22 Obesityand Body Composition

RACHELBALLARD-BARBASH, CHRISTINE FRIEDENREICH, MARTHA SLATTERY, AND INGERTHUNE

An" explanationfor the increasingincidence of cancersin devel- papersor books were also identified and reviewed. Summary figures rr.-/oping countries has been the hypothesis that a chronic state of of risk estimatesfor the highestcompared to lowest quantileof BMI positive energy balance promotes tumor growth. (Tannenbaum and the major cancer sites were generated.Because of the very large l940a,b) first explored this hypothesisin animal models and demon- number of studiesin breastcancer, studies with at least 100 cases strated that both increasedcalories and fat increasedthe occurrence each of pre- and postmenopausalbreast cancer casesare included and number of breastcancer tumors in mice. Similarly, epidemiologic in the breast cancer figures. Becauseof differencesin risk for colon studiesin humansin the 1970sdemonstrated that heavierwomen were and rectal cancer and the limited number of studies on rectal cancer at increasedrisk of breast and endometrial cancer (De Waard et al., alone, only data on colon cancer were summarized in the colon 1974;Blitzer e[ al., 1976).Other population-basedstudies in the early cancer figure. Biological mechanismspotentially involved in the 1980ssuggested that risk increasedwith increasingweight for several associationbetween obesity and specific cancers were summaized other cancer sites, specifically colon, prostate, gallbladder (among in a table. In addition, population attributable risks estimated for women), and kidney (Lew and Garfinkel, 1979: Hartz et al., 1984; the EuropeanUnion by Bergstromet al. (2001) were comparedto Garfinkel, 1986).Since then, numerousepidemiologic studies have currentestimates for the United Statesand summarizedin a table.A examined the associationof weight and other anthropometric meas- summarystatement for eachcancer site was developedfollowing the ures with cancer incidenceand have explored potential biological criteria for convincing, probable, possible, and no associationas mechanismsto explain observedassociations. The majority of these summarizedin the 1997World CancerResearch Fund and American studies have focused on breast, colon, endometrial, renal cell, Institue for Cancer Research(World Cancer ResearchFund, 1997) esophageaUgastric,and prostatecancer; limited data are available on report on Food, Nutition and the Prevention of Concer: A Global body size associationswith other cancer sites. Perspective. The aim of this review is to provide a comprehensiveoverview of the stateof scientificevidence for.the associationbetween obesity- related risk factors and cancer, with a focus on those cancers with METHODOLOGICAND MEASUREMENTISSUES sufficientevidence for review: thoseof the colon, rectum,esophagus/ gastriccardia, renal cell, endometrium,ovary, breast,prostate, thyroid, Existingmethods for assessingthe etiologicand prognosticinfluences lung, and head and neck. The specific objectives of this review of energybalance and body size on cancerhave limitations that should are to: be consideredwhen evaluatingthe epidemiologic evidenceand plan- ning future research.Briefly summarized,these limitations fall into . Summarizethe epidemiologic literature for the associationbetween threeareas: validity of exposureassessment, difficulty of comparisons cancer risk and obesity-relatedmeasures of weight or body mass acrossstudies, and insufficientsample sizes to explorefully the many index (BMI; definedmost commonly as weight in kilograms [kg] factorsthat relateto body sizeand fat massand that could potentially divided by height in meterssquared), fat depositionpatterns (com- modify observedassociations. monly approximatedby waist circumferenceor by the ratio of waist- The measurementof anthropometricindices is generallywell stan- to-hip circumferences[WHR] or by ratios of truncal to extremity dardizedand documented(Lohman et al., 1988;WHO, 1995).In fact, skinfolds), and weight change; the standardization,accuracy, and reliability of these measuresare . Highlight gaps in scientific knowledge for the association of betterthan for many other cancer-relatedexposures. Weight and height obesity-relatedfactors and cancerrisk; are the most standardizedmeasurements and least subject to variabil- . Identify the possibleunderlying biological mechanismsfor these ity. More recent investigationshave used interviewersto measure associations: variousanthropometric indices from study participants,hence reduc- . Comment on areasfor future research;and ing the possibility of random and systematic error. However, many . Estimate the public health impact of obesity on cancer risk at the of cancer have relied on weight and internationallevel. studies etiology self-reported height.Studies on self-reportedheight suggest a reasonabledegree of Becauseof spacelimitations, the epidemiologicevidence on height accuracywith a bias in overreportingheight that is somewhatgreater and cancerrisk is not includedin this review but has beenwell sum- in men comparedto women and that increaseswith age (Rowland, marizedin severalprevious reviews (Ballard-Barbash, 1999; Gunnell 1990)or is limited to peopleover age 60 (Kuczmarskiet a1.,2001), et al..2001). presumablydue to age-relatedloss in height.Studies on self-reported weight indicate the presence of significant misreporting at the extremesof weight, with heavier people underreportingand lighter METHODS people overreportingtheir weight (Rowland, 1990; Stevenset al., 1990;Must et al., 1993).Therefore, studies of chronicdisease that rely A search was conductedon MEDLINE and PUBMED for all publi- on self-reportedheight and weight will underestimatethe risk associ- cations on weight, body mass index, anthropometric factors, and ated with thesemeasures (Gunnell et al., 2000). Nonetheless,com- specific cancersin human populationsand was supplementedby a pared to estimatesof correlationbetween reported versus measured manual search of all major relevant journals. The literature search assessmentof otherexposures, such as dietary intake, correlation coef- includedall publicationsup to February2003. Studies included in rhis ficientsbetween recalled and measuredweight suggestthat weight is review focused on some aspect of anthropometric risk factors in recalledwith a reasonabledegree of accuracy.For example,in one relation to cancer risk for the cancersnoted above. Maior review U.S. study.correlations between reporled and measuredweishts fbr . =.; 422 '.;:ti ri; ,F*::- Obesity,end BodyCcmposition 423 elderly subjectsasked to recall weight currently and from 4 and 28 COLORECTALCANCER years previously were reponed to be 0.98 for current,0.94 for 4-year, and 0.82 for 28-year recall (Stevenset al., 1990).Measurement of Summaryof Findings skinfoldsand circumferencesis lessreliable than the measurementof Data from case-control and cohort studies provide convincing evi- weightand height (Lohman et al., 1988;WHO,1995). However,reli- dence of an approximately twofold increasein risk of colon cancer ability is improvedwith standardizationin measurementtechnique. It among men with a BMI of 30 or more. Risk estimatesfor men are is not feasible to obtain these anthropometric measuresfrom self- most commonly reported to be about 2. Among women, less exten- report.Waist circumference is currentlyconsidered to be the mostcon- sivedata provide evidence of a probablyincreased risk of coloncancer venient and simple measureof abdominal or central adiposity for from a high BMI, with risk estimatesusually between 1.0 and 1.5. epidemiologicresearch (WHO, 1995,2000). Percentbody fat can be Among obese women who are estrogen positive (defined as pre- estimatedfrom bioelectricimpedance but only has been reportedin menopausal or postmenopausaland taking hormone replacement one studyof canceretiology (Lahmann,2003). Bioelectric impedance therapy [HRT]), a twofold increasein risk has been reported.Most of is currently the most promising field method for estimating body the literaturehas found no associationbetween BMI and rectal cancer. composition in large epidemiologic studies becauseit is portable, Meta-analysesthat provide a more quantitativestatement of the asso- inexpensive,easy to use, and highly precise.Issues related to use of ciation acrossseveral studies have not beenpublished. different methodsfor estimatingbody compositionfor chronic disease epidemiology are well summaized by Baumgartner and colleagues (l99s). Overview Comparisonacross studies is difficult becausemost studieshave Many aspectsof the associationsbetween body weight, obesity, and examined risk by quantile distributions(most commonly tertile or colorectal cancer have been examined. Colorectal cancer has been quartile) for the BMI or other anthropometricmeasures used. As the studied by examining risk associatedwith colon and rectal cancers distributionof BMI varies acrosspopulations, these quantile groups combined, as well as for each site separately.Although many studies are not comparableacross studies, making comparisondifficult. With have examined associationswith colon cancer,few have considered increasinginterest in understandingthe risk for many chronicdiseases rectal cancer specifically.Additionally, unlike many cancers,associa- by standard WHO BMI categories,investigators have begun to tions betweenbody size and the precursorlesion, adenomatouspolyps, examinerisk by thesecategories: underweight as BMI of less than have been reported. Several consistentdifferences in associationsby 18.50,normal weight as BMI of 18.5G-24.99,overweight as BMI of gender, site, and location within the colon have been observed. 25.00 or higher. This latter overweight category is further subdivided Stronger associationshave been observedfor men than for women; into four categories: preobese, 25.00-29.99, obese class I, for colon than for rectal cancer; and, within the colon, for distal than 30.00-34.99;obese class II, 35.00-39.99;and obeseclass III, >40.00 for proximal tumors. (WHO, 2000).More recentmeta-analyses have the ability to examine Studiesexamining the associationsbetween body size, obesity,and risk by thesebroad weight categories;however, analyses should not colorectal cancer have used many different indicators of body size. be limited to thesecategories as risk may vary within them, depend- Most studies have relied on BMI, usually from several years before ing on the chronic diseasesand the populationsexamined. diagnosis.Few studies have used the current WHO criteria for over- Researchto date suggestsoften complex and varying associations weight (BMI 2 25) and obesity (BMI > 30), making comparisonsand of body size with different cancersites. Although canceras a whole interpretationsacross studies difficult. Most studieshave categorized now exceedscoronary artery diseaseas a causeof deathin the United BMI basedon distribution in the population, and although the lower Statespopulation under age 85, the incidenceof site-specificcancer limit for the upper category of BMI is often near 30, at other times it is much lower than diseasessuch as diabetesmellitus and coronary is around 26. Limited data exist on waist and hip circumferencesas arterydisease, limiting the samplesizes of studiesfor less common indicatorsof fat pattern and its impact on colorectalcancer risk. Most cancers.Therefore, delineation of the associationof body size and studiesfocus on adult BMI, with little information availableon weight cancerrequires site-specific studies, which can often be done only changeor early life body size as possiblepredictors of risk. Because through multicollaborative efforts for less common cancers.Further- of differencesin reportedassociations for colon and rectal cancer,evi- aa more, as more factors are identified as potentially influencing body dence for these is E sites summarizedseparately. 'i: size and cancer associations,delineation of the underlying mecha- * nisms requires further subgroup analysis, thereby increasing sample i1 Colon Cancer !r: sizerequirements. :ii The "gold standard"technologies required, such as doubly-labeled Epidemiology i'' :€.i water for measuringenergy expenditureor computenzedtomography Many studies have evaluated associationsbetween body size and for measuring body composition, are expensive,not easily portable, colon cancer, and reported associationshave been fairly consistent A'j and, thus, not feasible for large samples.Therefore, it is not possible across case-control and cohort studies. Studies that have examined i?,,'' 14- to directly measure positive energy balance in large epidemiologic associationswith obesity report slightly stronger associationsthan i;: studiesof cancerrisk or prognosis.In addition,although it is possible those for overweight. For men, most studies report significant T.i '.&.*::'. to estimate energy balance from its components-energy intake and increasedrisk, with risk estimatesof 2.0 or greater,although risk esti- gri. expenditure-current methodsto assessdiet and physical activity rely matesrange from 1.2 to 3.0 (Grahamet al., 1978;West et al., 1989; i*-f, on self-reportof food intake and physical a+ activity, a methodologywith Gerhardssonde Verdier et al., 1990 Le Marchand et al., 1992: I*e '7t,. '*,. substantialreporting error. Consequently,because large epidemiologic and Paffenberger,1992a; Giovannucci et al., 1995; Caan et al., 1998; *:,. studiescannot measureenergy balancethrough either doubly-labeled Singh et al., 1998; Schoenet al., 1999) (Figures22-Ia and 22-lb). sj'a.: water techniquesor through accurateestimation of energy intake and Most studieshave adjustedfor important confounding factors,includ- :!' !4. cxpenditure from self-reports,the most valid measureof either per- ing physical activity, dietary intake, and smoking, although some *,*-- sistent or recurrent statesof positive energy balance is weight gain report age-adjustedestimates of associationonly (Lund Nilsen and % during adult life. Given findings from clinical metabolic researchof Vatten,2001). Associationsobserved for women are generallyweaker ls' g.f: complex interactionsamong different steroid hormones,such as sex than those observedfor men and are often not statisticallysignificant. ffi steroidsand insulin, ranging from receptor cross-reactivityto postre- For women, risk estimates have generally ranged from 0.7 to 2.7 ceptor potentiation, statistical modeling methods that allow for this (Ponerand McMichael, 1983;Graham et al., 1988;Kune et al., 1990; F complex interactionare neededand have not been used.In the future, Bostick et al., I994;Dietz et al., I995;Le Marchand et al., 1997;Mar- fr- statisticalapproaches for complex systems,including factor analysis tinez et al., 1997; Caan et al., 1998; Singh et al., 1998;Ford,1999; and assessmentof effect modification and interaction, may advance Terry et al., 2001, 2002). Some studies report a similar twofold understandingof the metabolic factors underlying body size and increase in risk of colon cancer for both men and women for the cancerassociations. highestBMI quantile examined (approximately30) (Westet al., 1989; 424 PARTlll: THECAUSES OF CANCER

Cohort Studies: Population-Based

Leeand Pafienbarger(1992a), USA, n=290

Le Marchander a/. ( l 997), USA, n=azl

Chyouelal. (1994),USA, n=289

Giovanntrci el ar.(1995), USA, n-203

Singh& Fraser(1998). USA, n=59

Ford(1999), 951, n=t04

Lund Nilsenand Vatten(2001), Norway, n=234

Case-Control Studies: Hospital-Based

Russoeral. (1998), ltaly, n=687

Case-ControlStudies: Population-Based

Grahamer a,.(1988), USA, n=205

Gerhardssonde Verdietet al. (1990),Sweden, n=233

wesr elal. (1989).UsA. n=112

Kuneet a/.(1990), Australia, n=388

Caaneral. (1998), uSA, n=1,095

A

Cohort Studies: Population-Based

Bostickelal. (1994),USA, n=212

Matlinezet al. (1997),USA, n=393

Singh& Fraser(1998), USA, n=83

Ford(1999). USA. n=ll8

fe(ry et al. (2001).Sweden, n=29t

lund Nilsenand Vatten(2001). NoMay, n=277

fefty et al. (2t02). Canada,n=383

Iefi,J el al. (2002).Canada, premenopausal, n=118

Te(ryel al. (2002),Canada, postmenopausal, n=184 CasecontrolStudies: Population-Based

Weslefa,. (1989), USA. n=119

Gratlamet al. (1988).USA, n=223

Gerhardssonde Vedie( el al. {1990},Sweden. n=189

Kuneet al.(1990), Ausrralia, n=327

Caal' et al. (1998),USA, n=888 Figure 22-1. A, Summaryof risk estimatesand Slatl.eryet al. (2003),USA, estroQeo+, n=301 confidenceintervals from epidemiologicstudies of BMI and colon cancerin men.B, Summaryof risk slaneryeral. (2003), usA, estrogen-,n=576 estimatesand confidenceintervals from epidemio- logic studiesof BMI and colon cancerin women. B

Ford, 1999). Few studieshave included African Americans. The one Schoen et al., L999). Similar results have been observed from both study reporting associationsfor African Americans was limited to 99 case-controland cohort studies, and suggest that a higher WHR casesand reported no significant associationbetween BMI and col- increasesrisk of colon cancer. orectal cancer(Dales et al., 1979). Some studies have had sufficiently large sample sizes such that BiologicalMechanisms and TumorMutations associationsfor proximal and distal colon tumors could be considered Associationsbetween body size and colon cancer risk may be modi- separately. Stronger associations for distal tumors than proximal fied by severalfactors. Interactions between BMI, physical activity, tumors are often seen(Dietz et al., 1995; Martinez et al., 1997: Caan and hormonesappear to provide the most insight into possiblebio- et al., 1998),although the studyby Le Marchandet al. (1997) showed logical mechanisms.High levels of vigorous physical activity have strongerassociations with proximal rather than distal tumors, and the been shown to modify the risk associatedwith obesity;for example, study of women by Terry et al. (2001) did not observe a significant risk is not as markedly increasedamong obesemen and women who associationfor either proximal or distal tumors. [n a study by Le are physically active (Lee and Paffenberger,1992a; Le Marchand et Marchand et al. (1992), the greatestrisk associatedwith a high BMI al.,1997:Slattery etal.,1997). Interactions between obesity and estro- was observedfor sigmoid tumors, which most studiesinclude as part gen or HRT may explain genderdifferences in risk associatedwith of the distal colon. BMI (Teny et al., 2002).Studies have shown that being overweight Some studieshave examinedwaist-to-hip ratio and colon cancer or obeseincreases risk of colon canceronly among women who are risk (Martinez et al., 19971'Caan et al., 1998; Russo et al., 1998; premenopausal(Terry et al., 2002). In another study that examined Obesityand BodyComposition 425 estrogen-positivewomen (defined as premenopausal or post- ADENOCARCTNOMASOFTHE ESOPHAGUS (AE) AND menopausaland taking HRT), risks associatedwith a high BMI were GASTRTCCARDTA (AC) similar to associationsobserved in men, with risk estimatesfor a BMI of 30 or more being twofold greaterthan that of women who are lean Summaryof Findings (Slatteryet al., 2003).However, women who were estrogennegative Data from case-controland cohort studiesprovide evidenceof a prob- (definedas postmenopausaland not taking HRT) did not experience able and fairly consistentassociation of a twofold or greaterincreased increasedrisk from being overweight or obese.In men it has been risk of adenocarcinomaof the esophagusand gastric cardiafrom over- observedthat, with advancingage, the risk associatedwith being weight and obesity.Most studieshave evaluated risk using a BMI of overweightor obesedeclines (Slattery et al., 2003).This changein around 30 or more as the upper level of BMI, although some studies risk over time could be the result of declining androgenlevels that use a much lower BMI as the upper level. The increasedrisk associ- operatein a similar fashion as estrogenin regulatingrisk associated ated with obesity has been observedin both men and women. Meta- with obesity.These interactions between physical activity and BMI, analysesthat provide a more quantitativestatement of the association and estrogenand BMI, suggestthat at leastone way BMI may influ- acrossseveral studies have not beenpublished. ence colorectalrisk is through its influenceon estrogenand on the interactionbetween estrogen and insulin. Evaluationsof the associationbetween BMI and colon cancer Epidemiology that have examinedspecific tumor mutationshave provided insight gastric into possiblemechanisms of action and the role of body size in pos- The incidence of adenocarcinomaof the esophagusand cardia siblespecific disease pathways. The few publishedstudies suggest that has risen dramaticallyin the past two decadesin developednations (Li BMI may be involvedin severalpathways. One studyreported that an andMorbahan, 2000; Walther et al.,2001). For reasonsnot entirely elevatedBMI was associatedwith Ki-ras mutationsin codons12 and clear, body size and obesity are associatedwith AE and AC, with the 13 (Slatteryet a1.,2001a).The BMI associationsappeared to be more increasesin the incidenceof thesediseases paralleling the increases specificfor Ki-ras mutationsin women than in men. BMI appearedto in obesity observedin Westerncultures. be equallyassociated with tumors with and without p53 mutations An important feature of the observedassociations between AE and (Slattery et al., 2002). Obesity was reported as being associatedonly AC and body size is that they are limited to adenocarcinomasand gen- with tumorsthat were stable(i.e., negative for microsatelliteinstabil- erally are not observedfor squamouscell canceror other histological (Vaughan ity) in women, although in men, obesity was associatedwith both types et al.,1995,Chow et al., 1998;Lagergren et al., 1999). stableand unstabletumors (Slattervet al.. 2001b). However, one study reporteda sixfold increasedrisk of squamouscell esophagealtumors and no increasedrisk of AE or AC adenocarcino- mas among individuals with a high BMI (Kabat et al., 1993). RectalCancer In some studies,tumor location seemsan important featurefor AC, with associationsstronger for tumors located more distally (Ji et a1., The small numberof studiesthat haveexamined associations between 1997).Associations have remainedafter adjustmentfor important con- body sizeand rectal cancergenerally report resultsfor few cases.From founding factors, such as cigarette smoking and diet (Brown et al., existing data it appearsthat body size is not associatedwith rectal 1995;Chow et al., 1998;Wu et a1.,2001). canceramong men or women (Grahamet al., 1988;Gerhardsson de Most studiesshow strongerassociations for AE than for AC (Figure Verdieret al., 1990;Dietz et al., 1995;Le Marchandet al.,1991:Russo 22-2); risk estimatesare inconsistentand rangefrom the null to sixfold et al., 1998; Terry et al., 2001). In contrast,one study of women or greater with BMIs of 27 or higher. The study by Lagergrenet al. reportedreduced risk of rectal canceramong thosewith a BMI of more (1999)revealed that associations were stronger forAE (OR = 7.6;95Vo than 25 relative to those with a BMI of less than 22 (Potter and C.I., 3.8-15.2)than for AC (OR = 2.3;95VoC.1., 1.5-3.6); among McMichael, 1986).Unlike colon cancer,studies of rectalcancer have peoplewith a BMI of more than 30, the odds ratio was 16.2 for AE not evaluatedinteractions between BMI and other indicatorsof body (95VoC.1., 6.341.4). In some instances,the reportedBMI level at size with diet and lifestyle factors.This lack of examinationof inter- which a risk is observedis modest.For example, in one study of AE action is most probably becauseof the relatively limited number of in British women, a sixfold increasedrisk was observedfor a BMI rectal cancercases available for analysis,making studiesof interac- higher than 22.7 (Cheng et al., 2000). In the study by Chow et al. tion impreciseand problematic.Evaluation of BMI and other indica- (1998), a three- to fourfold increase in risk of AE and a twofold tors of body size with specific mutations in tumors have not been increasein risk of AC was observed with a BMI of about 27.5 or reported. greater.Limited data supporta dose-responseeffect, with higher BMIs resulting in higher risk. Wu et al. (2001) observeda significant dose- response,with risk increasingas BMI increasedduring all periods of ColorectalAdenomas adult life (i.e., for BMI at ages20, 40, and recent adult BMI). Weight predictor Studiesevaluating the associationbetween body size and the occur- gain of 46 pounds or more was shown to be as important a (Ji renceof adenomashave examined associations by polyp sizeand type. as weightitself et al., 1997;Chow et al., 1998). Most studiesof colorectaladenomas have examinedBMI alone and Body size at young age and 20 years before diagnosis also appear lr'il (Ji have observedsimilar associationsas for colon cancer(Neugut et al., to be important contributors to risk et al., 1997;Lagergren et al., 1991;Shinchi et al., 1994:Davidow et al., 1996;Giovannucci et al., 1999).In some studies,the risk estimatesobserved for body size at younger 1996). Some studies do not detect associationswith adenomasand ages are larger than those for recent body size. In the study (2000) years body size (Terry et al., 2001,2002). In a study that included both of British women reportedby Cheng et al. BMI at age20 cancerand adenomas,an increasedrisk with a high BMI was observed appearedto be an important contributor to risk. In addition, associa- for large adenomasonly, but not for colorectaltumors (Boutron-Ruault tions havebeen reported as strongerfor nonsmokersthan smokers,and (<59) etal.,2001). Most studieshave observed a 1.5-to 2.5-fold increasein for younger people than older people, and as similar for those risk of adenomasamong the group with the highestBMI. A largeWHR with and without gastroesophagealreflux disease,for men and women, (Lagergren Tretli also has been associatedwith adenomas,with stronger associations and for various levels of education et a1., 1999; being observedfor larger adenomas(Kono et al., l99l; Shinchi et al., et al., 1999). 1994; Kono et al., 1999). Similar to studies of rectal cancer, studies of adenomashave been limited in terms of examininginteractions with BiologicalMechanisms other diet and lifestyle factors,primarily becauseof small samplesizes availableto examineinteractions. One studyby Martinezet al. (1999) The underlying biological mechanismsfor the associationbetween observedno differencesin risk for adenomasamons those with and BMI and AE/AC are not clear. Although it has been proposed that the without Ki-ras mutations. mechanisminvolves esophagealreflux, limited data show that the risk 426 PARTIII: THECAUSES OF CANCER

Gohorl Studies

Tretli& Robsahm(1999), Norway, males, n.94

Tretli& Robsahm(1999), Norway, females, n=25

Case-ControlStudies: Hospital-Based

Kabatelal.(1993), USA, males, n=121

Thanget al-(1996), USA, males and females, n=95

Gaiusetal (2001),ltaly, Switzerland. tefiales, n=114

Inoueelal. (2002), Japan, females. n=365

Case0ontrol Studies: Population-Based

Btownet al. (1995), USA, males, n=l6l

Vaughanefal. (1995), USA, AE, males and fema,es, n=131

Vaughaneral 0995),USA, AC, males and females. n=164

Jiela,. (1997), China, AC. males, n=]48

Jie, al 0997),China, AC, females, n=37

cnowet al. (1998), USA, AE, males, n=244

chonet al. (1998), usA, AE, females, n=48

chowelal. (.|998), usA,Ac, males, n=223

Chowetal 0998),USA.AC, females, n=38

Lagergrenetal. (1999),Sweden.AE, males and females,n=]89 (RR=7.6)

Lagergrenet at (1999),Sweden, AC, malesand females,n=262 € .3 ,4 Cheng el r/. (2000).United Kingdom. AE, females,n=68 (RR=6 0) .i 'a Figure 22-2. Summary of risk estimates and confi- i': Wu eI al. (2001),USA, AE, malesand lemales. n=2'12 dence intervals from epidemiologic studies of BMI and esophageal or gastric cardia adenocarc.inomain Wu ef ai (2001),USA, AC. males and females,n=266 men and women.

,".t .:t

associatedwith BMI is the samefor thosewith and without sympto- et al., 1995).In that study,the relativerisk for women in the top quar- E matic reflux disease(Lagergren et al., 1999). Other hypothesesfor tile of BMI was 2.2 comparedto a risk of 1.2 for men in the top quar- t) obesity potentially increasing risk include effects of diabetes or tile of BMI. A similar patternwas observedin the very obesegroup, insulin-relatedmechanisms (Cheng et al., 2000), and the delay in with higherincreases in risk observedfor womencompared with men. :'f esophagealmotility associatedwith obesity (Li and Mobarhan, 2000). For very obesewomen with a BMI in the top 57o (>38.1), risk was markedly increasedto 3.6. In contrast,for a similar group of very obesemen, risk was 1.6.In a meta-analysisincluding I I studies,risk RENALCELL CANCER of incident renal cell cancer was increasedby 67o and 7Vo for each Summaryof Findings unit increasein BMI for men and women, respectively(Bergstrdm et a1.,2001). This increasein risk correspondsto a 367o increase Data from case-control and cohort studies provide convincing evi- in risk for an overweight person and an 84Vo tncreasefor an obese denceof a consistentassociation of increasedrisk of renalcell cancer person. from overweight and obesity that is greater in women than in men. Data arelimited on adultweight gain andrenal cell cancer,and they Estimates from a meta-analysisthat calculated risks for men and have not beenreported except for threestudies. In a 1995multicenter women combinedsuggest this increasein risk correspondsto a 36Vo study,Mellemgaard et al. (1995) examinedthe slope and variability increasefor an overweight person and an 84Voincrcase for an obese (coefficientof variation)of BMI, two measuresthat investigatorshave person(Bergstrom et al., 2001). conceptualizedas representingthe rate of weight gain and frequency of weight fluctuations,respectively. In this study,the rate of weight Epidemiology changeappeared to be an independentpredictor of risk for women but not for men, though the coefficient of variation of BMI was not asso- Risk for renal cell cancer is increasedin heavier men and women ciatedwith renalcell cancerafter adjustmentfor BMI in women.One in virtually all studies(Wynder et al., 1,974:McLaughlinet al., 1984, otherstudy. by Chow et al. (2000),examined the associationof change 1992; Goodmanet al., 1986; Yu et al., 1986; Asal et al.. 1988: in BMI over a periodof 6 yearsand renalcell cancerrisk in men and Kadamini et al., 1989; Maclure and Willett. 1990; Partanenet al., found a non-statisticallysignificant increasedrisk of 1.6 with an l99l; McCredie and Stewart, 1992; Benhamouet al., 1993; Finkle increasein BMI of 2.5 units.Additionally, Bergstrdm et al. (2001) et al., 1993;Kreiger et al., 1993;Hiatt et al., 1994; Lindblad er al., observedthat weight gain. especiallyamong subjectswith high BMI 1994; Mellemgaardet al., 1994, 1995; Muscat et al., 1995; Chow at age20 (OR = 1.9,C.I., 1.2-3.0),was associatedwith an increased et al., 1996,2000; Boeing et al., 1997;Yuan et al., 1998).The one renal cancerrisk. exceptionis a hospital-basedcase-control study (Talamini et al., 1990) A numberof other factors,including hypertension,diabetes melli- (Figures 22-3a and 22-3b). In contrast,no associationbetween body tus, and tobaccouse, have been consistentlyassociated with risk of size and tumors of the renal pelvis has been demonstrated(McCredie renal cell cancer.Each of thesefactors is relatedto obesity:Obesity and Stewart, 1992; Chow et al., 2000). Most studiesdemonstrate a is positively and causallyrelated to hypertensionand diabetesmelli- linear dose-responserelationship between body weight or BMI and tus, and tobaccouse is associatedwith leanness.A lirnited numberof renal cell cancer.However, the increasein relative risk appearsto be studiesof obesity and renal cell cancerhave adjustedfor hyperten- higher for women than for men, perhaps best demonstratedin the sion and/orsmoking (Kreiger et al., 1993:Behnamou et al., 1993; largestmulticenter study that included men and women (Mellemgaard Mellemgaardet al., 1994;Muscat et al., 1995;Chow et al., 1996,

.a . 'l:. ...ti Obesityand BodyComposition 427

Cohort Studies

Hiattet al.(1994), USA, n=163

Chowelal (2000),Sweden, n=759

Case-Control Studies: Hospital-Based

Goodmanelal (1986),USA, n=173

Benhamouelal. (1993), France, n=138

Muscater al ('1995).USA. n=543

Case-Control Studies: Population-Based

McLaughlinet al.(1984), USA, n=310

Asaletal (r988),USA, n=209

Maclure& Wille((1990), USA, n=135

McCredie& Stewart(1992), Ausrralla, n=307

McLaughlinet al.(1992),China, n=76

Kreigereral (1993),Canada. n=282

Lindbladet al. (1994). Sweden, n=207

Mellemgaardet al.(1994), Denmark, n=225

Chowe1rl (1996).USA. n=27a

Yuanet at (1998),USA, n=781

A CohortStudies

Finkleet al. (19€3),USA, n=161

Hia[ er al (199a), USA,n=88

Case4onhol Studies: Hospital-Based

Goodmanet al. (1986),USA, n=71

Benhamouel at ('1993),France, n=58

Muscarelal (1995),USA. n=2a5

CaseControl Studies: Population.Based a:":.:. McLaughlineta,.0984), USA,n= 178

Asalet at. (1988),USA, n=100

Maclure& Wilbn (1990),USA, n=68

McCredie& Steilarl(1992), Australia, n=I73

McLaughlinel ar.(1992). China, n=58

KJeigetet al.(1993). Canada, n=181

Linduadefal (1994),Sweden. n=172

Mellemgaardelat (1994),Denmark, n=141 Figure 22-3. A, Summary of risk estimates and

Chof/ et at. (1996),USA. n=163 confidence intervals from epidemiologic studies of BMI and renal cell cancer in women. B, Summary Yuanet al.(1998). USA, n=a23 of risk estimates and confidence intervals from epi- demiologic studies of BMI and renal cell cancer in B men.

2000). However, none has fully explored potential interactionsamong et al., 1987), or through paracrine growth factors (Concolino et al., these factors. Limited data suggesrthat risk from obesity is higher 1989).In a 1996 review on nutrition and renal cell cancer,Wolk et al. among heavier men and women with hypertension(Muscat et al., (1996) also proposedthe hypothesisthat the metabolic syndromeasso- 1995)and who arenonsmokers (Chow et al., 2000).A limited number ciated with upper body obesity,including hypertension,glucose intol- of studieshave found amphetamines,used in the treatmentof obesity erance, increasedlevels of insulin-like growth factor (IGF), and in and other conditions,to be positively associatedwith renal cell cancer women, anovulation and increasedandrogen production, might result (Wolk et al., 1996). in renal damage that would increasethe susceptibility of the kidney to other carcinogens. BiologicalMechanisms The precisemechanisms by which higher body massmight influence ENDOMETRIALCANCER renal cell carcinogenesisare not clearly defined. Hormonal factors, Summaryof Findings particularly sex steroids,have been found to promote renal cell pro- liferation and growth by direct endocrine receptor-mediatedeffects Data from case-controland cohort studies provide convincing evi- (Ronchi et al., 1984),by regulation of receptorconcentrations (Traish dence of a consistent association of increased risk of endometrial 428 PARTlll: THECAUSES OF CANCER cancerfrom overweight and obesity with a doubling to tripling of risk Tdrnberg and Carstensen(1994) observed a stronger association observedamong obesewomen. The lifetime risk of being diagnosed betweenbody massindex and endometrialcancer risk with increasing with endometrial cancer in the United States is about 3Vo, whereas agethan haveothers (Furberg and Thune,2003). In a studyof 47,003 the cumulative risk increasesto 97o to lj%o in obese women women who were older than 55 years at entry and followed for 25 (Schottenfeld,1995). Based on the prevalenceof obesity (Flegal et al., years,Tornberg and Carstensen(1994) observeda threefold increased 2002) and estimatesof relative risk in U.S. women, excessiveweight risk (RR =3.16; P < 0.0001)when comparingheavy to leanwomen and a central pattern of fat distribution may account for ITVoto 467o (BMI 2 28 vs. BMI > 22). of endometrialcancer incidence in postmenopausalwomen. Data Consistentwith the observationfrom cohort studies,mosf case- suggest that obesity may account for about 40Vo of the worldwide control studies(Blitzer et al., 1976, La Vecchiaet al., 1984, l99l', variation in cumulative rates of endometrial cancer (IARC, 2002). Lapiduset al., 1988;Austin et al., l99l; Shu et al., l99l; Brinton Meta-analysesthat provide a more quantitative statementof the et al., 1992:Leviet al., 1992;Shu et al., 1992;Swanson et al., 1993: associationacross several studies have not beenpublished. Inoueet al., 1994;Olson et al., 1995;Baanders-van Halewyn et al.. 1996;Kalandidi et al., 1996;Shoff and Newcomb et al., 1998)except (Koumantaki Parslov al., 2000;Beard et al., 2000) Epidemiology three et al., 1989; et observedthat womenwith a BMI of 25 or higherhave a two- to three- Adenocarcinomaof the endometrium is the most common cancer of fold increasein risk (Figure224). High BMI was associatedwith an the female reproductiveorgans. International variation in endometrial even greaterexcess risk in a study in Hawaii that includedJapanese, cancer rates may representdifferences in the distribution of known Caucasian, native Hawaiian, Filipino, and Chinese populations risk factors, which include obesity, hormone replacement therapy (Goodmanetal.,l99l). Comparedto leanwomen (BMI <21,.1),over- (HRT), ovarian dysfunction, diabetesmellitus, infertility, nulliparity, weight women (BMI > 27.3) had a fourfold increasedrisk (RR = 4.3, and tamoxifen use (Purdieet al., 2001). Comparisonsof cumulative P < 0.0001).Another large, informative Swedishcase-control study rates (ages 0-84) of endometrial cancer with reported data on the with 709 endometrialcancer cases found an even largerincreased risk prevalenceof obesity among women in developedand developing of endometrialcancer with increasing weight (Weiderpasset al., countries suggest that obesity may account for about 40Vo of the 2000). Comparedto lean women (BMI < 22.5), obesewomen (BMI worldwide variation in cumulative rates of endometrial cancer of 3G-33.99)had a threefoldincreased risk (OR =2.9, C.1.,2.0-4.0). (Akhmedkhanovet al., 2001).Thus, a substantialportion of the inter- and markedlyobese women (BMI > 34) a sixfold increasedrisk (OR national variation in the incidenceof endometrialcancer could be = 6.3, C.L, 4.2-9.5). The effect of BMI did not vary by age. explained by differencesin the prevalenceof obesity. menopause,or use of contraceptives.Thus, a linear increasein risk A consistentpositive association between obesity and endometrial with increasingBMI has beenobserved. cancer risk has been observedin most cohort studiesfocusing on Weight gain during adulthoodis especiallyinteresting, and studies this association(Ewertz et al., 1984:'Folsom et al., 1989, Tretli show a linear dose-responserelationship between weight gain and and Magnus, 1990;Le Marchandet al., l99l; Moller et al., 1994; endometrialcancer (Le Marchand et al., l99l; Shu et al., 1992, Tcirnbergand Carstensen,1994:'Teny et al., 1999;Jain et al., 2000; Swansonet al., 1993;Olson et al., 1995;Teny et al., 1999).Two of Furbergand Thune,2003). Excess risk persistseven after adjustment thesestudies did not adjustfor young adult weight or BMI (Swanson for several factors, such as parity, HRT, and smoking (Le Marchand et al., 1993;Olson et al., 1995).The studiesthat adjustedfor young et al.. l99l: Jainet a1..2000).Both Le Marchandet al. (1991)and adult or baselineweieht or BMI found differine results.[n two studies

Cohort Studies: Population-Based

Ewenzet al. (1984),Denmark. n=t t 5

Tretli& Magnus(1990). Norway, n=2,208

LeMarchand etal- (1991), USA, n=214

Malle(et al. (1994).Denmark, n=il 4

Ieffy et al- (1999),Sweden, n=133

)ain et al. (2000),canada, n=221

Case-Control Studies: Hospital-Based

La Vecchiaelal (1984),ltaly, Postmenopausal, n=283

La Vecchiael a/.(1 984). ltaly, Premenopausal, n=283

Koumantakieral. (1989), Greece, n=83

Austinerrl. (1991),USA. n=l68 =___-_{-

Inoueel al (1994),Japan, n=143 -

CaseControl Studies: Population-Based

Casagrandee( al.(1979), USA. n=150 t

Shu etal (1992),China. n=268 _-_------{-

Brinronelal (1992),USA, n=405 ------t-

Swansonel al (1993),USA, n=405 --|-

Olsoneral (1995),USA, n=232 -

Sholl& Newcomb(1998), USA, n=723 ------f-

Weiderpasset al. Qcf,q, Sweden,n=709 (RR=6.3)

Bea'd et al. (2000),USA. n=1 1 7 Figure22-4. Summaryof risk estimatesand con- fidenceintervals from epidemiologicstudies of Xu(2002), Chrna, n=497 BMI andendometrial cancer. Obesityand BodyComposition 429 that adjustedfor either adolescentor early adult weight or BMI. the OVARIANCANCER association between adult weight ,eain and endometrial cancer remainedafter adjustment (Le Marchandet al., 1991;Xu et a1.,2002). Summaryof Findings ln oneother study, adjustment for baselineweight eliminated the asso- The results from cohort and case-controlstudies do not indicate ciationbetween adult wei-eht gain andendometrial cancer (Teny et al., any clear associationbetween overweight or obesity and epithelial 1999).In one study in Shanghai.weight loss from ages20 to 30 was ovarian cancerrisk. Severalrecent studies,however, suggesta possi- inverselyassociated with endometrialcancer risk (Xu et al., 2002). ble positiveassociation. The distributionof body fat, including such measuresas skinfold (subscapular)and WHR, also has been examinedin severalstudies Epidemiology (Folsomet al., 1989;Elliott et al.. 19901Austin. 1991 ; Schapiraet al., l99l;Shu et al., 1992:Swanson et al.. 1993).Evidence suggests that Epithelial ovarian cancer incidence varies between countries and fat distributionmay be importantin endometrialcancer, with upper- within countriesand ethnic groups,but only a small percentagehas body obesity particularlyincreasing risk. A case-controlstudy from been attributed to family history. The majority of ovarian cancer is Shanghaifound that the distributionof subscapularskinfolds was a sporadic,suggesting that part of the diseasemay be explainedby mod- better predictor for endometrialcancer risk than WHR, even after i fi able lifestyle-relatedfactors. adjustmentfor BMI (Shu et al., 1992). In contrastto somecohort studiesthat found no associationbetween Interactionbetween obesity and physical activity has been proposed older adult obesityand ovariancancer (Mgller et al., 1994;Tornberg asone explanation for the observationthat somestudies have observed et a1.,1994; Mink et al., 1996;Fairfield et a1.,2002),two studies(Lew a strongerpositive association between obesity and endometrial cancer andGarfinkel,1979, Fairfield etal.,2002) have observed a directasso- risk in older than youngerwolnen (Le Marchandet al., 1991:TOrn- ciationbetween obesity at someages and ovariancancer incidence. In berg and Carstensen.1994). This finding is consistentwith somepop- a recentreport from the NursesHealth Study with 402 casesof epithe- ulationevidence that older womenare less active than younger women lial ovarian cancer (Fairfield et al., 2002), a twofold increasein pre- (IARC, 2002). Recently,the establishedlink betweendiabetes and menopausalovarian cancerrisk was found among women with a BMI endometrialcancer risk suggeststhat hyperglycemia(Furberg and at age 18 of 25 or higher comparedto a BMI lessthan 20. However, Thune, 2003) and hyperinsulinemia (Weiderpasset al., 2000; no associationbetween recent BMI or adult weight changeand ovarian Andersonet al., 2001) mav influencerisk. cancerrisk was observedin this study. An inconsistentassociation also has beenobserved in case-control BiologicalMechanisms studies(Figure 22-5). Only six of thesestudies (Casagrande et a1., 1979;CASH, 1987;Farrow et al., 1989;Purdie et al., 1995;Mori et Changesin metabolismand hormonal activity that occur in obesity al., 1998;Lubin et al., 2003) found a directassociation, whereas other may be importanlmechanisms that explain the biologically plausible studiesfound no association(Shu et al., 1989; Hartge et al., 1989; link between increasedendometrial cancer risk and obesity. The Mink et al.,1996: Nesset al., 2000; Greggi et al., 2000; Lukanovaet normal menstrualcycle reflectsthe complex balancebetween the pro- al., 2002) and some a negative association(Byers et al., 1983; Par- liferative actions of estrogen and the antiestrogenicand secretory razzini et al., 1997;Riman et a1.,2001).One case-controlstudy that transformingactions of progesteroneon the endometrium(Hale et al., found no associationbetween BMI and ovarian cancer did observea 2002).A shift to a positive energybalance might contributeto an unfa- statisticallysignificant increased risk with elevatedWHR (Mink et al., vorable sex hormoneprofile in women (Key et a1.,2001;Jasienska 1996). et al.,2001). Among premenopausalwomen, obesity may inducea progesterone deficiencyduring the luteal phaseas a result of anovulatorycycles, BiologicalMechanisms amenorrhea,and irregular menstrualperiods. This situation may lead Severalreproductive factors, including the number of pregnancies, to an increasedproliferation and decreaseddesquamation of endome- early menarche,late menopause,and late first pregnancy,influence trial cells.Among postmenopausalwomen, severallines of evidence ovarian cancer risk and could potentially play a role in the suggestthat obesityis associatedwith increasedlifetime exposureto obesity-ovariancancer risk association.All of thesefactors exceptfor estrogenthat may also increaserisk: increasedaromatization of andro- the numberof pregnanciesmay increaserisk. Thus,it is plausiblethat gens (androstenedione)to estronein adiposetissue (Key and Pike, obesity may influence ovarian cancer risk through changesin sex I 988a; Zeleniuch-Jacquotte,200 I ), increasedsecretion of androstene- steroidslevels. However, at presentthe epidemiologic data have not dione in adrenalglands, and decreasedsex hormonebinding globulin established consistent direct association between overweisht or (SHBG), resultingin an increasein bioavailableestradiol (free and obesity and ovarian cancer. albumin bound).Therefore, through different mechanismsfor pre- and postmenopausalwomen, obesity alters sex steroid concentration and metabolism in ways that increaseendometrial cancer risk. BREASTCANCER Experimentaland epidemiological evidence suggest that insulin-like Summaryof Findings growth factor-l (IGF-I), a hormone with strong mitogenic and anti- apoptotic actions,may be important in endometrialcarcinogenesis. A Extensivedata from case-controland cohort studiesprovide convinc- high BMI may alter IGF-I blood levels, althoughthis has not been ing evidenceof a modest increasedrisk of postmenopausalbreast cell examined in studies of endometrial cancer risk. In addition, energy cancerfrom overweight and obesity and a larger twofold increasein restriction is known to enhance DNA repair, moderate oxidative risk for adult weight gain. Estimatesfrom a meta-analysisof cohort damageto DNA, andreduce oncogene expression (Lipman et al., 1989). studies found gradual increases in risk of postmenopausalbreast Anotherpossible biological mechanism is relatedto the established cancerto a BMI of 28, after which risk did not increasefurther. The link between diabetesand endometrial cancer (Furberg and Thune, relative risk for a BMI of 28 compared with a BMI of less than 21 2003) and the fact that obesity is associatedwith insulin resistance, was I .26 (van den Brandt et al., 2000). However, risk estimatesvary hyperinsulinemia,and diabetesmellitus. Evidenceis increasingthat by age at diagnosis,history of HRT, and estrogenreceptor statusof insulin is a growth factor for tumor formation (Yu and Rohan, 2000). the tumor. Another meta-analysisfound that this increasedrisk for The mechanismsunderlying insulin-mediatedneoplasia may include postmenopausalbreast cancer correspondedto a I2Vo increase for enhancedDNA synthesiswith resultant tumor cell growth, inhibition an overweight woman and a 25Vo increase for an obese woman of apoptosis,and an alteredsex hormone milieu. Hence,in overweight (Bergstrdm et al., 2001). Data show a stronger associationfor adult and obesewomen, coexisting metabolic disturbancesmay act syner- weight gain, with a doubling of risk among women who have never gistically to facilitate malignant transformationof glandular endome- used HRT and who gained over 20 kg from age 18 (Huang et al., trial cells. 1997).A meta-analysisof cohort studiesfound an inverse association 430 PARTlll: THECAUSES OF CANCER

Cohort Shdies: Poputation-Based

Mollet el al. (1994),Denmart, n=58

Tdrnberg& Cartensen(1994), Sweden, n=330

Mink er a/. (1996),USA, n=97

Faifield et al. (2002).USA, n=4OZ

Case'Control Studies: Hospital.Based

llton et al. (1988).Japan, n= 1t0

Paraztinielal. (1997),ltaly, n=971

Mon et al. (1998),Japan, n=89

Gase'Control Studies: Population-Based

Faftw eI al. (19$1, USA. n=27t

Shu el a/. (1989),China, n=1 ?2

Pu'die et al. (19951,Australia, n=824

Nesseral. (2000),usA. n=767

Rimanet a/.(2001), Sneden, AI histdogb types,n-.|93

Rirnanetar. (200'l),Sweden, Serous, n=l93 (RR=6.5)

Rimanef a/.(2001), Sweden, Mucinous, n=193

Lukanovaetal.(20021, USA, Sweden, ltaty, n=122 Figure 22-5. Summaryof risk estimatesand confi- denceintervals from epidemiologicstudies of BMI Lubinetal (2003),lsrael. n=1,269 and ovariancancer.

betweenBMI and premenopausalbreast cancer,with a relative risk of weight or BMI is assessedat the time of diagnosisor at earliertimes 0.54 for women with a BMI higher than 3l compared with women during childhood, adolescence,or adulthood (Hislop et al., l9g6; with a BMI of less than zl (van den Brandt et a1.,2000).This esti- Kolonel et al., 1986;London er al., 1989;Brinton er al., lgg}). Early mate is consistentwith the reduction in risk of 0.6 to 0.7 observedin studiessuggested that the protectiveeffect among heavier women was many studiesand does not appear to be presentfor BMIs lessthan 2g. limited to early-stagedisease due to poorer detectionof small tumors An extensive literature on breast cancer prognosis and survival sug- (Willett et al., 1985; Swanson et al., 1989). However, subsequent gests that women who are heavier at the time of diagnosis and galn studies in these same groups suggest that detection bias does not w_eightfollowing diagnosisexperience a worse prognosisinespective explain the increased risk for breast cancer observed among lean of menopausalstatus at diagnosis. premenopausalwomen (London et al., 1989; Brinton et al., 1992: swanson et al., 1996).A large case-controlstudy of 1588 casesfound Epidemiology that risk was increasedabout twofold among women who were tall and thin comparedwith women who were heavy and short (swanson Animal model research in the 1930s first examined the hypothesis et al., l'996). A mera-analysisof seven cohorts comprising 723 inci- that a positive energy balance increased risk for breast cancer dent casesof invasivebreast cancer in premenopausalwomen found (Tannenbaum, 1940a, 1940b). The epidemiological evidence on an inverseassociation between BMI andpremenopausal breast cancer; weight or BMI and breast cancer risk varies by menopausalstatus the relativerisk was0.54 for women with a BMI higherthan 3l com- (Figures 224a and 224b), ageat diagnosis, hormoner...pto, status paredto women with a BMI lessthan2l (van den Brandtet al., 2000). of the breast cancer, and exposure to exogenousestrogens. Because This estimateis consistentwith the reductionin risk of 0.6 to 0.7 studies of breast cancer have used so many different BMI cutpoints observedin many studiesand doesnot appearto be presentfor BMIs that are not consistentwith current wHo criteria of overweight and lessthan 28. obese, the following review uses the term heavier women ro conversely, most studies have found that heavier women are at describethe upper BMI groups rather than the terms overweight and increasedrisk of postmenopausalbreast cancer (Valaoras et al., 1969; obese. The most informative studies have distinguished between de Waard et al., 1974; Choi et al., 1978; paffenbergeret al., l9g0; pre- and postmenopausalbreast cancer; examinedthe effect of weight, Kalish, 1984;Lubin et al., 1985;Hislop er al., 1986;Kolonel er al., weight gain, and central body fat at various ages;and have examined 1986;Le Marchander al., 1988a;Negri et al., 1988;Tao et al., lggg; the differential effects of endogenous and exogenousestrogens. with Ingram et al., 1989; Tretli et al., 1989; Folsom et al., 1990; Hseih the emergence of a possible lGF-mediated pathway fbr several et al., 1990;Parazziniet al., 1990;Chuet al., l99l;Harris etaI.,1992; cancers, recent studies have also begun to explore the potential Pathaket al., 1992;Sellers et al., 1992,2002;Radimer er al., 1993; interactionsof IGF with body size. Francheschi et al., 1996; Yong et al., 1996; Ziegler et al., 1996; Most studies find that heavier women have a decreasedrisk of Trentham-Dietzet al., 1997;Chie et al., 1998: Galaniser al., l99g; premenopausalbreast cancer (Paffenbarger et al., l9g0; willett et al., Magnussonet al., 1998;Enger er al., 2000; Hall er al., 2000; Lam 1985; Hislop et al., 1986;Le Marchand et al., lggga; London er al., et al., 2000: Li et al., 2000: Trenrham-Dietzet al., 2000: Shu et al., 1989; Swansonet al., 1989; Tretli, l9g9; Bouchardy et al., 1990; 2001; Yoo et al., 2001;Friedenreich et al., 2002;Wenten et al., 20A2: Brinton et al, 1992 Harris et al., 1,992:pathaket al., rigz; vanen and Lahmannet al., 2003).A meta-analysisof sevencohorts comprising Kvinnsland, 1992;Tdrnberg and carstensen,1994; Francheschi et al., 3208 casesof invasivepostmenopausal breast cancer found gradual 1996; Swansoner al., 1996: Huang et al., 1991; Chie et al., l99g; increasesin risk to a BMI of 28 after which risk did not increase Coateset al., 1999; Peacocket al., 1999; Enger et al., 2000; Hall further; the relative risk for a BMI of 28 comparedwith a BMI of less et a1.,2000;de vasconceloset a1.,2001; Friedenreichet a1.,2002). than2l was I .26 (vanden Brandtet al., 2000).The majority of studies other studies find no association between BMI and premenopausal on BMI and breastcancer risk have adjustedfor major breastcancer breastcancer (Hirose et aI.,2001;Shu yob et al., 2001; et a1.,20011. risk factors,including reproductivefactors. Few studieshave exam- Relativerisks of approximately0.6 to 0.7 havebeen reported whether ined in detail the effect of confoundinsor interactionswith diet and *::lri1:1+';'t ii,ii

Cohoil Srudies: Hospital-Based

CohortStudies: Hospital.Based Yooetal (2001),Japan, n=l,154

Yooel al.(2001), Japan, n=1,151 Cohort Sludies: Population-Based Le Marchandel a/.(1988a), USA, age 50-51, n=145 Cohorl Studies: Population-Based Le Marchandet at (1988a),USA, age 55-65, n=l 35 LeMarchandel a,.(1988b), USA, age 30,49, n=149 Tretli(1989). Norway. n=5,122 I LeMarchandet a/.(1988b). USA,aqe 45-49,n=141 Sellersetal (19S2),USA, no familyhistory n=469, _--t- Tretli(1989), Norway, n=5,1 22 __+_ Sellerset al (1992),USA. family history, n=469. --'|- (1997), Vatlen& Kvinnsland(1992), Norway. n=291 -._|- Huangel al USA,n=1.517 Manjerel a[ (2001),Sweden, n=112 Huangel al (1991),USA, n= 1,000 --|- Sellerset al (2002),USA, no familyhistory, n=1.874 Manjeret al. (2001),Sweden. n=112 Sellersel al (2002),USA, familyhistory, n=1.874 Van Den Brandt et al. (20N\, Pooledanalysis of I studies.n=3,208 -_.|- Lahmannet al. (2W31,Sweden. n=246

Case.ControlStudies: Hospital.Based Van Den Erandtet al. (200fJ1,Pooled analysis of I studies,n=3,208

Hsiehelal. (1990).Intemational, n=3,993 Case'Control Studies: Hospital-Based Hsiehet a/.(1 990), International, n=3,993 Franceschietal. (1996),ltaly, n=988 -__t- Hamser al. (1992),USA, n=412 Chie et al. (1998),Taiwan, n=334 Francheschiel rt (1996),ltaly, n=988 -_|- de Vasconceloset a/.(2001 ). Brazil,n=1 77 Chieel al. (1998),Iaiwan. n=334

Hiroseet a/. (2001).Japan. no family history,n=861 Lamet al.(2000), USA, n=529

Hirosee, a/. (2001),Japan, family history n=65 de Vasconceloselal (2001),Erazil, n=177

Hiroseet al (2001),Japan. no lamilyhistory. n=604 _--+|- Case-Control Studies: Population-Based Hiroseel al (2001), Japan,family history, n=44 Hislopet al (1986),Canada, n=306 Case-Control Studies: Population-Based Chuer al (1991),USA. n=2.053 Hislopet al (1986),Canada, n=306 (1 Brinton& Swanson 992),USA, n=41 4 Kolonelel al (1986).USA, Japanese, n=272

Swansonetdl (1996),USA, n=I,588 ---l_ Kolonelel al (1986),tlSA, white, n=272 (1991), -|-. Yongel al (1996),USA, n=226 chu et al. usA, n=2.053 Erinton& Swanson(1992), USA, n=414 Liegleret al.(1996), USA, n=421 Yongef al (1996),USA, n=226 Coalesel al (1999),USA, n=1,590 _.----|-- Galanisef a,.(1998), USA, n=292 Peacocketal (1999),USA, n=845 Magnussonet at (1998). Sweden, n=2,904

Engeretal. (2000),usA. n=400 Engerer al (2000).USA, n=a00

Hallelal (2000).USA, 8lack. n=389 Hallet al.(2000), USA, 8lack, n=3gl

Hall et al. (2000),USA, White, n=391 Hall et al. (2000),usA, whire, n =389 _--l-- Li et al.(20001, usA, n=479 Shu eral.(2001), China, n=1,459 Trentham-Dieizet al.(20M1,. USA, n=5,659 Friedenreichel al.(2002\,Canada, n=462 Shue,al (2001).China, n=1.459

Wentenelal (2m2),USA. Hispanic, n=315

A Wentenet al (2002),USA, non-Hispanic White, n=372

Friedenreichet al.(20021, Canada, n=771

B

Figure 224. A, Summary of risk estimates and confidence intervals from epidemiologic 100 cases).B, Summary of risk estimates and confidence intervals from epidemiologic studies studies of BMI and breast cancer in premenopausal women (limited to studies with at least of BMI and breast cancer in postmenopausal women (limited to studies with at least 100 cases). 432 PARTlll: THECAUSES OF CANCER

physical activity. only one study in vermont that had data on breast (Paffenbargeret al., 1980;Willett et al., 1985;London et al., 1989; density from screening mammograms controlled for the effect of Folsomet al., 1990;Lund et al., 1990;Chu er al., l99l:Brinton and breastdensity (Lam et al., 2000).Because BMI is inverselyrelated to Swanson.1992; Sellers et al., 1992;Ursin et al., 1995;Huang er al., breastdensity, adjustment for breastdensity resultedin an increasein 1997:'Magnusson et al., 1998; Coateset al., 1999; Peacocket al., the risk estimationsat all levels of BMI; the oR increasedfrom 1.9 1999, Enger e[ al., 2000; Hirose et al., 2001; de Vasconcelos to 2.5 after adjustmentfor breastdensity among obesewomen. et al., 2001; Wentenet al., 2002). During the middle decadesof life, when examined,risk estimatesfor the associationbetween obesity the risk associatedwith BMI remainsinverse for premenopausalbreast and breastcancer vary by age at diagnosis,history of HRI estrogen cancerand increaseswith age for postmenopausalbreast cancer. receptor status of the tumor, and possibly family history of breast lncreasesin centraladiposity have been associatedwith a L4- to cancer.Risk has been found to increasewith age at diagnosisin some 2.O-foldincrease in breastcancer risk amongpostmenopausal women studiesthat include a substantialnumber of postmenopausalwomen in most studies(Ballard-Barbash et al., 1990a;Folsom et al., 1990: older than 65 years (Yong et al., 1996;van den Brandt et al., 2000; Schapiraet al., 1990;Bruning et al., 1992a,1992b; Sellers et al., 1992 Friedenreichet al., 2002).In one study,risk estimatesincreased from Ng, 1997;Shu et a1.,2001;Friedenreichet a1.,2002).However, not l.l amongwomen youngerthan 60 yearsto 1.8 amongwomen older all studies show this association(den Tonkelaar et al., 1992: Petrek. than 65 years (Yong et al., 1996). 1993; Lahmann et al., 2003). Data on cenrral adiposity and pre- The effect of exogenous estrogen or estrogen receptor status of menopausalbreast cancer do not suggesta consistentassociation tumors has been examined with stratified analyses in more recent betweenmeasures, such as waist circumferenceor WHR, and breast studies.In these studies,obesity-related risk has been higher among cancerrisk (Francheschiet al., 1996;Mannisto et aI., 1996:Swanson women who have never used HRT (Huang et al., 1997, Lam et al., et al., t996; Ng et al., 1997;Kaaks et al., 1998;Huang et al., 1999; 2000; Friedenreich et al., 2002; Lahmann et al., 2003). One of the Sonnenscheinet al.,1999;Hall et al.,20001Shu et al.,2001: Frieden- largest cohort studies in the United Statesfound a statistically sig- reich et al., 2002). only five of these studies observed statistically nificant BMI and estrogenreplacement interaction, with no increase significantincreases in risk (Mannisto et al., t996; Ng et al., 1997 in risk (RR of l.l) among all women bur an increasein risk (RR of Sonnenscheinet al., 1999;Hall et al., 2000; Shu et al., 2001).The 1.6) among heavier women who had not used HRT (Huang et al., associationin postmenopausalwomen may be modified by a family 1997).At leastthree studies have examined risk by BMI and estrogen history of breastcancer and ovarian cancer.In the Iowa Women's receptorstatus of the breasttumor (Engeret al., 2000;yoo et al., 2001; Health Study,among women with elevatedWHR, only women with Sellerset al., 2002).In one U.S. study,risk for a BMI of 27 compared a positive family history of breastcancer were at increasedrisk. The with a BMI of 22 was2.4 for tumorsthat were both estrogenand pro- combination of a high WHR with a family history of breast and gesteronereceptor positive (Enger et al., 2000).In anotherstudy, risk ovarian cancerwas associatedwith a more than fourfold increasein estimateswere 2.0 and 2.2 for a BMI of 30.7 comparedwith a BMI risk of breastcancer (Folsom et al., 1990;Sellers et al., l9g3). of 23 for estrogenreceptor positive and progesteronepositive rumors, The most consistentbody size predictorof postmenopausalbreast respectively(sellers et al., 2ooz). In both of thesestudies, obesity- cancerrisk is adultweight gain (Lubin, 1985;Le Marchandet al.. 1988b; related risk was not increasedfor estrogenand progesteronereceptor Ballard-Barbash,1990b; Folsom et al., 1990; Brinron et al., l99Z: negative tumors. In a Japanesestudy, risk did not vary by estrogen Radimeret al., 1993;Barnes-Josiah et al., 1995;Ziegleret al., 1996; or progesteronestatus of the tumor (Yoo et al., 2001). However, Huang et al., 1997;Magnusson et al., 1998;Jernstrdm and Barrett- the women in this studywere lean;the upperquartile of BMI of 22 in Connor, 1999; Enger et al., 2000; Li er al., 2000; Trenrham-Dietz this study is lower than the lowest quartile of BMI in most U.S. et a1.,2000;Shoff et a1.,2000;Hirose er al., 2001:Shu et al.,2001; studies. Friedenreichet al., 2002: Wenten et al., 2002: Lahmann et al., 2003). Data are very limited on variation in BMl-related risk for post- This associationhas been seenin cohort studiesthat found no asso- menopausalbreast cancer by family history. In several studies from ciation between BMI at baseline and subsequentdevelopment of the Iowa women's Healthstudy, heavier postmenopausal women with breastcancer and that also adjusted for baselineBMI (Ballard-Barbash a family history of breast cancer have a greater risk of deveroping et al., 1990a:Folsom et al., 1990;Huang et al., 1997).Findings from breastcancer than do heavierwomen without a family history(Sellers one of the largestcohort studiessuggest that a doubling of risk was et al., 1992;Sellers et al., 2OO2).Inone srudyin Japan,no differences associatedwith a weight gain of more than 20kg after the age of l8 were observedin associationsof weight, BMI, or changein BMI by years,but only among women who had never usedpostmenopausal family history for premenopausalbreast cancer, although somewhat HRT (Huanget al., 1997).Other studies have observed similar results, strongerassociations were seenfor weight and changein BMI and with increasesin risk eitherlimited to or much largeramong women postmenopausalbreast cancer among postmenopausalwomen with a who have gained more weight and who have never used HRT com- family history of breast cancer,confirming earlier results by Sellers pared to current users(Harris et al., 1992: Magnussonet al., 1998; and colleagues(Hirose et al., 20Ol). Only one study has examined Trentham-Dietzet al.. 2000: Friedenreichet al., 2002:Lahman et al., variation in BMl-related risk for premenopausalbreast cancer by 2003). Consistentwith findings for BMI and premenopausalbreast family history of breastcancer (Swerdlow et al., 20021.rn that study, cancer,weight gain appearsto be associatedwith a reducedor no sig- the inverseassociation commonly observedbetween BMI and breast nificantincreased risk of premenopausalbreast cancer in most studies cancerrisk was only observedin women without a family history of (Le Marchandet al., 1988a;London et al., 1989;Brinton et al.. 1992.. breastcancer. Huanget al., 1997;Coateset al., 1999;Peacock et al., 1999:Hirose Weight or BMI at birth, during childhood,and early in adulthood et a1.,2001;Shuet aI.,2001;de Vasconceloset a1.,2001;Friedenreich have beenexamined relative to breastcancer. The data on birth weight et al., 2002, Wentenet al., 2002). However, a study by Wenten et al. and breastcancer are limited by a very small numberof cases,with (2002) in New Mexico found differencesin risk for Hispaniccom- most studieshaving fewer than 100cases. Some studies find no asso- paredto non-Hispanicwhite women.[n that study,no associationwas ciation (Le Marchandet al., 1988b;Ekbom et al., 1997).or a non- found betweenweight gain and risk of premenopausalbreast cancer significantincreased risk (Hilakivi-Clarke er al., 2001); others find amongnon-Hispanic white women; in contrast,a non-statisticallysig- an increasein risk with increasingbirth weight for premenopausal nificantbut nearlytwofold increasedrisk wasobserved with morethan but not postmenopausalbreast cancer (Berstein, 1988; Ekbom et al.. l4kg of weightgain among Hispanic white women. 1992 Michels et al., 1996;Sanderson et al., 1996;Innes et al., 2000) The only study that has examinedthe effect of percentbody fat or a stronger increasein risk for premenopausalcompared to post- measuredby bioelectricimpedance as well as severalother measures menopausalbreast cancer (De Stavola et al., 2000: Kaijser et al., of body size,fat mass,and distribution found the strongestassociation 2001). In most studies,heavier weight or BMI during teenageand for percentbody fat, with a doublingof risk for women with a percenr young adulthood (18-20 years) is associaredwith a l\vo to 30vo body fat of over 36Vo comparedto women with a percent body fat decreasein breastcancer risk for both pre- and postmenopausalbreast of less than 27Vo.Similar to results reported for BMI and weight cancer.This decreasedrisk is most often not statisticallysisnificant gain. risk for percentbody fat was strongeramong wornen who had Obesityand BodvComposition 433 never used HRT. Amon-ewomen with a percentbody fat of more BiologicalMechanisms than 367c, the risk estimateamong women who did not use HRT was 3.4 comparedto 1.0 amon_swomen using HRT (Lahmannet al., OvarianHormone Metabolism 2003). The major focus of researchon mechanismsunderlying body sizeand breast and endometrial cancer risk has been the effects of adiposity on endogenous hormonal, predominantly estrogen, metabolism; BreastCancer Prognosis this researchhas been extensivelyreviewed (Kirschneret al., l98l; Extensivedata indicate that heavier women experience poorer survival Key and Pike, 1988b;Bernstein and Ross, 1993; Key et al., 2001). and increasedlikelihood of recurrencein most studies,irrespective Hypotheseshave evolved from a focuson estrogenexcess, to the com- of menopausalstatus and after adjustmentfor stage and treatment bined effects of estrogen and progesterone,and most recently, to (Greenberget al., 1985: McNee et al.. 1987;Hebert et al., 1988: attemptsto delineatefactors defining bioavailability and eff-ectsof Mohle-Boetaniet al., 1988;Lees et al., 1989;Verrault et al., 1989; estrogensand androgensand their metaboliteson specificend organs. Coateset al., 1990;Tretli et al., 1990;Kyogoku er al., 1990;Vanen Increasesin overall and central adiposity have been associatedwith et al.. l99l: Senieet al., 1992;Giuffrida et al., 1992 Bastarrachea increasesin insulin,androgens, and triglycerides;decreases in SHBG; (Evans et al., 19911'Zhanget al., 1995:den Tonkelaaret al., 1995:Maehle and increasesin total and free estradiol et al., 1984;Haffner and Tretli, 1996).The effect of weight or BMI on prognosisappears et al.. 1989;Kirschner et al., 1990;Kaye et al., l99l;Bruning et al., to be limited to or more pronouncedamong women with stageI and 1992a,1992b, Potischman et al., 1996).A numberof thesehormonal II disease(Verreault et al., 1989;Tretli and Magnus, 1990),estrogen changesincrease the bioavailabilityof estradioland its metabolites receptorand progesterone receptor positive status (Coates et al., 1990; and may also directly promotetumor growth. Giuffrida et al., 1992: Maehle and Tretli, 1996),and negative nodes The bioavailability of estradiol is dependent on the degree of (Mohle-Boetaniet al.. 1988;Newman et al.. l99l). The mostprecise binding and the strengthof binding to severalprotein carriers.SHBG risk estimatesfor BMI and breastcancer prognosis are derived from is the predominant protein carrier of estradiol and the percentageof large population-basedcohorts of breastcancer cases. In the largest free estradiol is generally inversely related to the level of SHBG. cohort of more than 8000 women with breast cancer,risk varied by However, estradiol is also transportedand bound, though less tightly, stageat diagnosis(Tretli and Magnus, I990). Among women with to albumin.Increases in free fatty acids,such as triglycerides, has been stage I disease,women in the upper quintile of BMI had a 70Vo reported to increasethe level of free estradiol by displacing estradiol increasedrisk of dying from breastcancer. Among women with stage from SHBG where it is tightly bound, to albumin, where it is less II disease,women in the upperquintile had a 407oincreased risk. BMI tightly bound. Therefore, both decreasesin SHBG and increasesin was not associatedwith risk amongwomen with late stageIII and IV triglyceridesmay result in increasesin free estradiol. (1988a, disease(Tretli and Magnus, 1990).In a subsetof 1238 women from Key and Pike 1988b) first proposedthat the effect of this cohort who had unilateralbreast cancer treatedwith modified adiposity on the bioavailability of estrogen was modulated by radicalmastectomy and were followed for l5 years,the risk of dying menopausalchanges in estrogenand progesteroneproduction, and so from breastcancer relative to BMI varied markedlyby hormonerecep- explained the apparentcontradictory findings for premenopausaland tor status(Maehle and Tretli, 1996).Although women with estrogen postmenopausalbreast cancer. Before menopause,ovarian estrogen receptorand progesteronereceptor positive tumorshad a 46Voreduced production overwhelmschanges in estrogenmetabolism related to the risk of dying from breast cancer, the risk within hormone receptor overall level of adiposity.Consequently, estradiol in ovulatorycycles positive and negativegroups varied by BMI. Among women with does not differ measurablyin obesecompared to lean women. In con- hormone receptor positive tumors, obese women had a threefold trast, estradiol levels are reducedin anovulatory cycles that are more higher risk of deaththan did thin women. Conversely,among women frequentin obesethan lean premenopausalwomen. Conversely,obese with hormone receptornegative tumors, thin women had a sixfold premenopausalwomen have been found to have markedly reduced higher risk of deaththan did obesewomen, evenafier adjustmentfor progesteronelevels, both due to increasedfrequency of anovulation lymph node status,tumor diameter,and meannuclear area. One study and decreasedprogesterone production in the luteal phase.With the that examined the associationof BMI with distant recurrenceand onsetof menopause,the decreasedrisk associatedwith premenopausal death found a 707oto SOVoincrease in risk of both distant recurrence obesitydeclines over time. In postmenopausalwomen, the overall level and death among women with a BMI of 21.8 or greater(Goodwin of adiposity results in increasesin estrogenicactivity by increasesin et al., 2002a).This study also examined the associationof fasting estrogenproduction from androgens(Kirschner et al., 1981),decreased (Bruning, insulin,IGF, and estradiolto theseoutcomes and found increasedrisk estrogen-proteinbinding 1987) due to decreasesin SHBG (Moore of distant recurrenceand death among women with elevatedinsulin, et al., 1987; Kaye et al., 1991; Bruning et al., 1992b)and (Bruning but not with elevatedIGF- l, IGF-II, and estradiol.A subsequentstudy increasesin triglycerides et al., 1992b).Furthermore, the C-2 in this samesample found that the binding proteinsof IGFBP-l and hydroxy metaboliteof estrogenhas beenproposed to be a lesstumori- IGFBP-3 were inversely associatedwith risk of distantrecurrence, but genic metaboliteof estrogen,and C-2 versusl6-hydroxylation of estra- only IGFBP-I was also inversely associatedwith risk of death diol is reportedto be decreasedin obesewomen (Schneideret al., 1983). (Goodwin et al., 2002b). A well-designedstudy using contemporary,high-quality assaysfor sex Weight gain is reported in the majority of women undergoing steroidsfound that the increasesin estrone,estradiol, free estradiol,and adjuvanttherapy for breastcancer (Heasman et al., 1985; Goodwin albumin-bound estradiol associatedwith increasesin BMI were not et al., 1988,1999;Camoriano et al., 1990;Demark-Wahnefried, 1993). present in premenopausalwomen but were statistically significant (Potischman Weight gain associatedwith treatmentis lowest among women not among postmenopausalwomen et a1.,1996). receiving systemictherapy, intermediate among women receiving hor- monal therapy,and more pronouncedamong women receiving adju- lnsulin,Insulin-like Growth Factors, vant chemotherapyand thosewho undergomenopause after diagnosis and Other Hormones andtreatment. To identify optimalinterventions to preventweight gain Insulin, insulin-like growth factors, and binding proteins (BPs) may during treatment,research has begun to examine whether changesin promote hormone-dependenttumors through direct effects on tumor energy intake and expenditureduring and after treatmentare associ- cells (Foekenset al., 1989; Yee et al., 1989; Turner et al., 1997), and ated with weight gain (Demark-Wahnefriedet al., 1993,2001: De through indirect effects on estrogensand possible interactions with Waard et al., 1993). Although data on the association of post- estrogenat the estrogenreceptor on breastcancer cells (Clayton et al., diagnosis weight gain and prognosis are limited, the largest study 1997; El-Tanani et al., 1997).A comprehensivereview of epidemio- of 646 premenopausalwomen found that women who gained more logical researchby Yu and Rohan (2000) summarizedthe mitogenic than 5.9kg were 1.5 times as likely to relapseand 1.6 times more and antiapoptoticactions of IGFs on various cancer cells, their syn- likely to die than women who gained less weight (Camoriano et al., gergistic effects with other growth factors and steroids,and the asso- 1990). ciations of IGFs and binding proteins with cancer.The researchon 434 PARTlll: THECAUSES OF CANCER breastcancer is briefly reviewed here. Five of sevenstudies that have PROSTATECANCER examined the effect of IGF-I have found an increased risk of pre- menopausal breast cancer associatedwith increased IGF-I (peyrat Summaryof Findings et al., 1993;Bruning et al.,1995;Del Giudiceet al., 1998;Hankinson Extensive data from cohort and case-controlstudies provide convinc- et al., 1998; Toniolo et al., 2000). However, only one of thesestudies ing evidenceof no associationbetween body massindex and prostate (Bruning et al., 1995) found a statistically significant increased cancer. Data on other anthropometric measures,such as abdominal risk. Three studies found no associationbetween IGF-I and post- obesity,muscle mass,or weight gain, are too limited to allow for any menopausalbreast cancer (Bruning et al., 1995; Hankinson et al., definitive conclusion. 1998;Jernstrlean body mass and adiposity. Other anthropo- Russellet al., 1998;Ho et al., 1999). metric measuresof body fat distribution may be more strongly related to prostatecancer risk becausethis canceris androgen-dependent,and Other Mechanisms lean body massis relatedto androgenlevels. Fewer studieshave exam- ined lean body massor abdominaladiposity, thereby timiting any con- with the recentemergence of studiessuggesting that physical activity clusive assessmentsof the associations.Severson et al. (1988) found may reduce risk of breast cancer,a limited number of analyseshave an increasedprostate cancer risk in a Japanesepopulation associated examinedthe possibility of effect modification or interactionbetween with the area of muscle in the arm and not with the fat in the arm. A physical activity and various weight-relatedmeasures (Thune et al., Norwegian cohort study found no increasedrisk with lean body mass 1997)- Although some studies have included physicar activity in a (Lund Nilsen and Vatten, 1999). Similarly, an American study found large list of covariateswithin multivariate moders to adjust for con- no associationwith lean body massor fat massarea as measuredin founding, no analyseshave reported whether effect modification by the arm (Clarke and Whittemore, 2000). physical activity is observedin risk estimatesfor weight-relatedmea- Weight gain throughout the lifetime and increased abdominal sures and breast cancer. Similarly, although diet has been proposed adiposity also are possibleprostate cancer risk factors, although few as a major mediator of the associationof weight or BMI with post- studieshave evaluated these variables. Schuurman et al. (2000)found menopausalbreast cancer risk, no analyseshave been designed to a significantpositive trend for increasingquantiles of BMI at age20, examine this issue.The large error in self-reportingof energy,greater but no increasedrisk for high weight gain from age 20 to time of inter- in heavier compared to lighter individuals (Heitman, 1993, Subar view. Hsing et al. (2000) found that abdominal adiposity,as measured et al., 2003), limits the ability to examine this issue with currently by WHR, increasedprostate cancer risk in Chinesemen. The observed availabledata. increasedrisk with WHR was independentof BMI, socioeconomic F ti t: :: Obesityand BodyComposition 435

Inodence Studies: Cohorl Studies

N{)rnuraelal (1985),Severson etal {1988),USA, n=174

['lll1,er r/. (1989),USA, n=180

t.ttyouet al.(,l99a), USA, n=306

I r. Marchandel al. (19941,USA, n=198

lhune E Lund(1994), Norway, n=220

Molle( et al. (1994), Denmark,n=96

(,rovanflucciel at (1997),USA, n=1,369

(.r{hanelal. (1997),USA, n=71

Anderssonet ai (1997),Sweden, n=2,368

l und Nilsen& Vatten(1999), No^,yay, n=642

t'utnan et al.(2000), usA, n=101

Schuurmanel al (Z0m),Netherlands, n=6,8 1

llabel elal (2000),USA, n=2,0?9

tee elal (2001),954,n=439

Case-Control Studies: Hospital-Based

Ialaminiel a/.(1986), ltaly, n= 166

Yuelal. (1988). USA, n=1,162

Menlineral (1989),USA, n=371

Demark-Wahnefriedetai (1997a).USA, n=l56

Spte et al.(2000), USA, n=103

{Jairarielal (2000).Canada, n=64

Case-Control Studies: Population-Based

Anderssonelal (1995),Sweden, n=256

Anderssonelal (1996),Sweden, n=256

Gronbergel a/.(1996). Sweden, n=406

l\ey et al.{1997). United Kingdom, n=328

Hayeset al.(1999). USA, Blacks. n=449

Ha,leset al.(1999). USA, Whites, n=483

Vrlleneuveet a/ (1999),Canada. n-1,623

Hsing (2000),China, n=239 elal Figure 22-7. Summaryof risk estimatesand confi- Sharpeand Siemiarycki (2001). Canada, n=399 denceintervals from epidemiologicstudies of BMI and prostatecancer.

status,physical activity, and dietary intake, including total calories. BiologicalMechanisms Though the associationswith WHR were strong and consistent in this study, additional empirical evidence is needed to confirm this Prostate cancer is a hormone-mediatedcancer in that endogenous association. hormones regulate the growth and function of the prostate gland Someeffect modificationby stageof diseasemay exist for obesity (Henderson et al., 1982), and administration of large quantities of and prostate cancer. In the Health ProfessionalsFollow-up Study, testosteronecan induce prostatecancer in rodents(Noble, 1977).Men Giovannucciet al. (1991) observeda 58Voincreased risk among men who have high estrogenlevels rarely develop prostatecancer (Glantz, with metastaticprostate cancer for those in the highest quintile of 1964). A Western lifestyle, charactenzedby a positive energy balance WHR, though no such increased risk was found for all stages of as a result of high-energy intake and low activity levels, has consis- prostate cancer combined. In this same study, preadult obesity tently been implicated as a risk factor for prostate cancer in inter- appearedto protect against the occurrence of prostate cancer, and national comparisonsand experimental studies (Hebert et al., 1998; these associationswere particularly evident for metastaticdisease Boslandet al., 1999;Mukherjee et al.,1999).The evidence,however, (Giovannucci et al., 1991). Further effect modification may occur from individual-level epidemiologic studies on the associations with race. In a study in the United Stateswith African American and between dietary intake, obesity, and physical activity and prostate white men, Hayes et al. (1999) found an increasedrisk of advanced cancerrisk is weaker (Friedenreichand Thune, 2001; Schulmanet al., prostatecancer among black men in the highestquartile of BMI at age 2001). Despite the lack of an associationbetween BMI and pro- 25. No similarly increasedrisks were found for white men or for men statecancer from epidemiologic studies,several studies of biological with less advancedprostate cancer for either race.Additional factors markers that are related to obesity suggest an associationof these that may modify the effect of this putative associationhave not been markers with prostate cancer. fully explored. One such factor is smoking. A case-controlstudy in Leptin, an adipocyte-derivedhormone that regulates satiety and Canada found an interaction between smoking and BMI. Men with energy expenditure,has been shown to increaseprostate cancer risk. high BMI who were smokers were at increasedrisk as compared to It is suggestedas a key link between a state of continual positive nonsmokerswithin the same BMI category (Sharpeand Siemiatycki, energy balance and the transition from premalignant lesions to overt 2001). prostatecancer (Stattin et al., 2001). In addition, higher serum insulin 436 PARTlll: THECAUSES OF CANCER

levels increaseprostate cancer risk independentof abdominal adipos- mones. Proposed mechanisms include increases in obesity-related ity (Hsing et al., 2001). Despite these associations,it is unclear postmenopausalestrogen based on evidencethat exogenousestrogens whether the concomitant hormonal and metabolic changes that occur are weakly associatedwith thyroid cancerrisk (La Vecchiaet al., 1999; becauseof abdominal adiposity and insulin resistanceare the biolog- Negri et al., 1999).In addition, detection bias due to more frequent ical mechanisms for the link between insulin and prostate cancer risk. examination of the thyroid gland becauseof screeningfor hypothy- Two other biomarkers for prostate cancer risk that may be related roidism among heavier women may be a factor in countries,such as to BMI and either fat or musclemass have beenidentified from studies the United States,where this practice may be more prevalent. in diverse populations: IGF-I (Mantzoros et al., 1997; Chan et al., 1998; Wolk et al., 1998) and serum testosterone(Shaneyfelt et al., 2000). A meta-analysisof publishedstudies on hormonal predictorsof prostatecancer risk found that men with serum testosteroneor IGF- I CANCERSOF THE LUNG AND THEHEAD AND NECK levels in the upper quartile of the population disrribution have an approximately twofold higher prostatecancer risk (Shaneyfeltet al., Overview 2000). Levels of dihydrotestosteroneand estradiol were not found to An extensiveliterature exists on the commonly observedinverse asso- have as strong an associationwith prostate cancer risk (Shaneyfelt ciation betweenobesity and cancersof the lung and the head and neck. et al., 2000). Overall, exposureto endogenousandrogenic hormones This literature was summarizedin the 2002International Agency for appearsto be positively associatedwith prostatecancer risk (Bosland, Researchon Cancer (IARC) report on weight control and physical 2000). A number of studies have explored the associationsbetween activity (IARC, 2002). In the past 10 years, studies examining the various sex steroidsand BMI in men (Glass, 1989; Kato et al., I99Z: inverse associationof obesity and these cancershave been designed Andersson et al., l99l;Demark-Wahnefried et al., 1997b Tymchuk to examine the extent of confounding by the effect of tobacco use or et al., 1998; Tamani et al., 200I; Sarma et al., 2002). However, how preexistingdisease that may result in weight loss. BMI may play a role in the associationof thesemeasures with prostate cancer is complex and not well understood. Summaryof Findings THYROIDCANCER Data from case-controland cohort studies suggestan inverse associ- ation betweenBMI Summaryof Findings and cancersof the lung and headand neck.This inverse associationmoves to the null after adjustmentfor confound- Data from case-control and cohort studies are suggestive of an ing by tobaccouse and,therefore, is not thought to be causalin nature. increasedrisk of thyroid canceramong heavier individuals (BMI cut- points not defined) that is presentonly for women. Estimatesfrom a meta-analysissuggest a 20Voincrease in risk in women (Dal Maso Lung Cancer et al., 2000). The majority of cohort (Knekt et al., 1991; Lee and Paffenbarger, 1992b: Chyou et al., 1994; Drinkard et al., 1995; Kark et al., 1995. Epidemiology Henley et al., 2002) and case-control studies (Kabat and Wynder, 1992;Goodman and Wilkins, 1993;Kubik et al., 2N2: Olson et al., A meta-analysis(Dal Maso et al., 2000) summarizedand reanalzyed 2002) have found an inverse associationbetween body size and lung evidencefrom the existing case-control studieson obesity and thyroid cancer,although it is generallyless striking after adjustmentfor poten- cancer(McTiernan et al., 1984;Preston-Martin et al., 1987;Ron et al., tial confounding by smoking status.In addition, severalrecent studies 1987;Linoset al., 1989;Kolonel et al., 1990;Francheschi et al., l99l; have examined the associationof BMI and lung cancer in subgroups Levi et al., 1991; Goodman et 1992; preston- al., Glattre et al., 1993; such as currentand nonsmokers(Rauscher et al., 2000; Henley et al., Martin et al., 1993; Wingren er al., 1993: Hallquist er al., 1994, 2002) and found little associationamong never or former smokers.A D'Avanzo et al., 1995;Galanri er al., 1997;Negri er al., 1999).The study limited to never and former smokersin New York (Rauscheret authors reportedthat relative risks for the upper tertile of BMI at the al., 2000) found a doublingof risk for lung canceramong obese men time of diagnosiswas above unity for 9 of the 12 studies in women, and women that was not statisticallysignificant. A cohort study from with a relative risk overall for the highest compared to the lowest the American Cancer Society's Cancer Prevention Study II, which tertile of 1.2 (95VoCl, 1.0-1.4), P for trendof 0.04.The overall rela- examinedthe effect of the potential for bias from tobaccouse and pre- tive risk for the highestcompared to the lowest tertile in men was r.0, existing disease,also evaluated the role of weight and BMI among with a nonsignificantP for trend. never smokers and found no association(Henley et al., 2002). This Similarly, in five case-control studiesthat haveexamined this factor, study showed a lesseningin the observed associationbetween BMI no associationhas been observed for thyroid cancer risk and BMI and lung cancer among the entire cohort after adjusting for preexist- betweenthe agesof 17 to 20 (McTiernan al., presron-Martin et 1984; ing diseaseand removing early follow-up as an additional method of et al., 1987;Linos et al., 1989; Preston-Martinet al., 1993;Wingren adjusting for preexisting disease.These findings are interpreted to et al., 1993). One case-control papillary study of thyroid cancer in indicate that leannessis unlikely to be a causalfactor in the develop- women (Rossing et al., 2000), published after the Dal Maso (2000) ment of lung cancer.However, it is likely that the causal factors for meta-analysis,found that women who weighed 185 pounds or more lung cancer among never smokers are very different than causal had a l.5-fold increasedrisk. Another case-controlstudy in women factors among current smokers.Therefore, it is not clear that analyses found no associationbetween BMI and (Mack thyroid cancer et al., of associationsbetween BMI and lung cancer among never smokers 2002) but did observe a non-statistically significant increasein risk can contribute much to understanding whether confounding by with weight gain. The only cohort study to examine the association tobacco use explains the majority of the inverse associationobserved between obesity and thyroid cancerrisk found no statistically signifi- betweenBMI and lung cancer among smokers. cant associationwith BMI at the baseline interview or with weisht The only study that has examined the associationof body fat dis- gain since age20 (Iribanen er al., 2001). tribution by waist circumferenceand lung cancerrisk found evidence of a differential associationbetween waist circumferenceand histo- BiologicalMechanisms logic subtypesof lung cancer(Olson et a1.,2002).In thatstudy among middle-agedlowa women, risk for small cell and squamouscell lung several mechanismshave been proposedfor the modest association cancer was increasedby threefold among women in the highest quin- observedin many case-controlstudies in women. Most relateto poten- tile of waist circumference;in contrast,no associationwas observed tial interactions between thyroid hormones and other steroid hor- betweenwaist circumferenceand adenocarcinomaof the lune. Obesityand BodyComposition 437 Headand NeckCancers to be at very low or high risk. In addition, more studieson weight and height at an early agein life with a long follow-up are neededto clarify Although tobacco and alcohol use account for more than 90Voof the role of weight and weight gain throughoutthe life span.For some cancers the head in (IARC, of and neck developingcountries 1986, cancer sites, a number of studiescorrected only for age, thus allow- 1988),leanness also has been associatedwith increasedrisk of head ing uncontrolled confounding to influence the results.Important con- in and neck cancers some case-controlstudies. No cohort studies founders to consider include physical activity, other anthropometric published. examining this associationhave been Six case-control factors such as height, smoking status, dietary intake, and other studies found an inverse associationbetween weight and/or BMI lifestyle factors.A full examinationof all anthropometricrisk factors (Mclaughlin and cancerof the oral cavity and pharynx et al., 1988; and canceris also needed.For somecancer sites, such as breastcancer, Marshall et al., 1992;;Day et al., 1993;Kabat et al., 1994,D' Avanzo additional measuressuch as body fat distribution, weight gain, and Francheschi et al., 1996; et al., 2001)and one founda non-statistically body composition have been examined. However, for most cancer significantrisk estimateof 1.5 among heavierwomen (Negri et al., T: very factor lr. sites, few studieshave consideredany other than BMI. 2000).Two other studiesfound weakerinverse associations between Moreover, studiesthat examine a full range of possibleeffect mod- ':. (Muscat BMI and laryngeal cancer and Wynder, 1992: D'Avanzo ifiers need to be conductedbecause relatively few investigationshave ::: et al., 1996).Similar to findingsfor lung cancer,several studies found t.: examined risks within subgroupsof the population defined on char- (Kabat eitherno association et al., 19941'Francheschiet aI.,2001) or acteristicssuch as ethnic origin/race,other lifestyle risk factors (e.g., (D'Avanzo weaker associations et al., 1996) betweenBMI and oral physical activity, dietary intake, smoking status),HRT, geneticfactors cancer among never smokers compared to current smokers. These (e.g., family history of cancer or other comorbid diseases,such as resultssuggest that BMI is not causallyrelated to risk of head and diabetesthat may influence the health consequencesof obesity), or neck cancers.However, similar to lung cancer,the causalfactors con- other possiblerisk factors, such as breastdensity in the caseof breast tributing to head and neck cancer among never smokersmay be dif- cancer.Studies that provide data specific to racial and ethnic groups ferent than those amons current or former smokers. are increasinglyimportant. At present,most of the data on obesity and cancer are based on population samples drawn from Europe, the United States, Canada, Australia, Japan, and China, and therefore CANCERSITES WITH INSUFFICIENT EVIDENCE havelargely includedwhites of Europeandescent, Chinese, Japanese, FORCONCLUSIONS and Hawaiians. Data are very limited for African-American men and women or those of North. Central. or SouthAmerican Latina descent Limited epidemiologicevidence exists on weight or BMI for a num- or for many populationsin Asia and Africa. In addition, data on these ber of other cancers, including leukemia, non-Hodgkin lymphoma, groups are seldom reported separatelyeven within large case-control malignant melanoma,and testicular,pancreatic, bladder, and cervical or cohort studiesthat may include thesepopulations. cancers. However, data for these and other cancer sites are too To clarify the extent to which anthropometric factors are risk factors limited to allow specific conclusions.tobe made and are not summa- for specific cancers,further researchis neededto delineate the role rized here due to space limitations. A summary of the evidence for of site-specific biological mechanisms,particularly for endogenous several of these other cancer sites is provided in the February hormones, insulin, insulin-related growth factors, leptin, cortisol, 2002 IARC publication Weight Control and PhvsicalActivirv (IARC, and relatedhormones that may influencethe developmentof body fat, 200D. skeletal structure, and musculature and that may underlie observed associationswith canceroutcomes. Evidence is evolvingrapidly about geneticfactors that may have particular importancein the metabolism DATAON CANCERMORTALIry of sex steroids and insulin or in the developmentof obesity. Future studies are neededto examine how specific genespredict underlying Becauseof space limitations and the difficulty of disentangling the metabolic profiles and subsequentcancer risk. effect of screeningand treatment from individual-basedrisk factors, An overview of the potential biological mechanismsthat have been such as obesity,when examining cancermortality as an outcome, this either hypothesizedor examinedas possibleexplanations for the asso- review doesnot provide a detailedoverview of the evidenceon obesity ciation of obesity with specificcancers is presentedin Table 22-l.Tltis and cancermortality. However, a recentstudy from a large U.S.-based overview demonstratesthat many of these cancersare hypothesized cohort provides severalnew insights (Calle et al., 2003). In this study, to be hormone based,either in terms of sex steroidsor, more recently, the effect of obesity on cancermortality was examinedin a cohort of in terms of insulin, leptin, or insulin-relatedgrowth factors.Although more than 900,000 adults who developedmore than 57,000 cancersin not explored to any extent in humans, research in animal models 16 yearsof follow-up. The study is unique in that it provided estimates suggestthat factors such as cortisol or vitamin D may also explain for multiple cancerswithin one cohort and one report and had a much some of the risk associated with obesity. Using animal models, larger sample size than any other single report to date. Therefore, these mechanismshave largely been explored by studies of caloric the study was able to provide estimatesfor multiple cancers,including restriction rather than by studies of increasesin energy expenditure more precise estimatesfor less common cancers,and examined the (Hurstinget al., 2003). effect of overweight and obesity among nonsmokers.The study found Two other areasof emerging researchshould be briefly noted. With that BMI was associatedwith higher rates of death in both men improved detection and treatment, many people are living longer and women for cancers of the esophagus,colon and rectum, liver, and healthier lives with cancer,and therefore,studies on the effect of gallbladder,pancreas, kidney, non-Hodgkin lymphoma, and multiple obesity on cancerprognosis are needed.At present,research in prog- myeloma. BMI was also associatedwith higher rates of death from nosis has been largely limited to breastcancer and should be extended cancersof the stomachand prostatein men and cancersof the breast, to other cancers.In addition, to achievea better understandingof what uterus,cervix, and ovary in women. Basedon ratesof overweight and interventionsmay reduce risk, studiesof the effect of specific dietary obesity present in the United Statesin 1999 to 2000, the study esti- and exercise programs to control weight are needed, particularly mated that l47o of cancerdeaths in men and 20Voof cancerdeaths in studiesthat can elucidatethe effect of such interventionson underly- women could be attributed to overweieht and obesitv. ing mechanismsthat may influence carcinogenesis.

FUTUREDIRECTIONS EXTENTOF THEOBESITY EPIDEMIC AND MANAGEMENTSTRATEGIES Studies are needed that have sufficient power and complete data on a numberof factorsto adequatelycontrol for all confoundersof the asso- The WHO 2000 report Obesity: Preventing and Managing the Global ciation between BMI and cancer,and to study subgroupspostulated Epidemic is the most recent comprehensive source summarizing the 438 PARTlll: THECAUSES OF CANCER

Table22-1. Biological Mechanisms Potentially Involved in the Association Between Obesity and Specific Cancers

Mechanism Explanation Cancer Sites

Altered endogenousestrogen levels Higher estrogen,higher testosterone.and lower SHBG Breast,endometrial, renal Altered endogenousandrogen levels Higher estrogen,lower testosterone,and lower SHBG Prostate Altered metabolic hormone levels Elevated leptin and insulin, increasedIGF, higher Prostate,breast, endometrium, colon, cholesterol esophageal,gastric Interaction between sex steroid and metabolic Prostate,breast, endometrium, colon hormones Altered nervous and immune systemfunctioning Greater sympatheticactivity, immune system dysfunction Prostate,breast Increasedcell number Larger pool of cells to undergo malignant transformation Breast,colon, theoreticallyrelevant to all Positive energy balance High dietary fat intake, Iower physical activity All Concentrationof adversefactors in adiposetissue Greater concentrationof growth factors or carcinogens All Increased gastric reflux Increasedintra-abdominal fat mass increasesintra- Esophageal abdominal pressureand hence increasesgastric reflux Delayed esophagealtransit time Obesity leads to delayed esophagealtransit time and Esophageal increasedexposure time Increasedincidence of metabolic syndrome Contributes to renal damageand susceptibility to Renal (hypertension,glucose intolerance,increased other carcinogens IGF, and in women anovulation and increased androgens)

IGF, insulin-like growth factor; SHBG, sex hormone binding globulin.

global epidemic of obesity. Within that report, the WHO MONICA demia, coronary heart disease,hypertension, osteoarthritis, sleep study provides a full spectrum of data comparing rates of obesity apnea,low-back pain, polycystic ovarian disease,impaired fertility, for period worldwide the 1983-1986.At that time, the international reproductive hormone abnormalities,surgical complications, and '&* prevalence of obesity, defined as a BMI of 30 and higher, ranged anesthesiacomplications. [nternational estimates of the economiccost .:*t-: from 5Voto 20Vofor men and l}Vo to 407o in women (WHO, 2000). have rangedfrom 2Voto about 77oof nationalhealth care costs, with The WHO 2000 report highlighted sgveral major features of the the United Statesreporting the highestcost of 7Vo(WHO, 2000). In :=.:,il internationalpatterns of body weight and associatedhealth outcomes. 1990,this costwas estimated to be $45.8billion dollarsfrom thedirect 4 For the first time, more peoplewere classifiedas experiencingobesity cost of obesity-associateddiseases and $23.3 billion from lost pro- ,s than were suffering from starvation in developing as well as devel- ductivity (Wolf and Colditz, 1994).In an updatedanalysis, Wolf and '.-i: oped countries. The report also noted the rapidly increasing rates Colditz (1998) estimatedthe total U.S. direct cost attributableto r'*; z of obesityin someAsian countries,such as Japanand China, where obesityto be $99.2 billion, which represented5.77o of the total U.S. .'t the prevalenceof obesity traditionally had been very low. Similar to national health expenditurethat year. .tr: '.&:.;t patternsobserved in the United States,rates of obesity in other coun- A comprehensiveoverview of the treatment and prevention of a€ tries are generally higher among women than men and in urban as obesityis beyondthe scopeof this review.These issues are addressed ',8 compared to rural communities.Although data on children are more in detail in the WHO 2000 report Obesity: Preventingand Managing ,:z; (Kopelman, .i.*, limited, evidencesuggests that obesity is also increasingin children the Global Epidenticand other reviews 2000).Advances tr€ worldwide in developed and developing countries. In 2000, the in the field of obesity,particularly in termsof the geneticsof obesity .4:,{ :::d WHO initiateda global strategyfor preventingand controlling non- and its implication for developing more targeted and effective :t communicablediseases that includes a focuson combattingthe world- treatment approaches,have been highlighted in issues of Science ' "$' wide rise in obesity through reducing unhealthy diets and physical (February 2003) andNature (April 2000). Preventingand controlling .,{t' inactivity. obesity requires comprehensivesocietal efforts and must target life- '':,b In the United States,the prevalenceof obesity has changed dra- long health habits at the individual level. In addition, continued :.li matically over the past 40 years. It was relatively stable at approxi- researchis neededto find better meansof identifying individualsat ,a .fr, mately L}Vo for men and l5%ofor women during the early 1960sto risk of developingobesity and its adverseconsequences because of ;l the late 1970s.During the late 1980sand early t990s ratesof obesity their genetic,familial, or other environmentalrisk factors.At present, :* t# increased, and the most current estimates from the 1999-2000 with the rapid increasein obesityin developedand developingcoun- ,-E: lai National Health and Nutrition Examination Survey (NHANES) indi- tries amongall agegroups, including children, prevention has become ,'{ cate that the prevalenceof obesity has increased to 27.5Vo for men and an urgent worldwide priority. lssuesof particularconcern in cancer ' !.s.:.,. ri 33.4Vofor women (Flegal eta1.,2002).Rates tend to increasewith age control include the need to develop better strategiesto facilitate the ,9. to about age70; at later ages,they declineslightly. Ratesare highest avoidanceof weight gain during adult life and to identify whether amongnon-Hispanic black women who experiencea 507oprevalence specific interventionsduring treatmentcan alleviatetreatment-related .$ of obesity. weight gain that has been identified to worsen breastcancer prog- iai When the prevalenceof overweight, or a BMI of 25 and higher, is nosis,but may also be an issuefor other cancers.Key public health ,,:E considered,about 65Voof the U.S. adult population is affected.Of par- recommendationsand strategiesrelevant to cancer are reviewed in '':::i.i:? i:ji: ticular concernare increasesin ratesof overweight among children the2002IARC reportonWeight Control cmd Ph,-sicalActiv,in (IARC, .tiF ,tj: and adolescents.Prevalence rates, which were approximately 57o 2002). :H during the 1960s, have tripled to over I5Vo in 1999-2000 among school-agedchildren and adolescents.Rates rose by 10 percentage JF points or more between 1988 and 1994 and between 1999 and 2000 CONCLUSIONSAND POPULATION i;9. for both Mexican American and non-Hispanic black adolescents v;; ATTRIBUTABLERISK '':',9€,,t;r:r: (Ogden et al., 2002). ::j:i Estimatingthe consequencesand cost of obesity internationallyis At present,data provide convincing evidence of a positiveassociation .:4:i:,i complexbecause obesity and overweightare risk factorsfor multiple of overweight and obesity with cancersof the colon (among men). .;5r .€a chronic diseasesother than cancer,including non-insulin-dependent renal cell, postmenopausalbreast, endometrium, and probable evi- diabetesmellitus, insulin resistance,gallbladder disease, hyperlipi- denceof a oositiveassociation with colon cancer(amons women). E ;.,4

':!l9& at:;: &ieE Obesityand BodyComposition 439 Table22-2. Populati on Attributable Rates(PAR in vc)f or weight-Related 7o cancersfor the EuropeanUnion (EU) from 1982to 1996and the united 8 Statesfrom 1999to 2000 r60 ,9 Men Women tr, o50 -----ry--l o Overweight Obese Overweight Obese o a E40 t"J .o EU* Colon 6.9 -.a1 5.0 5.'7 iJU Renal 15.2 r0.3 il.t 13.4 c Endometrium 17.2 22.0 .9 zo Breast (postmenopausal) 4.1 4.5 o f-', ..,l i 7G USA d. 10 o L Colon 9.1 8.5 8.5 9.8 0. 13 Kidney 19.4 19.0 18.3 21.7 Endometrium 26.8 33.4 10 20 30 40 50 60 70 Breast (postmenopausal) 6.9 7.6 Overweight/Obesity Prevalence (%) overweight/obesity prevalence rates used: EU overweight and obesity rates were, respec- tively,35vo and l9va for men, and 50vo and l3vo for women. For the United Statesfrom Figure 22-8. Populationattributable risk (PAR) by overweight/obesity 1999 to 2000, the rates were increased to 6'lvc and 28vo for men and 62va and 33va fot women. prevalencerates and levels of relative isk. Source:Population Attribut- *Source of RR and EU PARs is Bergstr

adenocarcinomaof the esophagus and gastric cardia, and thyroid cancer(among women). on the other hand, epidemiologic studiesdo not demonstratean associationof weight or BMI with prostatecancer incidence.In addition,limited evidencedoes not suggestany consis- relatedcancers is further demonstratedin Figure 22-9, which demon- tent direct association between overweight or obesity and ovarian stratesthe relatively strong associationbetween rates of obesity and cancer. The consistent inverse association between overweight or endometrial cancer prevalence internationally. Similar associations obesity and lung, and head and neck, cancersis reducedwith adjust- can be seen for other cancers as well. With the expectationthat the ment for tobacco use; therefore,overweight and obesity are not con- epidemic of obesity is likely to continue, if not accelerate,in the near sidered to be protective for these cancers.The inverse association term, overweight and obesity will become increasingly important between overweight and obesity and premenopausalbreast cancer contributorsto cancer risk internationally. persists after adjustment for multiple potential confounding factors. Acknowledgments Extensive data provide convincing evidenceof a positive association betweenoverweight and obesity and adversechanges in breastcancer The authorswould like to acknowledgethe contributionsof the following prognosisand mortality. peopleto thisreview: Anita Ambs, National Cancer Institute, for managingthe distributionof materialsfor thereview and the Estimatesof the population attributable risk (PAR) of cancer due developmentof figures;David Tran,Scientific Consulting Group, for literaturereviews; Darrell Anderson, to overweight and obesity have been summarizedin the 2002 IARC ScientificConsulting Group, for developmentof summarytables, figures, and report and are modest for some cancers, such as colon and post- final preparationof the manuscript;and Anne Rodgers,for editing. menopausalbreast cancer. Approximately 9voto llvo of thesecancers are attributableto overweight or obesity.PAR estimatesnre more sub- stantialfor renalcell, esophageal,and endometrialcancer:25vo,37vo, and 39vo,respectively. However, theseestimates were basedon inter- national ratesof overweight and obesity from the 1990s(IARC, z00z) and are higher in the United Statestoday given the continuedincrease in the prevalenceof overweight and obesity. Table 22-2 provides a comparison of PARs for five major weight-related cancers for the 20 Czech Republico EuropeanUnion as estimaredby Bergstrcimand colleagues(2000) and for the united states basedon rates of obesity and overweight from s .i a ltaly the 1999-2000 NHANES survey. The t PARs estimated by Bergstrtim sweden. ?1s oc.nuoYsA' and colleagues(2001) were basedon relative risk estimatesfrom meta- o o Finland o yugosravia analysesand from rates of obesity in Europeanunion countriesfrom o cobmbia. f:1 ctl *tz]*"rr. . stuilrrno 1982 to 1996.Because (! TheNerherrandsl of increasesin ratesof overweight and obesity I '10 in the United Srates,PARs are higher for rhe united Statesthan those o IE uK,EngtandaFfance' estimatedfor the European union. Figure 22-8 shows the potential tr, for c, substantial increasesin PARs with increasesin rates of over- o a Zimbabwe tr weight/obesityand relative risk estimatesof overweight/obesitywith o5 JaPan' cancer.For example, p o Peru in a country such as Japan,which has obesity o O SaudiArabia rates of less than 3vo, the PAR for a cancer with a twofold increase tr Chinao due to obesity, such as endometrial cancer, is less than a Mali 3vo.T\e PAR estimatefor obesity and endometrialcancer within the European Union published by Bergstr0m and colleagues(2000) was 0 10 20 30 40 50 22vo when the prevalenceof obesity was l3vo for women within the Obesity Prevalence,7o European Union. In contrast, in the United States, whose rate of obesityin women is 33vo,the PAR for endometrialcancer is 33vo.The Figure 22-9. Associationbetween obesity prevalence (BMI > 30kg/m2) potential for marked increasesin the PAR for obesity and weieht- and endometrialcancer cumulative ratesin selectedcountries. MO PARTlll: THECAUSES OF C,ANCER

References Boutron-Ruault MC, SenesseP, Meance S, Belghiti C, Faivre J. 2001. Energy intake, body mass index, physical activity, and the colorectal adenoma- Agurs-Collins T, Adams-Campbell LL, Kim KS, Cullen zu. 2000. Insulin-like carcinoma sequence.Nutr Cancer 39:50-57. growth factor- l and breast cancer risk in postrnenopausal African- Brinton LA, Berman ML, Mortel R, Twiggs LB, Barrett RJ, Wilbanks GD, American women. Cancer Detect Prev 24:199-206. Lannom L, Hoover RN. 1992.Reproductive, menstrual, and medical risk Akhmedkhanov A, Zeleniuch-Jacquotte A, Toniolo P. 2001 . Role of exogenous factors for endometrial cancer: Results from a case-control study. Am J and endogenoushormones in endometrial cancer: Review of the evidence Obstet Gynecol 167:1317 -1325. and researchperspectives. Ann N Y Acad Sci 943:296-315. Brinton LA, Swanson CA. 1992. Height and weight at various ages and risk Anderson KE, AndersonE, Mink PJ,Hong CP, Kushi LH, SellersTA, Lazovich of breastcancer. Ann Epidemiol2:597409. D. Folsom AR. 2001. Diabetes and endometrial cancer in the Iowa Brown LM, Swanson CA, Gridley G, Swanson GM, Schoenberg JB, women's health study. Cancer Epidemiol Biomarkers Prev 10:6ll-616. Greenberg RS, Silverman DT, Pottern LM, Hayes RB, Schwartz AG. AnderssonSO, Baron J, Wolk A, Lindgren C, BergstrcimR, Adami HO. 1995. 1995.Adenocarcinoma of the esophagus:Role of obesity and diet. J Natl Early life risk factors for prostate cancer: A population-based case-control Cancer lnst 87:104-109. study in Sweden.Cancer Epidemiol Biomarkers Prev 4:187-192. Bruning PF. 1987. Endogenousestrogens and breast cancer a possible rela- AnderssonSO, Baron J, BergstrcimR, Lindgren C, Wolk A, Adami HO. 1996. tionship betweenbody fat distribution and estrogenavailability. J Steroid Lifestyle factorsand prostatecancer risk: A case-controlstudy in Sweden. Biochem 27:487492. Cancer Epidemiol Biomarkers Prev 5:509-513. Bruning PF, Bonfrer JM, van Noord PA, Hart AA, Jong-Bakker M, Nooijen Andersson SO, Wolk A, Bergstrtim R, Adami HO, Engholm G, Englund A, WI. 1992a. Insulin resistance and breast-cancerrisk. Int J Cancer Nyren O.1997. Body size and prostatecancer: A1}-year follow-up study 52:511-516. among 135006 Swedish construction workers. J Natl Cancer Inst Bruning PF, Bonfrer JM, Hart AA, van Noord PA, van der Hoeven H, Collette 89:385-389. HJ, Battermann JJ, Jong-Bakker M, Nooijen WJ, de Waard F. 1992b. Asal NR, Risser DR, Kadamani S, Geyer JR, Lee El Cherng N. 1988. Risk Body measurements,estrogen availability and the risk of human breast factors in renal cell carcinoma: I. Methodology, demographics,tobacco, cancer:A case-controlstudy. Int J Cancer51:14-19. beverageuse, and obesity.Cancer Detect Prev ll:359-377. Bruning PF, Van Doorn J, Bonfrer JM, van Noord PA, Korse CM, Linders TC, Austin H, Austin JM, Jr, Parridge EE, Hatch KD, Shingleton HM. 1991.Endome- Hart AA. 1995.Insulin-like growth-factor-bindingprotein 3 is decreased trial cancer,obesity, and body fat distibution. Cancer Res 5l:568-572. in early-stage operable pre-menopausal breast cancer. Int J Cancer Bairati I, Larouche R, Meyer F, Moore L, Fradet Y. 2000. Lifetime occupa- 62:266-270. tional physical activity and incidental prostate cancer. Canada. Cancer Byers T, Marshall J, Graham S, Mettlin C, Swanson M. 1983.A case-control CausesControl ll:7 59-7 64. study of dietary and nondietary factors in ovarian cancer. J Natl Cancer Ballard-Barbash R, Schatzkin A, Carter CL, Kannel WB, Kreger BE, Inst 7l:681-686. D'Agostino RB, Splansky GL, Anderson KM, Helsel WE. 1990a.Body Caan BJ, CoatesAO, SlatteryML, PotterJD, QuesenberryCP Jr, EdwardsSM. fat distribution and breast cancer in the Framingham Study. J Natl Cancer 1998.Body size and the risk of colon cancerin a large case-controlstudy. Inst 82:286-290. Int J Obes Relat Metab Disord 22:178-184. Ballard-BarbashR, SchatzkinA, Taylor PR, Kahle LL. 1990b.Association of Calle EE, Rodriguez C, Walkr-Thurmond K, Thune MJ. 2003. Overweight, change in body mass with breastcancer. Cancer Res 50:2152-2155. obesity and mortality from cancer in a prospectively studied cohort of Ballard-Barbash R. 1999. Energy balance, anthropometrics and cancer. In: U.S. adults.N Engl J Med 348:1625-1638. Heber D, Blackburn GL, Go VLW, Holland JL, eds. Nutritional Oncol- Camoriano JK, Loprinzi CL, Ingle JN, Therneau TM, Krook JE, Veeder MH. ogy. San Diego: Academic Press,pp. 137-153. 1990.Weight changein women treatedwith adjuvanttherapy or observed Barnes-JosiahD, Potter JD, SellersTA, Himes JH. 1995. Early body size and following mastectomy for node-positive breast cancer. J Clin Oncol subsequentweight gain as predictors of breast cancer incidence. Iowa, 8:1327-1334. United States.Cancer Causes Control 6:112-118. CasagrandeJT, Louie EW Pike MC, Roy S, Ross RK, HendersonBE. 1979. BastarracheaJ, Hortobagyi GN, Smith TL, Kau SW, Buzdar AU. 1994.Obesity "lncessantovulation" and ovarian cancer.Lancet 2:170-173. as an adverseprognostic factor for patients receiving adjuvant chemother- CASH STUDY. 1987.The reduction in risk of ovarian cancer associatedwith apy for breastcancer. Ann Intern Med 120:18-25. oral-contraceptive use. The Cancer and Steroid Hormone Study of the BaumgartnerRN, Heymsfield SB, RocheAF. 1995. Human body composition Centers for DiseaseControl and the National Institute of Child Health and the epidemiology of chronic disease.Obes Res 3:73-95. and Human Development.N Engl J Med 316:650-655. Beard CM, Hartmann LC, Keeney GL, Crowson CS, Malkasian GD, O'Brien Cerhan JR, Torner JC, Lynch CF, Rubenstein LM, Lemke JH, Cohen MB, PC, Melton LJ, III. 2000. Endometrial cancer in Olmsted County, MN: Lubaroff DM, Wallace RB. 1997. Association of smoking, body mass, Trendsin incidence,risk factors and survival. Ann Epidemiol l0:97-105. and physical activity with risk of prostatecancer in the Iowa 65+ Rural Benhamou S, Lenfant MH, Ory-Paoletti C, Flamant R. 1993. Risk factors for Health Sfudy. United States.Cancer CausesControl 8:229-238. renal-cell carcinoma in a French case-control study. Int J Cancer Chan JM, Stampfer MJ, Giovannucci E, Gann PH, Ma J, Wilkinson P, 55:32-36. HennekensCH, Pollak M. 1998. Plasmainsulin-like growth factor-l and Bennett FI, McFarlane-AndersonN, Wilks R, Luke A, Cooper RS, Forrester prostatecancer risk: A prospectivestudy. Science279:563-566. TE. 1997. Leptin concentrationin women is influenced by regional dis- Cheng KK, Sharp L, McKinney PA, Logan RF, Chilvers CE, Cook-Mozaffari tribution of adiposetissue. Am J Clin Nutr 66:1340-1344. P, Ahmed A, Day NE. 2000. A case-control study of oesophageal BergstrcimA, Pisani P, Tenet V Wolk A, Adami HO. 2001. Overweight as an adenocarcinomain women: A preventabledisease. Br J Cancer 83:127- avoidable causeof cancer in Europe. Int J Cancer 9l:421430. 132. Bernstein L, Ross RK. 1993. Endogenoushormones and breast cancer risk. Chie WC, Li CY, Huang CS, Chang KJ, Lin RS. 1998. Body size as a factor Epidemiol Rev l5:48-65. in different ages and breast cancer risk in Taiwan. Anticancer Res Berstein LM. 1988. Newborn macrosomy and cancer. Adv Cancer Res 50: 18:565-570. 23r-278. Choi NW, Howe GR, Miller AB, Matthews V, Morgan RW, Munan L, Burch Blitzer PH, Blitzer EC, Rimm AA. 1976.Association betweenteen-age obesity JD, FeatherJ, Jain M, Kelly A. 1978.An epidemiologic study of breast and cancerin 56,11l women: all cancersand endometrialcarcinoma. Prev cancer.Am J Epidemiol 107:510-521. Med 5:20-31. Chow WH, Mclaughlin JK, Mandel JS, Wacholder S, Niwa S, Fraumeni Boeing H, SchlehoferB, Wahrendorf J. 1997. Diet, obesity and risk for renal JF Jr. 1996. Obesity and risk of renal cell cancer. Cancer Epidemiol cell carcinoma: results from a case control-study in Germany. Z BiomarkersPrev 5:l7 -21. Ernahrungswiss36:3-1 l. Chow WH, Blot WJ, Vaughan TL, Risch HA, Gammon MD, Stanford JL, Bosland MC, Oakley-Girvan I, Whittemore AS. 1999.Dietary fat, calories,and Dubrow R, SchoenbergJB, Mayne ST, Farrow DC, Ahsan H, West AB, prostatecancer risk. J Natl Cancer Inst 91:489-491. Rotterdam H, Niwa S, Fraumeni JF Jr. 1998. Body mass index and risk Bosland MC. 2000. The role of steroid hormones in prostatecarcinogenesis. of adenocarcinomasof the esophagusand gastric cardia. J Natl Cancer J Natl Cancer Inst Monogr 27:39-66. Inst 90:150-155. Bostick RM, Potter JD, Kushi LH, SellersTA, Steinmetz KA, McKenzie DR, Chow WH, Gridley G, Fraumeni JF Jr, Jarvholm B. 2000. Obesity, hyper- Gapstur SM, Folsom AR. 1994. Sug*, meat, and fat intake, and non- tension, and the risk of kidney cancer in men. N Engl J Med 343: dietary risk factors for colon cancer incidence in lowa women. United 1305-l3ll. States.Cancer CausesControl 5:38-52. Chu SY Lee NC, Wingo PA, SenieRT, GreenbergRS, PetersonHB. 1991.The Bouchardy C, Le MG, Hill C. 1990. Risk factors for breastcancer according relationshipbetween body massand breastcancer among women enrolled to age at diagnosis in a French case-control study. J Clin Epidemiol in the Cancer and Steroid Hormone Studv. J Clin Eoidemiol 44: 43:267-275. rr97-1206. Obesityand BodyComposition M1

Chyou PH, Nomura AM. Stemmermann GN. 1994. A prospective study of Dietz AT, Newcomb PA, Marcus PM, StorerBE. I 995. The associationof body weight, body mass index and other anthropometric measurementsin size and large bowel cancer risk in Wisconsin (United States)women. relation to site-specificcancers. Int J Cancer 57:313-317. Cancer CausesControl 6:3G-36. Clarke G, Whittemore AS. 2000. Prostate cancer risk in relation to anthro- Drinkard CR, SellersTA, PotterJD, Zheng W, Bostick RM, Nelson CL, Folsom pometry and physical activity: the National Health and Nutrition Exam- AR. 1995.Association of body mass index and body fat distribution with ination Survey I Epidemiological Follow-Up Study. Cancer Epidemiol risk of lung cancer in older women. Am J Epidemiol 142:60H07. BiomarkersPrev 9;875-88l. Ekbom A, Trichopoulos D, Adami HO, Hsieh CC, Lan SJ. 1992. Evidence of Clayton SJ, May FE, WestleyBR. 1997.Insulin-like growth factorscontrol the prenatalinfluences on breastcancer risk. Lancet 340:1015-1018. regulation of oestrogenand progesteronereceptor expression by oestro- Ekbom A, Hsieh CC, Lipworth L, Adami HQ, TrichopoulosD. 1997.Intrauter- gens.Mol Cell Endocrinol 128:57-68. ine environment and breast cancer risk in women: a population-based Coates RJ, Clark WS, Eley JW, GreenbergRS, Huguley CM Jr, Brown RL. study.J Natl CancerInst 89:71-76. 1990. Race, nutritional status, and survival from breast cancer. J Natl El Tanani MK, Green CD. 1997. Interaction between estradiol and growth Cancer Inst 82:1684-1692. factors in the regulation of specific gene expression in MCF-7 human CoatesRJ, Uhler RJ, Hall HI, PotischmanN, Brinton LA, Ballard-BarbashR, breastcancer cells. J Steroid Biochem Mol Biol 6O:269-276. Gammon MD, Brogan DR, Daling JR, Malone KE, Schoenberg JB, Elliott EA, Matanoski GM, RosensheinNB, Grumbine FC, Diamond EL. 1990. SwansonCA. 1999.Risk of breastcancer in young women in relationto Body fat patterning in women with endometrial cancer.Gynecol Oncol body size and weight gain in adolescenceand early adulthood. Br J 39:253-258. Cancer8l:16'7-174. Enger SM, Ross RK, Paganini-Hill A, CarpenterCL, BernsteinL. 2000. Body Colleni RB, CopelandKC, Devlin JT, RobertsJD, McAuliffe TL. 1991.Effect size, physical activity, and breastcancer hormone receptorstatus: results of obesity on plasma insulin-like growth factor-I in cancerpatients. Int J from two case-control studies. Cancer Epidemiol Biomarkers Prev Obes 15:523-527. 9:681-687. ConcolinoG, LubranoC, SantonatiA, Di SilverioF, CatizoneA, Conti C. 1989. Evans DJ, Hoffmann RG, Kalkhoff RK, KissebahAH. 1984. Relationship of Epidermal growth factor (EGF) binding by normal and neoplastichuman body fat topography to insulin sensitivity and metabolic profiles in pre- renal tissue.J Tumor Marker Oncol 4:89-97. menopausalwomen. Metabolism33:68-75. D'Avanzo B, La Vecchia C, FranceschiS, Negri E, Talamini R. 1995. History Ewertz M, Machado SG, Boice JD Jr, JensenOM. 1984. Endometrial cancer of thyroid diseasesand subsequentthyroid cancerrisk. CancerEpiderniol following treatmentfor breastcancer: A case-controlstudy in Denmark. BiomarkersPrev 4: 193-199. Br J Cancer 50:687492. D'Avanzo B, La Vecchia C, Talamini R, FranceschiS. 1996.Anthropometric Fairfield KM, Willett WC, RosnerBA, Manson JE, SpeizerFE, Hankinson SE. measuresand risk of cancersof the upper digestiveand respiratory tract. 2002. Obesity, weight gain, and ovarian cancer. Obstet Gynecol Nutr Cancer 26:219-227 . 100:288-296. Dal Maso L, La Vecchia C, FranceschiS, Preston-Martin S, Ron E, Levi F, Farrow DC, Weiss NS, Lyon JL, Daling JR. 1989.Association of obesity and Mack W Mark SD, McTiernan A, Kolonel L, Mabuchi K, Jin F, Wingren ovarian cancer in a case-controlstudy. Am J Epidemiol 129:1300-1304. G, Galanti MR, Hallquist A, Glattre E, Lund E, Linos D, Negri E. 2000. Finkle WD, Mclaughlin JK, Rasgon SA, Yeoh HH, Low JE. 1993.Increased A pooled analysis of thyroid cancer studies.V. Anthropometric factors. risk of renal cell cancer among women using diuretics in the United Cancer CausesControl ll:137-144. States.Cancer CausesControl 4:555-558. Dales LG, Friedman GD, Ury HK, GroSsmanS, Williams SR. 1979. A case- Flegal KM, Carroll MD, Ogden CL, JohnsonCL.2002. Prevalenceand trends control study of relationshipsof diet and other traits to colorectal cancer in obesity among US adults, 1999-2000. JAMA 288:1723-1727. in American blacks. Am J Epidemiol 109:132-144. Foekens JA, Portengen H, JanssenM, Klijn JG. 1989. Insulin-like growth Davidow AL, Neugut AI, JacobsonJS, Ahsan H, Garbowski GC, Forde KA, factor-l receptorsand insulin-like growth factor-l-like activity in human Treat MR, Waye JD. 1996.Recunent adenomatouspolyps and body mass primary breast cancer. Cancer 63:2139-2147. index. CancerEpidemiol BiomarkersPrev 5:313-315. Folsom AR, Kaye SA, Potter JD, Prineaszu. 1989.Association of incident car- Day GL, Blot WJ, Austin DF, Bernstein L, GreenbergRS, Preston-Martin S, cinoma of the endometrium with body weight and fat distribution in older SchoenbergJB, Winn DM, Mclaughlin JK, FraumeniJF Jr. 1993.Racial women: Early findings of the Iowa Women's Health Study. Cancer Res differencesin risk of oral and pharyngealcancer: Alcohol, tobacco, and 49:6828-683l. other determinants.J Natl Cancer Inst 85:465-473. Folsom AR, Kaye SA, PrineasRI, Potter JD, Gapstur SM, Wallace RB. 1990. De StavolaBL, Hardy R, Kuh D, Silva IS, WadsworthM, Swerdlow AJ. 2000. Increasedincidence of carcinoma of the breast associatedwith abdomi- Birthweight, childhood growth and risk of breast cancer in a British nal adiposity in postmenopausalwomen. Am J Epidemiol 131:794-803. cohort. Br J Cancer 83:964-968. Ford ES. 1999.Body massindex and colon cancerin a national sampleof adult de VasconcelosAB, Azevedo e Silva Mendonca G, Sichieri R. 2001. Height, US men and women.Am J Epidemiol 150:390-398. weight, weight changeand risk of breastcancer in Rio de Janeiro,Brazil. FranceschiS, Levi F, Negri E, FassinaA, La VecchiaC. 1991.Diet and thyroid Sao Paulo Med J 119:6246. cancer: A pooled analysis of four European case-control studies.Int J de Waard F, Baanders-vanHalewijn EA. 1974.A prospectivestudy in general Cancer48:395-398. practice on breast-cancerrisk in postmenopausalwomen. Int J Cancer Franceschi S, Favero A, La Vecchia C, Baron AE, Negri E, Dal Maso L, l4:153-160. Giacosa A, Montella M, Conti E, Amadori D. 1996. Body size indices de Waard F, Ramlau R, Mulders Y, de Vries T, van WaverenS. 1993.A feasi- and breast cancer risk before and after menopause. Int J Cancer bility study on weight reduction in obese postmenopausalbreast cancer 67:181-186. patients.Eur J Cancer Prev 2:233-238. FranceschiS, Dal Maso L, Levi F, Conti E, Thlamini R, La VecchiaC. 2001. Del Giudice ME, Fantus lG,Ezzat S, McKeown-EyssenG, Page D, Goodwin Leanness as early marker of cancer of the oral cavity and pharynx. Ann PJ. 1998.Insulin and relatedfactors in premenopausalbreast cancer risk. Oncol 12:331-336. Breast Cancer Res Treat 4'1:l1l-120. Friedenreich CM, Thune L 2001. A review of physical activity and prostate Demark-WahnefriedW, Winer EP, Rimer BK. 1993.Why women gain weight cancer risk. Cancer CausesControl 12:461475. with adjuvant chemotherapy for breast cancer.J Clin Oncol 11:1418- Friedenreich CM, Courneya KS, Bryant lfE. 2002. Case-control study of 1429. anthropometric measuresand breast cancer risk. Int J Cancer 99:445452. Demark-WahnefriedW' Conaway MR, RobertsonCN, Mathias BJ, Anderson Furberg AS, Thune I. 2003. Metabolic abnormalities (hypertension,hyper- EE, PaulsonDF. 1997a.Anthropometric risk factors for prostatecancer. glycemia and overweight), lifestyle (high energyintake and physicalinac- Nutr Cancer 28:302-307 . tivity) and endometrial cancer risk in a Norwegian cohort. Int J Cancer Demark-Wahnefried W, Rimer BK, Winer EP. 1997b. Weight gain in 104:669476. women diagnosed with breast cancer. J Am Diet Assoc 97:519-526, Galanis DJ, Kolonel LN, Lee J, Le Marchand L. 1998. Anthropometric pre- s29. dictors of breast cancer incidence and survival in a multi-ethnic cohort of Demark-Wahnefried W, PetersonBL, Winer EP, Marks L, Aziz N, Marcom PK, female residents of Hawaii, United States. Cancer Causes Control Blackwell K, Rimer BK. 2001. Changes in weight, body composition, 9:217-224. and factors influencing energy balance among peremenopauslabreast Galanti MR, HanssonL, Bergstriim R, WolkA, HjartakerA, Lund E, Grimelius cancerpati€nts receiving adjuvant chemotherapy.J Clin Oncol 19:2367- L, Ekbom A. 1997. Diet and the risk of papillary and follicular thyroid 2369. carcinoma:a population-basedcase-control study in Swedenand Norway. den TonkelaarI, de Waard F, Seidell JC, FracheboudJ. 1995.Obesity and sub- Cancer CausesControl 8:205-214. cutaneousfat patterning in relation to survival of postmenopausalbreast Gallus S, Boseni C, FranceschiS, Levi F, SimonatoL, Negri E, La VecchiaC. cancerpatients participating in the DOM-project. BreastCancer Res Treat 2001. Oesophagealcancer in women: Tobacco, alcohol, nutritional and 34:129-137. hormonal factors. Br J Cancer 85:341-345. M2 PARTlll: THECAUSES OF CANCER

Garfinkel L. 1986. Overweight and mortality. Cancer 58:1826-1829. Haffner SM, Katz MS, Stern MP, Dunn JF. 1989.Relationship of sex hormone Gerhardssonde Verdier M, Hagman U, SteineckG, RiegerA, Norell SE. 1990. binding globulin to overall adiposity and body fat distribution in a bieth- Diet, body mass and colorectal cancer: a case-referentstudy in Stock- nic population.Int J Obes 13:l-9. holm. Int J Cancer46:832-838. Haffner SM, Miettinen H, KarhapaaP, Mykkanen L, Laakso M. 1997. Leptin Giovannucci E, Ascherio A, Rimm EB, Colditz GA, StampferMJ, Willett WC. concentrations,sex hormones, and cortisol in nondiabetic men. J Clin 1995. Physical activity, obesity, and risk for colon cancer and adenoma Endocrinol Metab 82:I 807-l 809. in men. Ann Intem Med 122:327-334- Hale GE, HughesCL, Cline JM.2002. Endometrial cancer:Hormonal factors, Giovannucci E, Colditz GA, Stampfer MJ, Willett WC. 1996. Physical activ- the perimenopausal"window of risk," and isoflavones.J Clin Endocrinol ity, obesity, and risk of colorectal adenoma in women. United States. Metab87:3-15. Cancer CausesControl 7:253-263. Hall IJ, Newman B, Millikan RC, Moorman PG. 2000. Body size and breast Giovannucci E. Rimm EB, Stampfer MJ, Colditz GA, Willett WC. 1997. cancer risk in black women and white women: The Carolina Breast Height, body weight, and risk of prostate cancer. Cancer Epidemiol Cancer Study.Am J Epidemiol l5l:754-764. Biomarkers Prev 6:557-563. Hallquist A, Hardell L, DegermanA, Boquist L. 1994.Thyroid cancer:repro- Giuffrida D, Lupo L, La Porta GA, La Rosa GL, Padova G, Foti E, Marchese ductive factors,previous diseases,drug intake. family history and diet. A V Belfiore A. 1992. Relation between steroid receptor starusand body case-controlstudy. Eur J Cancer Prev 3:481-488. weight in breastcancer patients. Eur J Cancer28:112-115. HankinsonSE, Willett WC, Colditz GA, Hunter DJ, Michaud DS, Deroo B, Glantz GM. 1964. Cirrhosis and carcinoma of the prostate gland. J Urol Rosner B, Speizer FE, Pollak M. 1998. Circulating concentrationsof 9l:291-293. insulin-like growth factor-I and risk of breastcancer. Lancet 351:1393- GlassAR. 1989.Endocrine aspects of obesity.Med Clin North Am 73:139-160. 1396. Glattre E, HaldorsenT, Berg JP,Stensvold I, Solvoll K. 1993.Norwegian case- Harris RE, Namboodiri KK, Wynder EL. 1992. Breast cancer risk: Effects of control study testing the hypothesis that seafood increasesthe risk of estrogenreplacement therapy and body mass.J Natl CancerInst 84:1575- thyroid cancer.Cancer Causes Control 4:ll-16. r582. Goodman MT, Morgenstern H, Wynder EL. 1986. A case-control study of Hartge P, Schiffman MH, Hoover R, McGowan L, Lesher L, Norris HJ. 1989. factors affecting the developmentof renal cell cancer.Am J Epidemiol A case-controlstudy of epithelial ovarian cancer.Am J Obstet Gynecol 124:926-941. l6l:10-16. Goodman MT, Kolonel LN, Wilkens LR. 1992.The associationof body size, HartzAJ, RupleyDC, Rimm AA. 1984.The associationof girth measurements reproductive factors and thyroid cancer.Br J Cancer 66: I 180-l 184. with diseasein 32,856women. Am J Epidemiol119:71-80. Goodman MT, Wilkens LR. 1993. Relation of body size and the risk of lung Hayes RB, Ziegler RC, Gridley G, SwansonC, GreenbergRS, SwansonGM, cancer. Nutr Cancer 20:179-186. SchoenbergJB, Silverman DT, Brown LM, Pottern LM, Litr J, Schwartz Goodman MT, Hankin JH, Wilkens LR, Lyu LC, McDuffie K, Liu LQ, Kolonel AG, Fraumeni JF Jr, Hoover RN. 1999. Dietary factors and risks for LN. 1997. Diet, body size, physical activity, and the risk of endometrial prostate cancer among blacks and whites in the United States.Cancer cancer.Cancer Res 57:5077-5085. Epidemiol Biomarkers Prev 8:25-34. Coodman-Gruen D, Barrett-Connor E. 1997. Epidemiology of insulin-like HeasmanKZ, SutherlandHJ, CampbellJA, ElhakimT. Boyd NF. 1985.Weight growth factor-I in elderly men and women. The Rancho Bemardo Study. gain during adjuvantchemotherapy for breastcancer. Breast Cancer Res Am J Epidemiol 145:970-976. Treat 5:195-200. Goodwin PJ, PanzarellaT, Boyd NF. 1988. Weight gain in women with local- Hebert JR, Augustine A, Barone J, Kabat GC, Kinne DW, Wynder EL. 1988. ized breast s4ngs1-{ descriptive study. Breast Cancer Res Treat Weight, height and body mass index in the prognosis of breast cancer: I l:59-66. Early results of a prospectivestudy. Int J Cancer 42:315-318. Goodwin PJ, Boyd NF. 1990. Body size and breastcancer prognosis:A criti- Hebert JR, Hurley TG, Olendzki BC, TeasJ, Ma Y, Hampl JS. 1998.Nutri- cal review of the evidence.Breast Cancer Res Treat 16:205-214. tional and socioeconomicfactors in relation to prostatecancer mortality: Goodwin PJ,Ennis M, Pritchard KI, McCready D, Koo J, Sidlofsky S, Trudeau a cross-nationalstudy. J Natl Cancer Inst 90:1631-1647. M, Hood N, Redwood S. 1999. Adjuvant treatment and onset of Heitmann BL. 1993.The influenceof fatness,weight change,slimming history menopausepredict weight gain after breastcancer diagnosis. J Clin Oncol and other lifestyle variableson diet reporting in Danish men and women 17:120-129. aged 35-65 years.Int J Obes Relat Metab Disord l7:329-336. Goodwin PJ, Ennis M, Pritchard KI, TrudeauME, Koo J, Hartwick W, Hoffma HendersonBE, Ross RK, Pike MC, CasagrandeJT. 1982.Endogenous hor- B, Hood N. 2002. Insulin-like growth factor binding proteins I and 3 and mones as a major factor in human cancer.Cancer Res 42:3232-3239. breastcancer outcomes.Breast Cancer Res Treat 74:65-76. Henley SJ, FlandersWD, Manatunga A, Thun MJ. 2002. Leannessand lung Goodwin PJ, Ennis M, Pritchard KI, Trudeau ME, Koo J, Madarnas Y. cancer risk: Fact or artifact'l Epidemiology 13:268-276. Hartwick W Hoffman B, Hood N. 2002. Fasting insulin and outcome in Hiatt RA, Tolan K, QuesenberryCP Jr. 1994. Renal cell carcinoma and thi- early-stagebreast cancer: results of a prospective cohort study. J Clin azide use: A historical, case-control study. California, USA. Cancer Oncol 20:42-51. CausesControl 5:319-325. GrahamS, Dayal H, SwansonM, MittelmanA, WilkinsonG. 1978.Diet in the Hilakivi-Clarke L, ForsenT, Eriksson JG, Luoto R, Tuomilehto J, Osmond C, epidemiology of cancer of the colon and rectum. J Natl Cancer Inst Barker DJ. 2001. Tallnessand overweight during childhood have oppos- 6l:709-714. ing effectson breastcancer risk. Br J Cancer85: l68G-1684. Graham S, Haughey B, Marshall J, Priore R, Byers I Rzepka T, Mettlin C, Hirose K, Tajima K, Hamajima N, Takezaki T, Inoue M, Kuroishi T, Miura S, PontesJE. 1983. Diet in the epidemiology of carcinoma of the prostate Tokudome S. 2001. Association of family history and other risk factors gland.J Natl CancerInst 70;687-692. with breast cancer risk among Japanese premenopausal and post- Graham S, Marshall J, Haughey B, Mittelman A, Swanson M, Zielezny M, menopausalwomen. Cancer Causes Control 12:349-358. Byers T, Wilkinson G, West D. 1988. Dietary epidemiology of cancer of Hislop TG, ColdmanAJ, Elwood JM, BrauerG, Kan L. 1986.Childhood and the colon in western New York. Am J Epidemiol 128:490-503. recent eating patterns and risk of breast cancer. Cancer Detect Prev GreenbergER, VesseyMP, McPherson K, Doll R, YeatesD. 1985. Body size 9:41-58. and survival in premenopausalbreast cancer. Br J Cancer 5l 691-697. Ho SC, Tai ES, Eng PH, Ramli A, Tan CE, Fok AC. 1999.A study in the rela- Greggi S, Parazzini F, ParatoreMP, Chatenoud L, Legge F, Mancuso S, La tionshipsbetween leptin, insulin, and body fat in Asian subjects.lnt J VecchiaC. 2000. Risk factors for ovarian cancer in central ltaly. Gynecol Obes Relat Metab Disord 23:246-252. Oncol 79:50-54. Hsieh CC, TrichopoulosD, KatsouyanniK, YuasaS. 1990.Age at menarche, Gronberg H, Damber L, Damber IE. 1996.Total food consumption and body age at menopause,height and obesity as risk factors for breast cancer: mass index in relation to prostate cancer risk: A case-control study in associationsand interactionsin an international case-controlstudy. Int J Swedenwith prospectivelycollected exposure data. J Urol 155:969-9-74. Cancer46:796-800. Gunnell D, BerneyL, Holland P, MaynardM, Blane D, FrankelS, Smith GD. Hsieh CC, Thanos A, Mitropoulos D, Deliveliotis C, Mantzoros CS, 2000. How accuratelyare height, weight and leg length reporred by the Trichopoulos D. 1999. fusk factors for prostate cancer: A case-control elderly, and how closely are they related to measurementsrecorded in study in Greece.Int J Cancer80:699-703. childhood? Int J Epidemiol 29:456464. Hsing AW, Deng J, SesterhennIA, Mostofi FK, Stanczyk FZ, Benichou J, Xie Gunnell D, Okasha M, Smith GD, Oliver SE, SandhuJ, Holly JM. 2001. I Gao YT. 2000. Body size and prostate cancer: a population-based Height, leg length, and cancerrisk: A systematicreview. Epidemiol Rev case-controlstudy in China.Cancer Epidemiol Biomarkers Prev 9:1335- ,.,i-: 23:313-342. 134r. Habel LA, Van Den EedenSK, FriedmanGD. 2000. Body size. age at shaving Hsing AW. Chua S Jr, Gao YT, GentzscheinE. Chang L, Deng J. StanczykFZ. tr initiation, and prostate cancer in a large, multiracial cohort. Prostate 2001. Prostatecancer risk and serum levels of insulin and leptin: A .'+...:l 43:I 36-t43. population-basedstudy. J Natl CancerInst 93:783-789. : i:l . rag" . :t" .t +:. .'i.:* :E', * Obesityand BodyComposition 443

Huang Z. HankinsonSE. Colditz GA. StampferMJ. Hunter DJ, Manson JE. Kalish LA. 1984.Relationships of body size with brcastcanccr. J Clin Oncol HennekensCH. RosnerB, SpeizerFE. Willett WC. 1997.Dual eff'ectsof 2:281-293. weight and weight gain on breas(cancer risk. JAMA 218:1401-l4ll. Kark JD, YaariS, RasoolyI. GoldbourtU. 199-5.Arc learrsrrrokcrs at increased Huang Z, Willett WC, Colditz GA. Hunter DJ. MansonJE, RosnerB. Speizer risk of lung cancer?The IsraelCivil ServantCarrcer Stuch. Arch Intern FE, HankinsonSE. 1999.Waist circumference.waist:hip ratio. and nsk Med 155:2409-2416. of breastcancer in the Nurses'Health Study.Am J Epidemiol 150: Kato I. NomuraA. StemmermannGN, Chyou PH. 1991.Dctenninilnts of sex 1316-1324. hormonelel'els in men as usefulindices in honnone-rclateddisorders. J IARC Working Group. 1986. lntemationalAgency for Researchon Cancer: Clin Epidemiol45: 141'1-1421. Tobaccosmoking. IARC monographson the evaluationof the carcino- Kaye SA, FolsomAR, Soler JT, PrineasRJ, PotterJD. 199l. Associltionsol genic risks to humans.vol 38. Lyon, France;IARC Press. body massand fat distributionwith sex hormoneconcentrations in posl- IARC Working Group. 1988. InternationalAgency for Researchon Cancer: menopausalwomen. Int J Epiderniol20:151-156. Alcohol drinking. IARC monographson the evaluationof the carcino- Kelly PJ,Eisman JA, StuartMC. PocockNA. SambrookPN, Gwinn1'll. 199() genic risks to humans.vol 44. Lyon. France:IARC Press. Somatomedin-C,physical fitness,and bone density.J Clin Enducrinol IARC Working Group. 2002. IARC Handbooksof CancerPrevention. vol. 6. Metab70:1)8-123. Weight control and physical activity. Lyon, France: IARC Press.pp. Key TJ, Pike MC. 1988a.The dose-effectrelationship between "unopposerl" l-315. oestrogensand endometrial rnitotic rate: its centralrole in explainingantl Ilic M. Vlajinac H. Marinkovic J. 1996.Case-control study of risk factorsfor predicting endometrialcancer nsk. Br J Cancer 57:205-212. prostatecancer. Br J Cancer74:1682-1686. Key TJ, Pike MC. 1988b.The role of oestrogensand progestagensin the cpi- Ingram D, NottageE, Ng S, SparrowL, RobertsA, Willcox D. 1989.Obesity demiology and preventionof breastcancer. Eur J Cancer Clin Oncol and breast disease.The role of the female sex hormones. Cancer 24:2943. 64:I 049-l 053. Key TJ, SilcocksPB, DaveyGK, ApplebyPN, BishopDT. 199'7. Acase-control InnesK, ByersT, SchvmuraM. 2000.Birth characteristicsand subsequentrisk study of diet and prostatecancer. Br J Cancer16:618-681. for breastcancer in very young women.Am J Epidemiol 152:1121-1128. Key TJ, Allen NE, VerkasaloPK, Banks E. 2001. Energybalance and cancer: Inoue M. OkayamaA. Fujita M. EnonrotoT. TanizawaO, UeshimaH. 1994. The role of sex hormones.Proc Nutr Soc 60:81-89. A case-control study on risk factors for uterine endornetrial cancer in Kirschner MA, Ertel N, SchneiderG. 1981.Obesity, hormones, and cancer. Japan.Jpn J CancerRes 8-5:346-3,50. CancerRes 4l :3-7ll-3'717. Inoue M, Ito LS. Tajima K. YamamuraX KoderaY TakezakiI HamajimaN, KirschnerMA, Samojlik E, Drejka M. Szmal E, SchneiderG, Ertel N. 1990. Hirose K, Kuroishi T, TominagaS. 2002. Height.weight. menstrualand Androgen-estrogenmetabolism in women with upperbody versuslower reproductive factors and nsk of gastric cancer among Japanesepost- body obesity.J Clin EndocrinolMetab 10:473419. rrenopausalwonren: Analysis by subsiteand histologic subtype.Int J Knekt P,Heliovaara M. RissanenA, AromaaA, SeppanenR. TeppoL. Pukkala Cancer97:833-838. E. 1991.Leanness and lung-cancerrisk. lnt J Cancer49:208-213. Iribarren C, HaselkornT. Tekawa IS. FriedmanGD.2001. Cohort study of Kolonel LN, NomuraAM, Lee J, HirohataT. 1986.Anthropometric indicators thyroid cancer in a San FranciscoBay area population.lnt J Cancer of breastcancer risk in postmenopausalwomen in Hawaii. Nutr Cancer 93:'745-'750. 8:247-256. Jain MG, Howe GR. Rohan TE. 2000. Nutritional factors and endometrial Kolonel LN, YoshizawaCN, Hankin JH. 1988.Diet and prostaticcancer: A cancerin Ontario,Canada. Cancer Control 1:288-296. case-controlstudy in Hawaii. Am J Epidemiol 127:999-1012. JanssenJA. Stolk RP,Pols HA. GrobbeeDE. de JongFH, LambertsSW. 1998. Kolonel LN, Hankin JH, Wilkens LR, FukunagaFH, Hinds MW. 1990.An Serumfree IGF-I, total IGF-I, IGFBP-l and IGFBP-3levels in an elderly epidemiologicstudy of thyroid cancerin Hawaii. CancerCauses Control population:relation to age and sex steroidlevels. Clin Endocrinol(Oxfl l:223-234. 48:4'71178. Kono S, Shinchi K, Ikeda N, Yanai F. lmanishi K. 1991.Physical activity, JasienskaC, ThuneI. 2001. Lifestyle,hormones, and risk of breastcancer. BMJ dietaryhabits and adenomatouspolyps of the sigmoid colon:A study of 322:586-587. self-defenseofficials in Japan.J Clin Epidemiol 44:1255-1261. JernstromH, Barrett-ConnorE. 1999.Obesity, weight change,fasting insulin. Kono S, Handa K, Hayabuchi H, Kiyohara C, lnoue H, MzrrugameT, proinsulin,C-peptide. and insulin-likegrowth factor-l levels in women ShinomiyaS. HamadaH, Onuma K, Koga H. 1999.Obesity. weight gain with and without breastcancer: The RanchoBernardo Study. J Womens and risk of colon adenomasin Japanesemen. Jpn J Cancer Res Health Gend BasedMed 8:1265-1272. 90:805-81l. Ji BT. Chow WH, Yang G, Mclaughlin JK, Gao RN, Zheng W Shu XO. Jin Kopelman PG. 2000. Obesity as a medical problem. Nature 404:635443. F. FraumeniJF Jr,Gao YT. \997. Body massindex and the risk of cancers Koumantaki Y, Tzonou A, Koumantakis E, Kaklamani E, Aravantinos D, of the gastric cardia and distal stomach in Shanghai,China. Cancer TrichopoulosD. 1989.A case-controlstudy of cancerof endometriumin Epidemiol BiomarkersPrev 6:48I --485. Athens. Int J Cancer 43:'795-'199. Johannsson G, Karlsson C, Lonn L, Marin P, Bjorntorp P, Sjostrom L, KreigerN, MarrettLD, DoddsL, Hilditch S, DarlingtonGA. 1993.Risk factors CarlssonB, CarlssonLM, BengtssonBA. 1998.Serum leptin concentra- for renal cell carcinoma:results of a population-basedcase-control study. tion and insulin sensitivity in men with abdominalobesity. Obes Res CancerCauses Control 4: l0l-l 10. 6:416421. Kubik AK. Zatloukal P, Tomasek L, Petruzelka L. 2002. Lung cancer risk Kaaks R, van Noord PA, den Tonkelaar I, PeetersPJ, Riboli E, Grobbee DE. among Czech women: a case-controlstudy. Prev Med 34:436 444. 1998.Breast-cancer incidence in relationto height,weight and body-fat Kuczmarski MF, Kuczmarski RI, Najjar M. 2001. Effects of age on validity of distributionin the Dutch "DOM" cohort.Int J Cancer16:64'1451. self-reported height, weight, and body mass index: findings from the Kabat GC, Wynder EL. 1992.Body mass index and lung cancerrisk. Am J Third National Health and Nutrition ExaminationSurvey, 1988-1994. J Epidemiol 135:169-71 4. Am Diet Assoc 101:28-34. KabatGC. Ng SK, Wynder EL. 1993.Tobacco, alcohol intake, and diet in rela- Kune GA, Kune S, Watson LF. 1990. Body weight and physical activity as tion to adenocarcinomaof the esophagusand gastric cardia. Cancer predictors of colorectal cancer risk. Nutr Cancer l3:9-l'7. CausesControl 4:123-132. Kyogoku S, Hirohata T, TakeshitaS, Nomura Y, ShigematsuT, Horie A. 1990. Kabat GC, Chang CJ, Wynder EL. 1994.The role of tobacco,alcohol use, and Survival of breast-cancerpatients and body size indicators.Int J Cancer body mass index in oral and pharyngeal cancer. lnt J Epidemiol 46:824-831. 23:ll3'7-1144. La VecchiaC, FranceschiS, DecarliA, GallusG, TognoniG. 1984.Risk factors KadamaniS, Asal NR, NelsonRY. 1989.Occupational hydrocarbon exposure for endometrialcancer at different ages.J Natl Cancer lnst73:66741l. and risk of renalcell carcinoma.Am J lnd Med l5:131-141. La VecchiaC,Parazzini F. Negri E, Fasoli M, GentileA, FranceschiS. 1991. Kaijser M, LichtensteinP, GranathF, ErlandssonG, CnattingiusS, Ekbom A. Anthropometric indicators of endometrial cancer risk. Eur J Cancer 2001. ln utero exposuresand breast cancer: a study of opposite-sexed 27:481490. twins. J Natl Cancerlnst 93:60-62. La VecchiaC, Ron E, FranceschiS. Dal MasoL, Mark SD. ChatenoudL, Braga KaklamaniVG, Linos A, Kaklamani E, Markaki I, MantzorosC. 1999.Age, C, Preston-MartinS, McTiernanA, Kolonel L, Mabuchi K, Jin F, Wingren sex, and smoking are predictorsof circulatinginsulin-like growth factor G, Galanti MR, HallquistA, Lund E, Levi F, Linos D, Negri E. 1999.A I and insulin-like growth factor-binding protein 3. J Clin Oncol pooled analysis of case-controlstudies of thyroid cancer.III. Oral con- 17:813-817. traceptives.menopausal replacement therapy and other female hormones. Kalandidi A, Tzonou A, Lipworth L, Gamatsi l, Filippa D, Trichopoulos D. CancerCauses Control 10:157-166. 1996. A case-control study of endometrial cancer in relation to LagergrenJ, BergstrdmR, Nyren O.1999. Associationbetween body massand reproductive, somatometric,and life-style variables.Oncology 53:354- adenocarcinomaof the esophagusand gastric cardia. Ann Intern Med 3s9. I 30:883-890.

:::,: ,-.i 4M PARTlll: THECAUSES OF CANCER

Lahmann PH, Lissner L, Gullberg B, Olsson H, Berglund G. 2003. A prospec- Lubin F, Chetrit A, FreedmanLS, Alfandary E, Fishler Y, Nitzan H, Zultan A, tive study of adiposity and postmenopausalbreast cancer risk: The Modan B. 2003. Body mass index at age l8 years and during adult life Malmo Diet and Cancer Study. Int J Cancer 103:246-252. and ovarian cancerrisk. Am J Epidemiol 157 ll3-120. Lam PB, VacekPM, Geller BM, Muss HB. 2000. The associationof increased Lukanova A, Toniolo P, Lundin E, Micheli A, Akhmedkhanov A, Muti P, weight, body mass index, and tissue density with the risk of breast car- Zeleniuch-Jacquotte A, Biessy C, Lenner P, Krogh V Berrino F, cinoma in Vermont. Cancer 89:369-375. Hallmans G, Riboli E, Kaaks R. 2002. Body mass index in relation to Landin-Wilhelmsen K, Wilhelmsen L, Lappas G, Rosen T, Lindstedt G, ovarian cancer: A multi-centre nested case-control study. Int J Cancer Lundberg PA, BengtssonBA. 1994.Serum insuiin-like growth factor I in 99:603-608. a random population sample of men and women: relation to age, sex, Lund E, Adami HO, Bergstr

MellemgaardA, EngholmG, Mclaughlin JK, OlsenJH. 1994.Risk factorsfor Ogden CL, Flegal KM, Canoll MD, JohnsonCL.2002. Prevalenceand trends renal--cellcarcinoma in Denmark. III. Role of weight, physical activity in overweight among US children and adolescents,1999-2000. JAMA and reproductive factors. Int J Cancer 56:.66-71. 288:1'728-l'732. MellemgaardA, Lindblad P, Schlehof'erB, BergstrcimR, Mandel JS, McCredie Olson JE, Yang P, Schmitz K, Vierkant RA, Cerhan JR, SellersTA. 2002. Dif- M, Mclaughlin JK, Niwa S, Odaka N, PommerW. 1995.International ferential associationof body mass index and fat distribution with three renal-cellcancer study. III. Role of weight, height,physical activity, and major histologic rypes of lung cancer: Evidence from a cohort of older useof amphetamines.Int J Cancer60:350-354. women.Am J Epidemiol 156:60ffi15. Mettlin C, SelenskasS, NatarajanN, Huben R. 1989.Beta-carotene and animal Olson SH, Trevisan M, Marshall JR, Graham S, Zielezny M, Vena JE, fats and their relationship to prostate cancer risk. A case-controlstudy. Hellmann R, FreudenheimJL. 1995.Body mass index, weight gain, and Cancer 64:605412. risk of endometrialcancer. Nutr Cancer 23:14l-149. Michels KB, TrichopoulosD, RobinsJM, RosnerBA, MansonJE. Hunter DJ, PaffenbargerRS Jr, Kampert JB, Chang HG. 1980.Characteristics that predict Colditz GA, HankinsonSE, SpeizerFE, Willett WC. 1996.Birthweight risk of breast cancer before and after the menopause.Am J Epidemiol as a risk factor for breastcancer. Lancet 348:1542-1546. 112:258-268. Mills PK, BeesonWL, Phillips RL, FraserGE. 1989.Cohort study of diet, Parazzini F, La Vecchia C, Negri E, Bruzzi P, Palli D, Boyle P. 1990. lifestyle, and prostatecancer in Adventist men. Cancer 64:598-604. Anthropometric variablesand risk of breastcancer. Int J Cancer 45:397- Mink PJ,Folsom AR, SellersTA, Kushi LH. 1996.Physical activity, waist-to- 402. hip ratio, and other risk factors for ovarian cancer: a follow-up study of Parazzini F, Moroni S, La Vecchia C, Negri E, dal Pino D, Bolis G. 1997. older women. Epidemiology7:3845. Ovarian cancerrisk and history of selectedmedical conditionslinked with Mohle-Boetani JC, Grosser S, Whinemore AS, Malec M, Kampert JB, female hormones.Eur J Cancer 33:1634-1637. PaffenbargerRS Jr. 1988. Body size, reproductive factors, and breast Parslov M, Lidegaard O, Klintorp S, PedersenB, Jonsson L, Eriksen PS, cancer survival. Prev Med 17:634442. Ottesen B. 2000. Risk factors among young women with endometrial Mgller H, MellemgaardA, Lindvig K, Olsen JH. 1994.Obesity and cancerrisk: cancer: A Danish case-control study. Am J Obstet Gynecol 182:23- A Danish record-linkagestudy. Eur J Cancer 30A:344-350. 29. Moore JW Key TJ, Bulbrook RD, Clark GM, Allen DS, Wang DY, Pike MC. PartanenT, Heikkila P, Hernberg S, KauppinenT, Moneta G, Ojajarvi A. 1991. 1987.Sex hormone binding globulin and risk factors for breastcancer in Renal cell cancerand occupationalexposure to chemical agents.Scand J a population of normal women who had never used exogenoussex hor- Work Environ Health 17:231-239. mones.Br J Cancer 56:661466. PathakDR, Whinemore AS. 1992. Combined effects of body size, parity, and Mori M, Harabuchi I, Miyake H, CasagrandeJT, Henderson BE, Ross RK. menstrual events on breast cancer incidence in seven countries. Am J 1988. Reproductive,genetic, and dietary risk factors for ovarian cancer. Epidemiol I 35:153-168. Am J Epidemiol 128:771-777. Peacock SL, White E, Daling JR, Voigt LF, Malone KE. 1999. Relation Mori M, Nishida T, SugiyamaT, Komai K, Yakushiji M, Fukuda K, TanakaT, between obesity and breast cancer in young women. Am J Epidemiol Yokoyama M, Sugimori H. 1998. Anthropometric and other risk factors 149:339-346. for ovarian cancerin a case-controlstudy. Jpn J CancerRes 89:246-253. PetrekJA, PetersM, Cirrincione C, RhodesD, BajorunasD. 1993.Is body fat l,t. Mukherjee P, Sotnikov AV, Mangian HJ, Zhou JR, Visek WJ, Clinton SK. 1999. topography a risk factor for breast cancer?Ann Intern Med I 18:356-362. Energy intake and prostate tumor growth, angiogenesis,and vascular Peyrat JP, Bonneterre J, Hecquet B, Vennin P, Louchez MM, Fournier C, endothelialgrowth factor expression.J Natl CancerInst 9l:512-523. Lefebvre J, Demaille A. 1993. Plasmainsulin-like growth factor-l (IGF- Muscat JE, Wynder EL. 1992. Tobacco, alcohol, asbestos,and occupational l) concentrations in human breast cancer. Eur J Cancer 29A:492497. risk factors for laryngeal cancer.Cancer 69:2244-2251. PotischmanN, SwansonCA, Siiteri P, Hoover RN. 1996. Reversalof relation Muscat JE, Hoffmann D, Wynder EL. 1995. The epidemiology of renal cell between body mass and endogenous estrogen concentrations with carcinoma.A secondlook. Cancer 75:2552-255'1. menopausalstatus. J Natl Cancer lnst 88:756-758. Must A, Willen WC, Dietz WH. 1993. Remote recall of childhood height, Potter JD, McMichael AJ. 1983. Large bowel cancer in women in relation to weight, and body build by elderly subjects.Am J Epidemiol 138:56- reproductive and hormonal factors: A case-controlstudy. J Natl Cancer &. Inst 7l:703-709. Negri E, La Vecchia C, Bruzzi P, Dardanoni G, Decarli A, Palli D, Parazzini Potter JD, McMichael AJ. 1986. Diet and cancer of the colon and rectum: A F, Rosselli del Turco M. 1988. Risk factors for breast cancer: Pooled case-controlstudy. J Natl Cancer Inst 76:557-569. resultsfrom threeItalian case-controlstudies. Am J Epidemiol128:120'1- Preston-MartinS, BernsteinL, Pike MC, MaldonadoAA, HendersonBE. 1987. 1215. Thyroid canceramong young women related to prior thyroid diseaseand Negri E, Ron E, Franceschi S, Dal Maso L, Mark SD, Preston-Martin S, pregnancyhistory. Br J Cancer55:l9l-195. McTiernan A, Kolonel L, Kleinerman R, Land C, Jin F, Wingren G, Preston-Martin S, Jin F, Duda MJ, Mack WJ. 1993. A case-controlstudy of Galanti MR, Hallquist A, Glattre E, Lund E, Levi F, Linos D, Braga C, thyroid cancer in women under age 55 in Shanghai (People'sRepublic La VecchiaC. 1999.A pooled analysisof case-controlstudies of thyroid of China). Cancer CausesControl 4:431-440. cancer.I. Methods. Cancer CausesControl 10:131-142. Purdie D, Green A, Bain C, Siskind Y Ward B, Hacker N, Quinn M, Wright Negri E, FranceschiS, Bosetti C, Levi F, Conti E, Parpinel M, La Vecchia C. G, Russell P, Susil B. 1995. Reproductive and other factors and risk of 2000. Selected micronutrients and oral and pharyngeal cancer. Int J epithelial ovarian cancer: An Australian case-control study. Survey of Cancer 86:122-127. Women's Health Study Group. Int J Cancer 62:678484. Ness RB, Grisso JA, Klapper J, SchlesselmanJJ, Silberzweig S, Vergona R, Purdie DM, GreenAC. 2001. Epidemiology of endometrialcancer. Best Pract Morgan M, Wheeler JE. 2000. Risk of ovarian cancerin relation to estro- Res Clin Obstet Gynaecol 15:341-354. gen and progestin dose and use characteristicsof oral contraceptives. PutnamSD, CerhanJR, ParkerAS, Bianchi GD, WallaceRB, Cantor KP, Lynch SHARE Study Group. Steroid Hormones and Reproductions. Am J CF. 2000. Lifestyle and anthropometric risk factors for prostate cancer in Epidemiol 152:233-241. a cohort of Iowa men. Ann Epidemiol 10:361-369. Neugut AI, Lee WC, Garbowski GC, Waye JD, Forde KA, Treat MR, Fenoglio- Radimer K, Siskind V Bain C, Schofield F. 1993. Relation betweenanthropo- Preiser C. 1991. Obesity and colorectal adenomatouspolyps. J Natl metric indicators and risk of breast cancer among Australian women. Am CancerInst 83:359-361. J Epidemiol 138:77-89. Newman SC, Lees AW, Jenkins HI. 1997. The effect of body mass index and Rauscher GH, Mayne ST, Janerich DT. 2000. Relation between body mass oestrogen receptor level on survival of breast cancer patients. Int J index and lung cancerrisk in men and women never and former smokers. Epidemiol 26:484490. Am J Epidemiol 152:506-513. Ng EH, Gao F, Ji CY, Ho GH, Soo KC. 1997. Risk factors for breast carci- Riman T, Dickman PW, Nilsson S, Correia N, Nordlinder H, MagnussonCM, noma in SingaporeanChinese women: The role of central obesity.Cancer PerssonIR. 2001. Risk factors for epithelial borderline ovarian tumors: 80:725-731. results of a Swedish case-controlstudy. Gynecol Oncol 83:575-585. Niskanen LK, Haffner S, Karhunen LJ, TurpeinenAK, Miettinen H, Uusitupa Ron E, Kleinerman RA, Boice JD Jr, LiVolsi VA, Flannery JT, Fraumeni JF Jr. MI. 1997.Serum leptin in obesity is relatedto genderand body fat topog- 1987. A population-basedcase-control study of thyroid cancer. J Natl raphy but does not predict successful weight loss. Eur J Endocrinol CancerInst 79:l-12. 13'7:6147. Ronchi E, Pizzocaro G, Miodini P, Piva L, Salvioni R, Di Fronzo G. 1984. Noble RL. 1977. T\e development of prostatic adenocarcinomain Nb rats Steroid hormone receptors in normal and malignant human renal tissue: following prolonged sex hormone administration. Cancer Res 37: Relationship with progestin therapy.J Steroid Biochem 2l:329-335. 1929-1933. Ross RK, Shimizu H, Paganini-Hil A, Honda G, HendersonBE. 1987. Case- Nomura A, Heilbrun LK, StemmerrnannGN. 1985. Body mass index as a control studies of prostate cancer in blacks and whites in southern predictor of cancer in men. J Natl Cancer lnst'74.,319-323. California. J Natl Cancer Inst 78:869-874. M6 PARTlll: THECAUSES OF CANCER

Rossing MA, Voigt LF, Wicklund KG, Daling JR. 2000. Reproductive factors Shu XO, Jin F, Dai Q, Shi JR, Potter JD, Brinton LA, Hebert JR, Ruan Z, Gao and risk of papillary thyroid cancer in women. Am J Epidemiol YT, Zheng W. 2001. Association of body size and fat distribution with l5L:765-772. risk of breast cancerurmong Chinese women. Int J Cancer 94:M9455. Rowland ML. 1990. Self-reported weight and height. Am J Clin Nutr 52: Singh PN, FraserGE. 1998.Dietary risk factors for colon cancer in a low-risk tr25-rr33. population. Am J Epidemiol 148:761-774. Russell CD, Petersen RN, Rao SP, Ricci MR, Prasad A,Zhang Y, Brolin RE, Slattery ML, Pofter J, Caan B, Edwards S, CoatesA, Ma KN, Berry TD. 1997. Fried SK. 1998. Leptin expression in adipose tissue from obese humans: Energy balanceand colon cancer-Beyond physical activity. CancerRes Depot-specific regulation by insulin and dexamethasone.Am J Physiol 57:75-80. 275:E507-E515. Slattery ML, Anderson K, Curtin K, Ma K, Schaffer D, Edwards S, Samowitz Russo A, Franceschi S, La Vecchia C, Dal Maso L, Montella M, Conti E, W. 2001a. Lifestyle factors and Ki-ras mutationsin colon cancertumors. Giacosa A, Falcini F, Negn E. 1998. Body size and colorectal-cancer risk. Mutat Res 483:73-81. lnt J Cancer78:161-165. Slattery ML, Potter JD, Curtin K, Edwards S, Ma KN, Anderson K, Schaffer Sanderson M, Williams MA, Malone KE, Stanford JL, Emanuel I, White E, D, Samowitz WS. 2001b. Estrogensreduce and withdrawal of estrogens Daling JR. 1996. Perinatal factors and risk of breast cancer. Epidemiol- increase risk of microsatellite instability-positive colon cancer. Cancer ogy 7:34-37. Res6l:126-130. SarmaAV Jaffe CA, Schottenfeld D, Dunn R, Montie JE, Cooney KA, Wei JT. Slattery ML, Curtin K, Ma K, Edwards S, Schaffer D, Anderson K, Samowitz 2002. Insulin-like growth factor-I, insulin-like growth factor binding W.2002. Diet activity, and lifestyle associationswith p53 mutations in protein-3, and body mass index: Clinical correlates of prostate volume colon tumors. Cancer Epidemiol Biomarkers Prev 1l:541-548. among black men. Urology 59:362-367. Slanery ML, Ballard-Barbash R, Edwards S, Caan BJ, Potter JD. 2003. Body Schapira DV Kumar NB, Lyman GH, Cox CE. 1990. AMominal obesity and mass index and colon cancer:An evaluation of the modifying effects of breast cancer risk. Ann Intern Med lt2:I82-186. estrogen(United States).Cancer CausesControl l4:75-84. Schapira DY Kumar NB, Lyman GH, Cavanagh D, Roberts WS, LaPolla J. SonnenscheinE, Toniolo P, Terry MB, Bruning PF, Kato I, Koenig KL, Shore 1991. Upper-body fat distribution and endometrial cancer risk. JAMA RE. 1999. Body fat distribution and obesiry in pre- and postmenopausal 266:1808-181 l. breast cancer.Int J Epidemiol 28:1026-1031. SchneiderJ, Bradlow HL, Strain G, Levin J, Anderson K. Fishman J. 1983. Spitz MR, Strom SS, Yamamura Y, Troncoso P, Babaian RI, Scardino PT, Effects of obesity on estradiol metabolism: Decreased formation of Wheeler T, Amos CI, von EschenbachA, Kagan J. 2000. Epidemiologic nonuterotropicmetabolites. J Clin Endocrinol Metab 56:973-978. determinants of clinically relevant prostate cancer. lnt J Cancer 89: SchoenRE, TangenCM, Kuller LH, Burke GL, Cushman M, Tracy RP, Dobs 259-264. A, SavagePJ. 1999. Increasedblood glucose and insulin, body size, and Stanin P, SoderbergS, Hallmans G, Bylund A, Kaaks R, StenmanUH, Bergh incident colorectal cancer.J Natl Cancer Inst 9l:1147-1154. A, Olsson T. 2001. Leptin is associatedwith increasedprostate cancer SchoftenfeldD. 1995.Epidemiology of endometrialneoplasia. J Cell Biochem risk: A nested case-referentstudy. J Clin Endocrinol Metab 86:1341- Suppl23:l5l-159. t345. Schulman CC, Ekane S, Zlotta AR. 2001. Nutrition and prostate cancer: StevensJ, Keil JE, Waid LR, Gazes PC. 1990. Accuracy of current, 4-year, Evidence or suspicion?Urology 58:318-334. and 28-year self-reportedbody weight in an elderly population. Am J Schuurman AG, Goldbohm RA, Dorant E, vari den Brandt PA. 2000. Anthro- Epidemiol 132:1156-1163. pometry in relation to prostate cancer risk in the Netherlands Cohort SubarAF, Kipnis V Troiano R, Terry MB, Midthune D, SchoellerD, Bingham Study. Am J Epidemiol 15l:541-549. S, Runswick S, SharbaughC, Ballard-BarbashR, SunshineJ, Schatzkin Segal KR, Landt M, Klein S. 1996. Relationship between insulin sensitivity A. 2003. The Observing Protein and Energy Nutrition (OPEN) Study and plasma leptin concentration in lean and obese men. Diabetes 2002.Am J Epidemiol 158:1-12. 45:988-991. Swanson CA, Brinton LA, Taylor PR, Licitra LM, Ziegler RG, Schairer C. SellersTA, Kushi LH, Potter JD, Kaye SA, Nelson CL, McGovern PG, Folsom 1989. Body size and breastcancer risk assessedin women participating AR. 1992. Effect of family history, body-fat distribution, and reproduc- in the Breast Cancer Detection DemonstrationProject. Am J Epidemiol tive factors on the risk of postmenopausal breast cancer. N Engl J Med 130:1133-l 141. 326:1323-1329. Swanson CA, Potischman N, Wilbanks GD, Twiggs LB, Mortel R, Berman SellersTA, GapsturSM, Potter JD, Kushi LH, Bostick RM, Folsom AR. 1993. ML, Banett RI, Baumgartner RN, Brinton LA. 1993. Relation of Association of body fat distribution and family histories of breast and endometrialcancer risk to past and contemporarybody size and body fat ovarian cancer with risk of postmenopausal breast cancer. Am J Epi- distribution. Cancer Epidemiol B iomarkers Prev 2:321-327. demiol 138:799-803. SwansonCA, CoatesRI, SchoenbergJB, Malone KE, Gammon MD, Stanford Sellers TA, Davis J, Cerhan JR, Vierkant RA, Olson JE, Pankratz VS, Potter JL, Shon IJ, PotischmanNA, Brinton LA. 1996. Body size and breast JD, Folsom AR. 2002. Interaction of waist/hip ratio and family history cancer risk among women under age 45 years. Am J Epidemiol on the risk of hormone receptor-defined breast cancer in a prospective 143:698-706. study of postmenopausal women. Am J Epidemiol 155:225-233. Swerdlow AJ, De Stavola BL, Floderus B, Holm NV, Kaprio J, VerkasaloPK, Senie RT, Rosen PP, Rhodes P, Lesser ML, Kinne DW. 1992. Obesiry at diag- Mack T. 2002. Risk factors for breastcancer at young agesin twins: An nosis of breast carcinoma influences duration of disease-freesurvival. international population-basedstudy. J Natl Cancer Inst 94:1238-1246. Ann Intern Med I 16:26-32. Talamini R, La Vecchia C, Decarli A, Negri E, FranceschiS. 1986. Nutrition, SeversonRK, Grove JS, Nomura AM, StemmerrnannGN. 1988. Body mass social factors and prostatic cancer in a Northem Italian population. Br J and prostatic cancer: a prospectivestudy. BMJ 297:713-715. Cancer53:817-821. Shaneyfelt T, Husein R, Bubley G, Mantzoros CS. 2000. Hormonal predicrors Talamini R, Baron AE, Barra S, Bidoli E, La VecchiaC, Negri E, Serraino D, of pros[atecancer: A meta-analysis.J Clin Oncol l8:847-853. Franceschi S. 1990. A case-control study of risk factor for renal cell SharpeCR, Siemiarycki J. 2001. Joint effects of smoking and body massindex cancerin northernItaly. CancerCauses Control l:125-131. on prostatecancer risk. Epidemiology 12:546-551. Tamimi R, Mucci LA, SpanosE, Lagiou A, BenetouV, Trichopoulos D. 2001. Shinchi K, Kono S, Honjo S, Todoroki I, SakuraiI Imanishi K, Nishikawa H, Testosteroneand oestradiol in relation to tobacco smoking, body mass Ogawa S, Katsurada M, Hirohata T. 1994. Obesity and adenomarous index, energy consumption and nutrient intake among adult men. Eur J polyps of the sigmoid colon. Jpn J Cancer Res 85:479-484. Cancer Prev 10:275-280. Shoff SM, Newcomb PA. 1998. Diabetes,body size, and risk of endometrial TannenbaumA. 1940a.The initiation and growth of tumors: Introduction. I. cancer.Am J Epidemiol 148:234-240. Effects of underfeeding.Am J Cancer 38:335-350. Shoff SM, Newcomb PA, Trentham-Dietz A, Remington PL, Minendorf R, TannenbaumA. 1940b.Relationship of body weight to cancerincidence. Arch GreenbergER, Willett WC. 2000. Early-life physical acriviry and post- Pathol30:509-517. menopausalbreast cancer: effect ofbody size and weight change.Cancer Tao SC, Yu MC, Ross RK, Xiu KW. 1988. Risk factors for breast cancer in Epidemiol Biomarkers Prev 9:59l-595. Chinese women of Beijing. Int J Cancer 42:495498. Shu XO, Gao YT, Yuan JM, Ziegler RG, Brinton LA. 1989. Dietary facrors T"rry P, Baron JA, WeiderpassE, Yuen J, Lichtenstein P, Nyren O. 1999. and epithelial ovarian cancer. Br J Cancer 59:92-96. Lifestyle and endometrial cancer risk: A cohort study from the Swedish Shu XO, Brinton LA, Zheng W, Gao YT, Fan J, Fraumeni JF Jr. 1991. A Twin Registry. Int J Cancer 82:3812. population-basedcase-control study of endometrial cancer in Shanghai, T"rry P, Giovannucci E, Bergkvist L, Holmberg L, Wolk A. 2001. Body weight China. Int J Cancer 49:3843. and colorectal cancerrisk in a cohort of Swedishwomen: Relation varies Shu XO, Brinton LA,Zheng W, SwansonCA, Hatch MC, Gao YT, Fraumeni by age and cancer site. Br J Cancer 85:346-349. IF Jr. 1992. Relation of obesity and body fat distribution ro endomerrial Terry PD, Miller AB, Rohan TE. 2002. Obesity and colorectal cancer risk in cancerin Shanghai,China. CancerRes 52:3865-3870. women.Gut 51:19l-194. v" 't

Obesityand BodyComposition M7

Thune I, Lund E. 1994.Physical activity and the risk of prostateand testicu- of the Swedish population between 1970 and 1997. Eur J Surg 167: lar cancer: A cohort study of 53,000 Norwegian men. Cancer Causes 148-75',7. Control 5:549-556. WeiderpassE, Perssonl, Adami HO, Magnusson C, Lindgren A, Baron JA. Thune I, Brenn T, Lund E, Gaard M. 1997. Physical activity and the risk of 2000. Body size in different periods of life, diabetesmellitus, hyperten- breastcancer. N Engl J Med 336:1269-1275. sion, and risk of postmenopausalendometrial cancer. Sweden. Cancer Toniolo P, Bruning PF,Akhmedkhanov A, Bonfrer JM, Koenig KL, Lukanova CausesControl I I : I 85-192. A, Shore RE, Zeleniuch-JacquotteA. 2000. Serum insulin-like growth Wenten M, Gilliland FD, Baumgartner K, Samet JM. 2002. Associations of factor-I and breastcancer. Int J Cancer88:828-832. weight, weight change,and body masswith breastcancer risk in Hispanic TornbergSA, CarstensenJM. I994. Relationshipbetween Quetelet's index and and non-Hispanic white women. Ann Epidemiol 12:4354. cancer of breast and female genital tract in 4'l ,000 women followed for West DW, Slattery ML, Robison LM, SchumanKL, Ford MH, Mahoney AW 25 years.Br J Cancer69:358-361. Lyon JL, SorensenAW. 1989.Dietary intake and colon cancer: Sex- and Traish AM, Wotiz HH. 1987. Prostaticepidermal growth factor receptorsand anatomicsite-specific associations. Am J Epidemiol 130:883-894. their regulationby androgens.Endocrinology l2l:1461-1467. West DW, Slattery ML, Robison LM, French TK, Mahoney AW. 1991. Adult Trentham-Dietz A, Newcomb PA, Storer BE, Longnecker MP, Baron J, dietary intake and prostatecancer risk in Utah: A case-controlstudy with GreenbergER, Willen WC. 1997. Body size and risk of breast cancer. specialemphasis on aggressivetumors. Cancer CausesControl 2:85- Am J Epidemiol145:1011-1019. 94. Trentham-DietzA, Newcomb PA, Egan KM, Titus-ErnstoffL, Baron JA, Storer Whittemore AS, Kolonel LN, Wu AH, John EM, Gallagher RP, Howe GR, BE, Stampfer M, Willett WC. 2000. Weight change and risk of post- Burch JD, Hankin J, Dreon DM, West DW. 1995. Prostatecancer in rela- menopausal breast cancer (United States). Cancer Causes Control tion to diet, physical activity, and body size in blacks, whites, and Asians ll:533-542. in the United Statesand Canada.J Natl CancerInst 87:652-661. Tretli S. 1989. Height and weight in relation to breast cancer morbidity and WHO Consultation on Obesity. 2000. Obesity: preventing and managing the mortality. A prospectivestudy of 570,000 women in Norway. Int J Cancer global epidemic.Report of a WHO ConsultationWHO TechnicalReports M:23-30. Series,No. 894. Geneva,Switzerland: World Health Organization. Tretli S, Magnus K. 1990. Height and weight in relation to uterine corpus WHO Expert Committee. 1995. Physical status:the use and interpretationof cancer morbidity and mortality. A follow-up study of 570,000 women in anthropometry.Report of a WHO Expert Committee. WHO Technical Norway. Int J Cancer 46:165-172. Reports Series, No. 854. Geneva, Switzerland: World Health Tretli S, Haldorsen T, OttestadL. 1990. The effect of pre-morbid height and Organization. weight on the survival of breastcancer patients. Br J Cancer 62:299-303. Willett WC, Browne ML, Bain C, Lipnick RI, Stampfer MJ, RosnerB, Colditz Tretli S, RobsahmTE. 1999. Height, weight and cancerof the oesophagusand GA, HennekensCH, SpeizerFE. 1985.Relative weight and risk of breast :'.: stomach:A follow-up study in Norway. Eur J Cancer Prev 8:ll5-122. cancer among premenopausalwomen. Am J Epidemiol122:731-740. Turner BC, Haffty BG, NarayananL, Yuan J, Havre PA, Gumbs AA, Kaplan Wingren G, HatschekI Axelson O. 1993.Determinants of papillary cancerof L, Burgaud JL, Carter D, Baserga R, Glazer PM. 1997. Insulin-like the thyroid. Am J Epidemiol 138:482491. growth factor-I receptoroverexpression mediates cellular radioresistance Wolf AM, Colditz GA. 1994. The cost of obesity: The US perspective. and local breast cancer recurrence after lumpectomy and radiation. Pharmacoeconomics5:34-37 . CancerRes 57:3079-3083. Wolf AM, Colditz GA. 1998.Current estimatesof the economic cost of obesity Tymchuk CN, TesslerSB, Aronson WJ, Barnard RI. 1998. Effects of diet and in the United States.Obes Res 6:97-106. exerciseon insulin, sex hormone-binding globulin, and prostate-specific Wolk A, Lindblad P, Adami HO. 1996. Nutrition and renal cell cancer.Cancer antigen. Nutr Cancer 3l:12'1-131. CausesControl 7:5-18. Ursin G, Paganini-Hill A, Siemiatycki J, Thompson WD, Haile RW 1995. Wolk A, Mantzoros CS, Andersson SO, Bergstrtim R, Signorello LB, Lagiou Early adult body weight, body mass index, and premenopausalbilateral P, Adami HO, Trichopoulos D. 1998. Insulin-like growth factor I and breast cancer: Data from a case-controlstudy. Breast Cancer Res Treat prostate cancer risk: A population-based,case-control study. J Natl 33:75-82. CancerInst 90:91l-915. Valaoras VG, MacMahon B, Trichopoulos D, Polychronopoulou A. 1969. WORLD Cancer ResearchFund and American Institute for Cancer Research. Lactation and reproductive histories of breast cancer patients in greater 1997. Food, Nutrition and the Prevention of Cancer: A Global Perspec- Athens, 196547.Int J Cancer4:350-363. tive. Washington,DC: American Instirute for Cancer Research. van Aggel-Leijssen DP, van Baak MA, TenenbaumR, Campfield LA, Saris Wu AH, Wan P, Bernstein L. 2001. A multiethnic population-basedstudy of WH. 1999. Regulation of average 24h human plasma leptin level; the smoking, alcohol and body size and risk of adenocarcinomasof the influence of exerciseand physiological changesin energy balance.Int J stomach and esophagus (United States). Cancer Causes Control ObesRelat Metab Disord 23:l5l-158. 12:721-732. van den Brandt PA, SpiegelmanD, Yaun SS, Adami HO, Beeson L, Folsom Wynder EL, Mabuchi K, Whitmore WF Jr. 1971. Epidemiology of cancer of AR, Fraser G, Goldbohm RA, Graham S, Kushi L, Marshall JR, Miller the prostate. Cancer 28:344-364. AB, Rohan T, Smith-Wamer SA, Speizer FE, Willett WC, Wolk A, Wynder EL, Mabuchi K, Whitmore WF lr. 19'74.Epidemiology of adenocar- Hunter DJ. 2000. Pooled analysisof prospectivecohort studieson height, cinoma of the kidney. J Natl Cancer Inst 53:1619-1634. weight, and breastcancer risk. Am J Epidemiol 152:514-52'7. Xu W Dai Q, Ruan Z, Cheng J, Jin F, Shu X. 2002. Obesity at different ages Vatten LJ. Foss OP. Kvinnsland S. 1991. Overall survival of breast cancer and endometrialcancer risk factors in urban Shanghai,China. Zhonghua patients in relation to preclinically determined total serum cholesterol, Liu Xing Bing Xue ZaZhi 23:347-351. body massindex, height and cigarettesmoking: a population-basedstudy. Yang G, Lu G, Jin F, Dai Q, Best R, Shu XO, Chen JR, Pan XY, ShrubsoleM, Eur J Cancer 27:641-646. Zheng W. 2001. Population-based,case-control study of blood C-peptide Vatten LI, Kvinnsland S. 1992. Prospectivestudy of height, body mass index level and breast cancer risk. Cancer Epidemiol Biomarkers Prev and risk of breastcancer. Acta Oncol 3l:195-200. 10:120'7-l2ll. Vaughan TL, Davis S, Kristal A, Thomas DB. 1995. Obesity, alcohol, and Yee D, Paik S, Lebovic GS, Marcus RR, Favoni RE, Cullen KJ, Lippman ME, tobacco as risk factors for cancersof the esophagusand gastric cardia: Rosen N. 1989.Analysis of insulin-like growth factor I gene expression adenocarcinomaversus squamous cell carcinoma. Cancer Epidemiol in malignancy:Evidence for a paracrinerole in human breastcancer. Mol Biomarkers Prev 4:85-92. Endocrinol3:509-517. VeldhuisJD, Liem AY, South S, WeltmanA, WeltmanJ, Clemmons DA, Abbon Yong LC, Brown CC, SchatzkinA, SchairerC. 1996. Prospectivestudy of rel- R, Mulligan T, JohnsonML, Pincus S. 1995. Differential impact of age, ative weight and risk of breast cancer: the Breast Cancer Detection sex steroid hormones, and obesity on basal versus pulsatile growth DemonstrationProject follow-up study, 1979 to 1987-1989. Am J Epi- hormone secretionin men as assessedin an ultrasensitivechemilumines- demiol 143:985-995. cence assay.J Clin Endocrinol Metab 80:3209-3222. Yoo K, Tajima K, Park S, Kang D, Kim S, Hirose K, Takeuchi T, Miura S. Verreault R, Brisson J, DeschenesL, Naud F. 1989. Body weight and prog- 2001. Postmenopausalobesity as a breastcancer risk factor according to nostic indicatorsin breastcancer. Modifying effect of estrogenreceptors. estrogenand progesteronereceptor status. Japan. Cancer l-ett 167:5743. Am J Epidemiol 129:26V268. Yu H, Harris RE, Wynder EL. 1988. Case-control study of prostate cancer and Villeneuve PJ, JohnsonKC, Kreiger N, Mao Y. 1999.Risk factors for prostate socioeconomicfactors. Prostate 13:317 -325. cancer: results from the Canadian National Enhanced Cancer Surveil- Yu H, Rohan T. 2000. Role of the insulin-like growth factor family in cancer lance System.The Canadian Cancer RegistriesEpidemiology Research developmentand progression.J Natl Cancer Inst 92:1472-1489. Group. CancerCauses Control l0:355-367. Yu MC, Mack TM, Hanisch R, Cicioni C, Henderson BE. 1986. Cigarette Walther C, Zilling T, Perfekt R, Mgller T. 2001. Increasingprevalence of ade- smoking, obesiry, diuretic use, and coffee consumption as risk factors for nocarcinomaof the oesophagusand gastro-oesophagealjunction: A study renal cell carcinoma.J Natl Cancer Inst 77:351-356. 48 PARTlll: THECAUSES OF CANCER

Yuan JM, castelao JE, Gago-DominguezM, Ross RK, yu MC. 199g. Hyper- Zhang ZF, Ktntz RC, Sun M, Karpeh M Jr, yu Gp, Gargon N, Fein JS, tension,obesity and their medicationsin relation to renal cell carcrnoma- GeorgopoulossK, Harlap S. 1996. Adenocarcino.ur Br J Cancer77:1508-1513. of the esophagus and gastric cardia: Medical conditions, tobacco, alcohol, and socioelo- Zeleniuch-JacquorreA, AkhmedkhanovA, Kato I, Koenig KL, shore RE, Kim nomic factors.Cancer Epidemiol Biomarkersprev 5:761_76g. MY, Levitz M, Mittal KR, Raju U, Banerjee S, Toniolo p. 2001. post_ Ziegler RG, Hoover RN, Nomura AM, Wesr DW, Wu AH, pike MC, Lake AJ, menopausal endogenous oestrogens and risk of endometrial cancer: Horn-Ross PL, Kolonel LN, Siiteri pK, Fraumeni JF Jr. 1996. Relative Resultsof a prospectivesrudy. Br J Cancer g4:975_9g1. weight,weighr change. height, and breastcancer risk in Asian-American zhang S, Folsom AR, Sellers TA, Kushi potter LH, JD. 1995. Berrer breasr women. J Natl Cancer Inst 88:650-660. cancer survival for postmenopausalwomen who are less overweight and eat less fat. The lowa women's Health Study. cancer l6:2is-2g3.