.8) 3 RBW = relative body weight (%), i.e., percent in comparison with an average in the population. b BMI = body mass index (Kg/m2). 0 BMI = body mass index (kg/m2 for men; kg/m15 for women). d Cut-points for quartiles not specified in the article. NA = not associated with risk. 12 Cancer Causes and Control.'Vol 7.1996 Nutrition and renal cell cancer Foods In a prospective cohort study of 34,198 Seventh-day Ad ventists from California (USA), 14 subjects developed RCC during a six-year period (Table 3). The investigators found a negative association with fruits (RR = 0.2 for > 3 fruits/wk) and vegetables (RR = 0.3 for green salads > 3/wk). Consumption of beef or fish more than once per week increased the risk to 1.6, while poultry decreased the risk to O.5.42 None of the risk estimates was statistically significant. Coffee Coffee (> 7 cups cf <2 cups/day) was found to have a negative association with RCC (RR = 0.3, P trend = 0.01) in a prospective cohort70 of 13,664 Norwegian men and 2,891 women; during 11.5 years of follow-up, 35 renal cell cancers were detected. However, this finding is not supported by the findings in a larger Norwegian cohort71 comprising 21,735 men and 21,238 women. Among these, 30 men and 13 women developed kidney cancer during 10 years of follow-up. RRs for drinking >7 cf <2 cups of coffee per day were 0.7 in men and 1.2 in women; the trends were not significant. In the analyses of both Nor wegian cohorts, adjustments were made for age and smoking. Tea In a cohort of 14,085 men from London,72 26 deaths due to kidney cancer occurred during 19 years of follow-up. A weak positive association was found between tea drink ing and kidney cancer (P trend = 0.04); > 10 cups of tea per day was associated with an 80 percent increase in risk.72 Contamination ofdrinking water In a cohort of 31,000 subjects from Washington County (USA),73 31 kidney cancers occurred (18 men, 13 women) during 12 years of follow-up. No association was observed between kidney cancer and drinking of chlori nated surface water at home compared with drinking of unchlorinated ground water (men: RR = 0.8, Cl = 0.3 2.7; women: RR = 1.0, Cl = 0.3-6.0). Although the postu lated relationship between organic chemical by-products of water chlorination and risk of kidney cancer was not confirmed, the CIs were wide.73 In an area of endemic chronic arsenicism in Taiwan,74 64 deaths from kidney cancer were noted in a cohort of 898,806 person-years. A significant dose-response relationship was found be tween the ingested amount of inorganic arsenic and mor tality from kidney cancer in both men and women; for increasing water-arsenic concentrations (170, 470, and 800 pg/1), the corresponding mortality rate ratios were 4.9,11.9, and 19.6 in men and 4.0,13.9, and 37.0 in women. with P-values for linear trend <0.001 in men and women.74'75 These high mortality rate ratios make con founding by some other risk factors unlikely, thus, the study suggests a causal relationship. Smith et al75 esti mated that more than 350,000 people in the US may be supplied with water containing more than 50 fig arsenic per liter (the US Environmental Protection Agency stan dard) and more than 2.5 million people may be supplied with water with arsenic concentrations above 25 fig/l. Alcohol No evidence of an increased risk associated with alcohol drinking was found in a population-based Swedish cohort76 of 8,340 men and 1,013 women with a discharge diagnosis of alcoholism followed up for an average of 7.7 years. This finding is in accordance with results from other cohort studies.77'80 Body weight The results of prospective studies of body weight and risk of kidney cancer are in conformity with those of case-control studies. In a nested case-control study,81 it was found that within a cohort of 47,261 men (students from Harvard and Pennsylvania University) followed up for 16 to 50 years, 77 subjects died of kidney cancer. Those who weighed more than 180 lbs in college had a 2.5-fold (Cl = 0.9-6.8) higher risk than those weighing un der 140 lbs.81 The risk expressed per 10 lbs additional body weight was RR = 1.2 (Cl = 1.0-1.3).82 In a long-term, population-based prospective cohort83 of 750,000 men and women in the USA followed-up for 13 years, the risk of death from renal cancer was twice as high in women with a relative body weight of over 140 percent of average weight as in those with a relative weight of 90 to 110 per cent. In a Danish cohort84 of 43,965 persons with obesity as a discharge diagnosis, 79 kidney cancers (21 men, 58 women) occurred during one to 10 years of follow-up. Compared with the national incidence rates, the RR in obese women was 2.0 (CI = 1.5-2.6) and in men 1.2 (Cl = 0.7-1.8).84 In a similar Swedish cohort (Wolk et al, unpublished), 17,547 obese subjects followed up for an average of 11.4 years, 75 kidney cancers developed (49 women, 26 men). The standardized incidence ratio for kidney cancer was 2.4 (Cl = 1.8-3.2) in women and 2.3 (Cl = 1.5-3.4) in men; specifically for RCC (66 cases), it was 2.4 (Cl = 1.9-3.1). The aggregated results from these four prospective studies provide strong evidence that high body weight is an important risk factor for kidney cancer both in men and women. Animal/mechanistic studies The interpretability of the results of epidemiologic stud ies is limited by the difficulties in accurately assessing past and current human nutrient consumption and of 13Cancer Causes and Control.Vol 7.1996 A. Wolk et al disentangling the possible impact of many dietary fac tors, such as energy, protein, fat, and animal products, which often vary in concert. Analysis of their interactive effects in animal models -- where multiple variables can be controlled simultaneously and precisely -- may contri bute to a better understanding of epidemiologic studies. Protein, fat, energy intake In the experiments in rats given diets with three different protein contents and three different contents of fat, the effects of protein and fat on the initiation and promotion of azoxymethane (AOM)-induced carcinogenesis in the kidney were studied.85 During the initiation phase, the low-protein diet was associated with an increased risk of renal adenocarcinoma (P < 0.001). McLean and Magee86 also reported that severe protein deficiency during dimethylnitrosamine (DMN) administration increased the risk of renal tumor development in rats. Subsequent studies showed that DMN metabolism, by liver slices, was reduced by 50 percent in rats on a protein-free diet.87 The reduction of DMN metabolism by the liver may have allowed a larger fraction of the administered dose to reach the systemic circulation with critical targets in the kidney. Dietary protein has profound effects on the metabolism and biological action of many xenobiotics.88'89 In a study by Clinton et al,Hi rats initially fed a lowprotein diet (eight percent) and switched to 16 percent protein diets following carcinogen (AOM) administra tion showed a compensatory increase in growth after the change in diet. Hypothetically, the nutritional and hormonal changes during this period of compensatory growth may have increased local proliferative stimuli in renal tissue to enhance the accumulation and expression of genetic lesions, contributing to an acceleration of the malignant transformation. In the promotion phase of the same study,85 the renal effects of dietary protein were less pronounced. High-protein diets (32 percent) fed to AOMtreated rats were associated with an increased frequency of inflammatory changes and renal atrophy, but not with occurrence of carcinoma. Variations in dietary fat had no effects on renal pathology, either during the initiation or the promotion phase in the study. Ad libitum energy in take was not associated with renal carcinogenesis. In con trast, in the same study, a significant positive association with energy intake was observed for intestinal carcino genesis (14 kcal/day difference in mean ad libitum intake corresponded to a 146 percent increase in the odds of developing an intestinal adenocarcinoma). Vitamin C, ot-tocopherol, and other antioxidants Long-term chronic administration of estradiol (E^ or diethylstilbestrol to male Syrian hamsters induces kidney tumors.90 The studies reviewed by Liehr91 clearly demon 14 Cancer Causes and Control.Vol 7.1996 strated that vitamin C inhibits the induction of renal tu mors by estrogen by about 50 percent. Moreover, there is evidence that vitamin C inhibits both the covalent mod ification of cellular macromolecules, i.e., direct covalent binding of estrogen quinones, and the damage by free radicals formed by redox cycling between quinones and their respective hydroquinones. A reduction of estrogen quinones by vitamin C, and hence a decrease in quinone concentrations, is a possible explanation for the decrease in both types of damage to cellular macromolecules. A reduction in estrogen quinones by vitamin C will result in decreased DNA and protein-adduct formation. Free radical damage to cellular macromolecules also may be diminished by the free radical scavenging action of vita min C. This protection of cells from covalent damage may result in a decrease in tumor initiation. The possibi lity of an influence of vitamin C on renal tumor promo tion and progression has not yet been investigated.91 Effects of the dietary antioxidants a-tocopherol, t-butylhydroquinone, propyl gallate, and butylated hydroxytoluene -- used widely throughout the world--were examined using a multi-organ carcinogenesis model in male rats. Histopathologic examination showed that atocopherol reduced the incidence and multiplicity of kidney atypical renal tubules, and propyl gallate was ef fective in reducing this latter multiplicity. Butylated hydroxytoluene decreased the incidence and multiplicity of renal cell tumors92 Quercetine, a major constituent of bioflavonoids in plant foods, has antioxidant properties.93 The influence of quercetine on kidney tumor induction by estradiol (Ej) in male Syrian hamsters was investigated by Zhu and Liehr94 in an attempt to understand estrogeninduced carcinogenesis and its prevention by dietary modification. The administration of a diet containing quercetine to hamsters does not induce tumors. How ever, co-treatment of hamsters with E2 and quercetine potentiated E2-induced renal tumorigenesis. Increased formation of 4-hydroxyestradiol together with inhibited inactivation of this catechol estrogen by catechol-Omethyltransferase, may result in elevated levels of this es trogen metabolite, specifically in the hamster kidney, where it may undergo metabolic redox cycling and gen erate mutagenic free radicals. Thus, the potentiaton of estradiol-induced tumorigenesis by quercitin in ham ster kidneys supports a role of 4-hydroxyestradiol in estrogen-induced carcinogenesis in this species. Other exposures Results from a study in rats suggest that high dietary cal cium intake in the presence of lead may increase the in cidence of renal tumors.95 Potassium bromate (KBrOj) is an oxidizing agent used as a food additive, mainly in bread-making. Although adverse effects have not been Nutrition and renal cell cancer evident in animals fed bread-based diets made from flour supplemented with KB1O3, the agent is carcinogenic in rats and hephrotoxic in both man and experimental ani mals when given orally. Potassium bromate is a complete carcinogen, possessing both initiating and promoting ac tivities for rat renal tumorigenesis. Its potential, however, seems to be weak in mice and hamsters.96 Summary and conclusions The cumulative evidence from analytical epidemiologic studies--both of case-control and cohort design -- is most consistent for the positive association of body weight with risk of RCC. The dose-response observed in these studies supports a causal relationship. However, there are still many aspects of overweight and obesity -e.g., weight development during lifetime, weight cycling, body fat distribution, potential mechanisms of action --that should be further investigated. The evidence accumulated from ecologic and casecontrol studies suggests that diet has an important role in the development of RCC. However, methodologic problems in most previous case-control studies (e.g., small numbers of subjects; use of hospital controls; not analyzed at nutrient level, thus without possibility of re cognizing whether total energy intake is a risk factor/ confounder) markedly limit the possibility of inferences concerning diet and RCC. One of the most intriguing findings, the apparent positive association of total energy intake with RCC, needs further investigation in prospec tive cohort studies to eliminate the risk of biased esti mates. Energy intake is the key problem, since if a highcalorie intake is a real risk factor for RCC it will entail methodologic consequences in most of the analyses, both at the nutrient and at the food level. The statistically significant positive associations be tween a low intake of vitamin E and magnesium and risk of RCC found in one large, multicenter, populationbased case-control study need further confirmation. A possible interaction of antioxidants--vitamin C, ^-caro tene, and those included in specific vegetables (allium, cruciferous, orange/dark-green) and fruits--with smok ing status also needs more detailed assessment. More over, confusing results concerning alcohol that suggest differences between men and women in response to these factors require more investigation. The potential risk associated with consumption of fried meats should be clarified in studies using a more quantitative approach to exposure to heterocyclic amines. In summary, several dietary factors should be consid ered as potentially involved in the development of renal cell cancer at different stages of tumorigenesis. However, they need more cumulative confirmation before recom mendations for public health can be made. References 1. Parkin DM, Pisani P, Ferlay J. Estimates of the world wide incidence of eighteen major cancers in 1985. Int J Cancer 1993; 54: 594-606. 2. Pisani P, Parkin DM, Ferlay J. Estimates of the world wide mortality from eighteen major cancers in 1985. Im plications for prevention and projections of future burden. IntJ Cancer 1993; 5: 891-903. 3. Devesa SS, Blot WJ, Stone BJ, Miller BA, Tarone RE, Fraumeni JF Jr. Recent cancer trends in the United States./NC/1995; 87; 175-82. 4. Engeland A, Haldorsen T, Tretli S, et al. Prediction of cancer mortality in the Nordic countries up to the years 2000 and 2010, on the basis of relative survival analysis. A collaborative study of the five Nordic Cancer Registries. APMIS1995; Suppl 49: 90-5. 5. Franceschi S, Levi F, La Vecchia C, Lucchini F, Negri E. Comparison of cancer mortality trends in major Eur opean areas, 1960-89. EurJ Cancer Prev 1994; 3:145-206. 6. The Cancer Registry. Cancer Incidence in Sweden 1992. Stockholm, Sweden: National Board of Health and Wel fare, 1995. 7. Whelan SL, Parkin DM, Masuyer E. Patterns ofCancer in Five Continents. Lyon, France: International Agency for Research on Cancer, 1990; IARC Sci. Pub. No. 102. 8. McCredie M, Pommer W, McLaughlin JK, et al. Interna tional renal cell cancer study. II. Analgesics. Int J Cancer 1995; 60: 345-9. 9. Mandel JS, McLaughlin JK, Schlehofer B, et al. Interna tional renal cell cancer study. IV. Occupation. Int ] Can cer 1995; 61: 601-5. 10. Lindblad P, Mellemgaard A, Schlehofer B, et al. Interna tional renal cell cancer study. V. Reproductive factors, gynecologic operations and exogenous hormones. Int J Cancer 1995; 61:192-8. 11. McLaughlin J, Chow WH, Mandel J, et al. International renal cell cancer study. VIII. Role of diuretics, other anti hypertensive medications and hypertension. IntJ Cancer, in press. 12. Mellemgaard A, Bergstrom R, Lindblad P, et al. Interna tional renal cell cancer study: III. Role of weight, height, physical activity and use of amphetamines. Int J Cancer 1995; 60: 350-4. 13. Wolk A, Gridley G, Niwa S, et al. International renal cell cancer study: VII. Role of diet. Int J Cancer 1995; 65: 67-74. 14. Armstrong B, Doll R. Environmental factors and cancer incidence and mortality in different countries, with spe cial reference to dietary practices. Int J Cancer 1975; 15: 617-31. 15. Wynder EL, Mabuchi K, Whitmore Jr WF. Epidemi ology of adenocarcinoma of the kidney. JNCI 1974; 53: 1619-34. 16. Shennan DH. Renal carcinoma and coffee consumption in 16 countries (Letter). Br] Cancer 1973; 28: 473-4. 17. Breslow NE, Enstrom JR Geographic correlations be tween cancer mortality rates and alcohol-tobacco con sumption in the United States.JNCI 1974; 53: 631-9. 15Cancer Causes and Control.Voi 7.1996 A. Wolk et al 18. Hinds MW, Kolonel LN, Lee J, Hirohata T. Associations between cancer incidence and alcohol/cigarette con sumption among five ethnic groups in Hawaii. BrJ Can cer 1980; 41: 929-40. 19. McLaughlin JK, Mandel JS, Blot WJ, Schuman LM, Mehl ES, Fraumeni JF Jr. A population-based case-control study of renal cell carcinoma.JNCI1984; 72: 275-84. 20. Maclure M, Willett W. A case-control study of diet and risk of renal adenocarcinoma. Epidemiology 1990; 1: 430 40. 21. Chow WH, Gridley G, McLaughlin JK, et al. Protein in take and risk of renal cell cancer. JNCI 1994; 86:1131-9. 22. Mellemgaard A, Engholm G, McLaughlin JK, Olsen JH. Risk factors for renal cell carcinoma in Denmark. I. Role of socioeconomic status, tobacco use, beverages, and fa mily history. Cancer Causes Control 1994; 5:105-13. 23. Lindblad P, Wolk A, Bergstrom R, Adami HO. Diet and risk of renal cell cancer: A population-based case-control study in Sweden. (Submitted) 24. Mellemgaard A, McLaughlin J, Overvad K, Olsen JH. Dietary risk factors of renal cell carcinoma in Denmark. (Submitted) 25. Willett W, Stampfer MJ. Total energy intake: implications for epidemiologic analyses. Am J Epidemiol 1986; 124:17 27. 26. Howe GR, Miller AM, Jain M. Re: "Total energy intake: implications for epidemiologic analyses."Am J Epidemiol 1986; 124:157-9. 27. Pike MC, Bernstein L, Peters RK. Re: "Total energy in take: implications for epidemiologic analyses." Am ] Epi demiol 1989; 129:1312-3. 28. Palmgren J, Kushi LH. Total energy intake: implications for epidemiologic analyses. Am J Epidemiol 1991; 133: 1291-3. 29. Kipnis V, Freedman LS, Brown CC, et al. Interpretation of energy adjustment models for nutritional epi demiology. AmJ Epidemiol 1993; 137:1376-80. 30. Brown CC, Kipnis V, Freedman LS, et al. Energy adjust ment methods for nutritional epidemiology: The effect of categorization. Am J Epidemiol 1994; 139: 323-38. 31. Wacholder S, Schatzkin A, Freedman LS, et al. Can en ergy adjustment separate the effects of energy from those of specific macronutrients? Am J Epidemiol 1994; 140: 848-55. 32. Das S. Vitamin E in the genesis and prevention of cancer. Acta Oncol 1994; 6: 615-9. 33. Elin RJ. Magnesium metabolism in health and disease. Dis Mon 1988; 34:161-218. 34. Blondell JM. The anticarcinogenic effect of magnesium. Med Hypotheses 1980; 6: 863-71. 35. Armstrong B, Garrod A, Doll R. A retrospective study of renal cancer with special reference to coffee and animal protein consumption. Br J Cancer 1976; 33: 127-36. 36. Yu MC, MackTM, Hanisch R, Cicioni C, Henderson BE. Cigarette smoking, obesity, diuretic use, and coffee con sumption as risk factors for renal cell carcinoma. JNCI 1986; 77: 351-6. 16 Cancer Causes and Control. Vol 7.1996 37. McCredie M, Ford JM, Stewart JH. Risk factors for cancer of the renal parenchyma. Int ] Cancer 1988; 42:13-6 38. Talamini R, Baron AE, Barra S, et al. A case-control study of risk factor for renal cell cancer in northern Italy. Cancer Causes Control 1990 1: 125-31. 39. Negri E, La Vecchia C, Franceschi S, D'Avanzo B, Parazzini F. Vegetable and fruit consumption and cancer risk. IntJ Cancer 1991; 48: 350-4. 40. McLaughlin JK, GaoYT, Gao RN, et al. Risk factors for renal-cell cancer in Shanghai, China. Int J Cancer 1992; 52: 562-5. 41. Kreiger N, Marrett LD, Dodds L, Hilditch S, Darlington GA. Risk factors for renal cell carcinoma: results of a population-based case-control study. Cancer Causes Control 1993; 4:101--10. 42. Fraser GE, Philips RL, Beeson WL. Hypertension, anti hypertensive medication and risk of renal carcinoma in California Seventh-day Adventisis. Int J Epidemiol 1990; 19: 832-8. 43. Munro IC, Kennepohl E, Erickson RE, et al. Safety as sessment of ingested heterocyclic amines: initial report. Regul Toxicol Pharmacol 1993; 17: si--sl09. 44. Outzen HC, Maguire Jr HC. The etiology of renal-cell carcinoma. Semin Oncol 1983; 10: 378-84. 45. Strube M, Dragsted LO, Larsen JC. Naturally Occurring Antitumorigens -- I. Plant Phenols. Copenhagen, Den mark: The Nordic Council of Ministers, 1993; Nordic Conference- and Progress Reports: 605. 46. Strube M, Dragsted LO, Larsen JC. Naturally Occurring Antitumorigens - II. Organic Isothiocyanates. Copenha gen, Denmark: The Nordic Council of Ministers, 1994; Nordic Conference- and Progress Report: 539. 47. Dorant E, van den Brandt PA, Goldbohm RA, et al. Garlic and its significance for the prevention of cancer in humans: a critical view. BrJ Cancer 1993; 67: 424-9. 48. Abdullah TH, Kandil O, Elkadi A, Carter J. Garlic revis ited: therapeutic for the major diseases of our times? J Natl Med Assoc 1988 80(4): 439-45. 49. Asal NR, Risser DR, Kadamani S, Geyer JR, Lee ET, Cherng N. Risk factors in renal cell carcinoma: I. Metho dology, demographics, tobacco, beverage use, and obesity. Cancer Detect Prev 1988; 11: 359-77. 50. Harrington JM, Whitby H, Gray CN, Reid FJ, Aw TC, Waterhouse JA. Renal disease and occupational exposure to organic solvents: a case-referent approach: see comments. Br] Ind Med 1989; 46: 643-50. 51. PartanenT, Heikkila P, Hernberg S, KauppinenT, Moneta G, Ojajarvi A. Renal cell cancer and occupational expo sure to chemical agents. Scand J Work Environ Health 1991; 17: 231-9. 52. Goodman MT, Morgenstern H, Wynder EL. A case-con trol study of factors affecting the development of renal cell cancer. Am J Epidemiol 1986; 124: 926-41. 53. Benhamou S, Lenfant MH, Ory-Paoletti C, Flamant R. Risk factors for renal-cell carcinoma in a French casecontrol study. IntJ Cancer 1993; 55: 32-6. 54. Hiatt RA, Tolan K, Quesenberry Jr CP. Renal cell carci Nutrition and renal cell cancer noma and thiazide use: a historical, case-control study (California, USA). Cancer Causes Control 1994; 5: 319-25. 55. Brownson RC. A case-control study of renal cell carcino ma in relation to occupation, smoking, and alcohol consumption. Arch Environ Health 1988; 43: 238-41. 56. Asal NR, Geyer JR, Risser DR, Lee ET, Kadamani S, Cherng N. Risk factors in renal cell carcinoma. II Medi cal history, occupation, multivariate analysis and conclusions. Cancer Detect Prev 1988; 13: 263-79. 57. Kadamani S, Asal NR, Nelson RY Occupational hydro carbon exposure and risk of renal cell carcinoma. Am J Ind Med 1989; 15:131-41. 58. McCredie M, Stewart JH. Risk factors for kidney cancer in New South Wales, Australia. II. Urologic disease, hy pertension, obesity, and hormonal factors. Cancer Causes Control 1992; 3: 323-31. 59. FinkleWD, McLaughlin JK, Rasgon SA, Yeoh HH, Low JE. Increased risk of renal cell cancer among women using diuretics in the United States. Cancer Causes Con trol 1993; 4: 555-8. 60. Lindblad P, Wolk A, Bergstrom R, Persson I, Adami HO. The role of obesity and weight fluctuations in the aetiol ogy of renal cell cancer. A population-based case-control study. Cancer Epidemiol Biomark Prev 1994; 3: 631-9. 61. Ronchi E, Pizzocaro G, Miodini P, Piva L, Salvioni R, Di Fronzo G. Steroid hormone receptors in normal malig nant human renal tissue: relationship with progestin therapy./ Steroid Biochem 1984; 21: 329-34. 62. Reznik-Schuller H. Carcinogenic effects of diethylstilbestrol in male Syrian golden hamsters and European hamsters.JNCI1979; 62:1083-8. 63. Li JJ, Li SA. Estrogen carcinogenesis in hamster tissues: a critical review.EndocrRev 1990; 11: 524-31. 64. Barrett JC, Huff J. Cellular and molecular mechanisms of chemically induced renal carcinogenesis. Renal Failure 1991; 13: 211-25. 65. Traish AM, Wotiz HH. Prostatic epidermal growth fac tors and their regulation by androgen. Endrocrinology 1987; 121:1461-7. 66. Concolino G, Lubrano C, Santonati A, Di Silverio F, Catizone A, Conti C. Epidermal growth factor (EGF) bind ing by normal and neoplastic human renal tissue./ Tumor Marker Oncol 1989; 4: 89-97. 67. Bouchard C, Despres JP, Mauriege P. Genetic and nongenetic determinants of regional flit distribution. Endocr Rev 1993; 14: 72-93. 68. Wald N, Idle M, BorehamJ, Bailey A. Low serum-vitaminA and subsequent risk of cancer. Preliminary results of a prospective study. Lancet 1980; 2: 813-5. 69. Kark JD, Smith AH, Switzer BR, Hames CG. Serum vita min A (retinol) and cancer incidence in Evans county, Georgia.JNCI 1966; 76: 7-16. 70. Jacobsen BK, Bjelke E, Kvale G, Heuch I. Coffee drink ing, mortality, and cancer incidence: results from a Nor wegian prospective study. JNCI 1986 76: 823-31. 71. Stensvold I, Jacobsen BK. Coffee and cancer: a prospec tive study of 43,000 Norwegian men and women. Cancer Causes Control 1994; 5: 401-8. 72. Kinlen LJ, Wollows AN, Goldblatt P, Yudkin J. Tea con sumption and cancer. BrJ Cancer 1988; 58: 397- 401. 73. Wilkins JR, Comstock GW. Source of drinking water at home and site-specific cancer incidence in Washington county, Maryland. AmJ Epidemiol 1981; 114:178-90. 74. Chen CJ, Chen CW, Wu MM, Kuo TL. Cancer poten tial in liver, lung, bladder and kidney due to ingested inorganic arsenic in drinking water. Br J Cancer 1992; 66: 888-92. 75. Smith AH, Hopenhayn-Rich C, Bates MN, et al. Cancer risks from arsenic in drinking water. Environ Health Perspect 1992; 97: 259-67. 76. Adami HO, McLaughlin JK, Hsing AW, et al. Alcoholism and cancer risk: a population-based cohort study. Cancer Causes Control 1992; 3: 419-25. 77. Schmidt W, De Lint J. Causes of death of alcoholics. Q / Stud Alcohol 1972; 33:171-85. 78. Pell S, DAlonzo CA. A five-year mortality study of alcoholics./ Occup Med 1973; 15(2): 120-5. 79. Monson RR, Lyon JL. Proportional mortality among alcoholics. Cancer 1975; 36:1077-9. 80. Jensen OM. Cancer morbidity and causes of death among Danish brewery workers. IntJ Cancer 1979; 23: 454-63. 81. Whittemore AS, Paffenbarger Jr RS, Anderson K, Lee JE. Early precursors of urogenital cancers in former college men./ Urol 1984; 132:1256-61. 82. Whittemore AS, Paffenbarger Jr RS, Anderson K, Lee JE. Early precursors of site-specific cancers in college men and women.JNCI 1985; 74: 43-51. 83. Lew EA, Garfinkel L. Variations in mortality by weight among 750,000 men and women. / Chronic Dis 1979; 32: 563- 76. 84. Moller H, Mellemgaard A, Lindvig K, Olsen JH. Obesity and cancer risk: a Danish record-linkage study. Eur J Cancer 1993; 30A: 344-50. 85. Clinton SK, Imrey PB, Mangian HJ, Nandkumar S, Visek WJ. The combined effects of dietary fat, protein, and en ergy intake on azoxymethane-induced intestinal and re nal carcinogenesis. Cancer Res 1992; 52: 857-65. 86. McLean AE, Magee PN. Increased renal carcinogenesis by dimethyinitrosamine in protein deficient rats. Br J Exp Pathol 1970; 51: 587-90. 87. Swan PR, McLean AE. Cellular injury and carcinogesis: the effect of a protein-free high-carbohydrate diet on the metabolism of dimethyinitrosamine in the rat. Biochem 1971; 124:283-8, 88. Campbell TC, Hayes JR. The effect of quantity and qual ity of dietary protein on drug metabolism. Fed Proc 1976; 35:2470-4. 89. Clinton SK, Truex CR, Visek WJ. Dietary protein, aryl hydrocarbon hydrozylase and chemical carcinogenesis in rats./ Nutr 1979; 109: 55-62. 90. Liehr JG, Sirbasku DA. Estrogen-dependent kidney tu mors. In: Taub M, ed. Tissue Culture of Epithelial Cells. New York, NY(USA). Plenum Publishing 1985; 205-234. 91. Liehr JG. Vitamin C reduces the incidence and severity of renal tumors induced by estradiol or diethyistilbestrol. AmJ Clin Nurt 1991; 54:1256S-60S. 17Cancer Causes and Control.Vol 7.1996 A. Walk et al 92. Hirose M, Yada H, Hakoi K, Takahashi S, Ito N. Modifi cation of carcinogenesis by alpha-tocopherol, t-butylhydroquinone, propyl gallate and butylated hydroxytoluene in a rat multi-organ carcinogenesis model. Carcinogenesis 1993; 14: 2359- 64. 93. Cholbi MR, Paya M, Alcaraz MJ. Inhibitory effects of phenolic compounds on CCl4-induced microsomal lipid peroxidation. Experimentia 1991; 47:195-9. 94. Zhu BT, Liehr JG. Quercetin increases the severity of es tradiol-induced tumorigenesis in hamster kidney. Toxicol Appl Pharmacol 1994; 125:149-58. 95. Bogden JD, Gertner SB, Kemp FW, McLeod R, Bruening KS, Chung HR. Dietary lead and calcium: effects on blood pressure and renal neoplasia in Wistar rats./ Nutr 1991; 121: 718-28. 96. KurokawaY, Maekawa A, Takahashi M, Hayashi Y. Toxi city and carcinogenicity of potassium bromate -- a new re nal carcinogen. Environ Health Perspect 1990; 87:309-35. 18 Cancer Causes and Control.Vol 7.1996