jurnal ca pnkreas

7/28/2019 Jurnal CA Pnkreas

1/8

Dietary Habits and Past Medical History as Related to FatalPancreas Cancer Risk Among Adventists

PAUL K. MILLS, PHD, W. LAWRENCE BEESON, MSPH, DAVID E. ABBEY, PHD,GARY E. FRASER, MD, PHD, AND ROLAND L. PHILLIPS, MD, DRPHt

Epidemiologic studies of diet and pancreas cancer are few, and include ecologic comparisons and alimited number of prospective and case-control studies. Foods and/or nutrients that have been sug-gested to be associated with increased risk of this cancer include total fat intake, eggs, animal protein,sugar, meat, coffee and butter. Consumption of raw fruits and vegetables has been consistently asso-ciated with decreased risk. Dietary habits and medical history variables were evaluated inaprospectivestudy of fatal pancreas cancer among34,000California Seventh-day Adventists between1976and1983.Forty deaths from pancreas cancer occurred during the follow-up period. Compared to all US whites,Adventists experienced decreased risk from pancreas cancer death (standardized mortality ratio [SMR]=72for men; 90 for women), which was not statistically significant. Although there wasasuggestiverelationship between increasing meat, egg, and coffee consumption and increased pancreatic cancer risk,these variables were not significantly related to risk after controlling for cigarette smoking. However,increasing consumption of vegetarian protein products, beans, lentils, and peasaswell asdried fruitwasassociated with highly significant protective relationships to pancreas cancer risk. A prior history ofdiabetes was associated with increased risk of subsequent fatal pancreas cancer, as was ahistory ofsurgery for peptic or duodenal ulcer. A history of tonsillectomywasassociated withaslight, nonsignifi-cant protective relationship as was history of various allergic reactions. These findings suggest that theprotective relationships associated with frequent consumption of vegetables and fruits high in protease-inhibitor content are more important than any increase in pancreas cancer risk attendent on frequentconsumption of meator other animal products. Furthermore, the previously reported positive associa-tions between diabetes and abdominal surgery and pancreas cancer risk are supported in these data.Cancer61:2578-2585, 1988.

ACH Y EAR in theUS, more than 24,000people dieE f pancreas cancer, making this the fifth leadingcauseof cancer death.' There is some evidence that theincidence of the disease has increased in the last 30years2 whereas survivorship remains very Due tothe extremely poor survival, incidenceis tantamount tomortality. Little is known concerning the cause of thedisease. Like other cancers of environmental cause, pan-creas cancer is prevalent in the Western industrializednations and shows a sharp increase with age beginningin the seventh decade of life. Men experience an increase

$ Deceased March 8, 1987.From the Department of Preventive Medicine, School of M edicine.Lorna Linda University, Lorna L inda, California.Supported by grant RO 1-CA 14703 from the National Cancer Insti-tute (R.L .P.).Address for reprints: Paul K. Mills, PhD, Department of PreventiveMedicine, School of Medicine, L orna Linda University, L oma Linda,CA 92350.Accepted for publication November 30, 1987.

in risk in comparison to women and, in theUS,blacksappear to be at higher risk than whites. The only estab-lished risk factor besides age, race and sex is cigarettesmoking. Although the pancreas is intimately involvedin the digestive process, relatively few studies have ad-dressed the role of diet in the cause of pancreas cancer.This isundoubtedly becauseof the extremely poor sur-vivorship associated with the disease which makes inter-viewing patients for retrospective studies problematic.Epidemiologic investigations of dietary factors thatmay relate to the cause of pancreas cancer include eco-logic corn par is on ^^^ and a limited number of prospec-tive6 and case-control studies.7-' Foods and/or nu-trients that have been suggested to be associated withincreased risk from pancreas cancer incidence or mor-tality include total fat intake,4 eggs, animal protein,sugar,5 meat,6,' ~offee,~,~-'beef, pork,"," and b~tter .~Foods and/or nutrients consistently associated with de-creased risk of pancreas cancer include raw fruits andvegetables?-" To date, however, the data are sparse, insome instances inconsistent (e.g., the coffee-pancreas

2578


2/8

No. 12 DIET,MEDICAL ISTORY,ND PANCREASANCER ISK - Mlls et al. 2579cancer relationship) and this field of inquiry should stillbe considered embryonic.Similarly, prior medical history has been investigatedin regard to pancreas cancer risk. DiabetesI2 and gas-trectomy'' appear to be associated with increased risk,while a history of tonsillectomy9 has been associatedwith decreased risk. In one study, pancreas cancer pa-tients were noted to have fewer allergies of any kind."Seventh-day Adventists (SDA) provide a unique op-portunity to study dietary relationships to pancreascancer. By church proscription Adventists do notsmoke, drink alcohol, or eat pork, and approximately50% follow a lacto-ovo-vegetarian diet. Therefore, notonly is the potentially confounding effect of cigarettesmoke nearly absent but there is considerable variationin exposure to meat and other dietary habits within thispopulation which is not found in other populations.Among Adventist vegetarians, many substitute soyproducts, legumes and peas for meat. These vegetablesare known to contain large amounts of protease inhibi-tors, which are thought to be protective against cancer atvarious sites. 3 However, some animal studies indicatethat these inhibitors also may serve as promoters of pan-creas cancer.I4 We, therefore, evaluated dietary andmedical history relationships to fatal pancreas cancerrisk in a cohort study involving 34,000 California Sev-enth-day Adventists who completed a detailed LifestyleQuestionnaire in 1976 and who were then followed forcancer incidence and mortality between return of thequestionnaire and the end of 1982.

MethodsStudy Population Identification and Enumeration

The Adventist Health Study (AHS) began preliminaryidentification of California Adventists in late 1973 byobtaining church directories from each of the 437 Cali-fornia SDA churches. A comprehensive list of namesand addresses of all California SDA households (N=63,530) was computerized in mid-1974 and served asthe first mailing list for the AHS. In August 1974, ademographic Census Questionnaire was sent via bulkrate mail to every household identified on a church di-rectory.The Census Questionnaire was designed to be com-pleted by one member of the household (usually thehead of the household) on behalf of the entire house-hold. Therefore, this person recorded demographic andother data on himself as well as on all other familymemberssothat, to a certain degree, information on theCensus Questionnaire was obtained by proxy.The Census Questionnaire was returned by 36,805households (58%) isting 95,196 persons of all ages andraces. One reason that the response rate was not higherwas that many households on our mailing ist should not

have been included because every member of thesehouseholds was no longer an active member of the spe-cific church and had not yet been removed from theofficial church directories. Persons under25 years of agewere removed from further analyses at this point (N=36,106) resulting in a study population of 59,090.In August 1976, a Lifestyle Questionnaire was mailedfirst class to every living member (N = 57,841) of thetotal study population and was returned by 40,398 indi-viduals (response rate, 69.8%). Among the46,031 livingnon-Hispanic whites who were sent the Lifestyle Ques-tionnaire, 34,556 (response rate, 75.1%) returned acompleted form. At this point, it was decided to dividethe study population as follows:

1. Non-Hispanic whites (25 years of age or older atthe time of Census Questionnaire completion and whoalsocompleted the Lifestyle Questionnaire) on whom allcausesof mortality (1974 to 1982) and the incidence ofcancer (1976 to 1982) would be monitored. This popu-lation (N = 34,198) became the incidence population,because cancer incidence determination would be com-pleted only on this group. Exposure information on in-dividuals within this population is available from boththe Census and the Lifestyle Questionnaire. Threehundred fifty-eight non-Hispanic whites were not in-cluded in the incidence population because of data pro-cessing errors that were not corrected until the caseascertainment phase of the study had been completed.The incidence population is further divided into twosubgroups-3 1,208 who are baptized into the church(Adventists) and 2990 who are considered non-SDA(primarily nonbaptized spouses of Adventists).2. All others in the study population (N =24,892).Exposure for this population is limited, for the mostpart, to the Census Questionnaire.

This report is concerned with fatal pancreas cancer(International Classification of Disease [ICD] = 157)which occurred in the incidence population,1976-1 982.Mortality Ascertainment i n the Study Population

Fatal outcomes in this population were detected byvarious mechanisms which overlapped with each otherbut which would provide for the most comprehensivecoverage of mortality during the entire period of fol-lowup(1974 to 1982). Three primary mechanisms wereused.1. Computerized record linkage with the Californiadeath certificate files.2. Computerized record linkage with the NationalDeath Index.3. Manual linkage with SDA church records.The last mechanism is probably least productive, in


3/8

2580 CANCERune15 1988 Vol. 61TABLE. Observed and Expected* Number of Pancreas Cancer Deaths in the Adventist Population, 1976-1982, by Age and Sex

Men WomenObserved Expected Observed ExpectedPerson no. of no. of Person no. of no. ofAge group years deaths deaths Obslexp years deaths deaths Obs/exp

4 4 37,6 I6 0 1.68 0.00 50,452 1 1.18 0.8555-59 8,592 1 I .69 0.59 11,682 I 1.so 0.6760-64 8,164 2 2.83 0.7 11,643 6 2.48 2.4265-69 7,620 0 2.90 0.00 11,779 3 3.48 0.8670-74 6,284 3 4.35 0.69 10,723 4 4.56 0.8875-79 4,4 I7 4 4.16 0.96 8,394 3 4.77 0.6380-84 2,878 3 2.91 I .03 6,094 2 4.1 I 0.4985-89 I ,581 1 1.60 0.62 3,44 2 2.39 0.8490-94 531 3 0.54 5.58 1,314 1 0.9 1.10295 77 0 0.08 0.00 203 0 0.14 0.00Total 77,760 17 23.74 0.72 115,723 23 25.52 0.90*Expected numbers are based upon age-specific mortality rates forU.S.whites, 1975-84.

that church records are known to be incomplete. How-ever, at beginning of follow-up, all SDA churches inCalifornia were requested to notify AHS whenever achurch member died. Church clerks supplied the AHSwith basic information on the decedent (particularlydate of death and place of death), which enabled theAHS to request death certificates from the State. Theprimary mechanisms used to monitor mortality in ourtotal study population was record linkage with the statedeath files and with the National Death Index.

Statistical AnalysisAge- and sex-specific mortality from pancreas canceramong white, non-Hispanic California Adventists wascompared with all US whites by applying age- and sex-specific mortality rates for US whites(1975 to 1984) toperson-years at risk in the Adventist cohort, using theprogram of Monson.'' Members of the study populationcontributed person years at risk beginning at the timethey returned the Lifestyle Questionnaire and endingDecember 31, 1982 or at the time of death. Age- andsex-adjusted incidence ratios of pancreas cancer werecalculated by maximum likelihood estimates by treatingthe outcome (fatal pancreas cancer) as a Poisson-distrib-uted variable and conducting Poisson regression forgrouped data using the GLIM statistical package soft-ware.16 Multivariate analyses were conducted by con-structing Cox proportional hazards regression models toobtain adjusted estimates of hazard ratios. Program 2Lin the BMDP package was used for these ana1y~es.l~Interactions in the data were evaluated by including allfirst-order cross-product terms in the initial Cox modelsand examining the change in the log-likelihood on re-moval of all interaction terms as a group.18

ResultsBetween the return of the Lifestyle Questionnaire and

the end of 1982 there were 40 deaths from pancreascancer in the Incidence Population cohort (17 men, 23women), of a total of 2776 deaths.The standardized mortality ratio for pancreas canceris 0.72 for Adventist men and 0.90 for Adventistwomen. Neither result is statistically significant (Table1).Current use of meat, poultry or fish is associated withincreasing risk (Table 2), which is of borderline signifi-

cance. Past use of these flesh foods is associated with anonsignificant increase in the point estimates. Similarly,current use of eggs is associated with increased risk.Current or past use of milk or cheese is not associatedwith significant elevation or depression of the point esti-mates. Conversely, among the fruits, vegetables and veg-etarian protein products examined, increasing con-sumption of the vegetable protein products, beans, len-tils, or peas and dried fruits are all associated withsignificant protective relationships to fatal pancreaticcancer risk. Also, use of tomatoes shows a suggestive,though nonsignificant, protective relationship to risk, asdoes the consumption of various types of fresh fruit(Table3).Use of soft margarine is associated with a nonsignifi-cant elevation in the relative risk, although increasinguse of this type of margarine on bread does not appear tobe associated with increased risk (Table 3). The use ofbutter on bread also bears little relationship to fatal pan-creatic cancer risk.Current coffee consumption bears a suggestive,though nonsignificant relationship to risk in these data,although past use is associated with decreased point es-timates. Current cigarette smoking is associated with in-


4/8

No. 12 DIET,MEDICAL ISTORY,ND PANCREASANCERISK Mlls et al. 2581TABLE. Age- and Sex-Adjusted RR forFatal Pancreas CancerAmong Adventists, Age25Years and Older, by SelectedDietary Variables,1976-1982

Frequency Age- andDietary of useor sex-adjusted No. of deaths1variable category RR (95%CI) person years*~

Current useof meat,poultryorfishCurrent useof eggs

VegetarianproteinproductsBeans, lentils,peasRaisins, dates,other driedfruits


5/8

2582 CANCERune 15 1988 Vot. 61Since Adventists who avoid meats often substitute

vegetarian protein products in their diet, in the secondseries of Cox models we included terms for current meatconsumption in addition to age, sex, and smoking.Again, none of the interaction terms proved significant.However, increasing consumption of the vegetarianprotein products was seen to be associated with a highlyprotective relationship to risk as was the use of beans,lentils, or peas as well as dried fruit consumption. Allthree of these items were highly statistically significantin a multivariate setting. However, if current meat,poultry, or fish is considered simultaneously with thevegetarian products, beans, and dried fruit, the excessrisk previously observed disappears (RR = 1.00). In-spection of log-log plots did not reveal any violation ofthe proportional hazards assumption.

DiscussionSeveral limitations in the nature of the study popula-

tion, the quality of the exposure data (particularly di-

TABLE . Age- and Sex-Adjusted RR for Fatal Pancreas CancerAmong A dventists, Age 25Y ears and Older, by MedicalHistorv and Al lersv Variables. 1976-1982Age- andSex-Adjusted No. of deaths/Variable Category RR (95%C.I .) person years

Ulcer of thestomach orduodenum

Peptic ulcer surgeryDiabetesHx of tonsillectomyHx ofHx of asthma

appendectomy

Hx of hay fever

Hx of Allergy*H x of reaction topoison ivy (oroak)or otherplantsHx of reaction tobee stingHx of reaction tomedicationHx of reaction tochemicals

NoY esNoY esNoY esNoY esNoY esNoY esNoY esN oY es

NoY esNoY esNoY es

1 oo1.86 (0.84-4.13)1.oo3.35 (1.27-8.86)1 oo3.76 (1.70-8.31)1.oo0.70 (0.39-1.41)1 oo1.74 (0.86-3.53)1 .oo0.87 (0.2 1-3.63)I .oo0.66 (0.23-1.87)1 oo0.99 (0.44-2.27)1 oo0.43 (0.06-3.16)I .oo0.67 (0.07-5.98)1 oo1.53 (0.37-6.30)

261 162,254812 1,748

271 173,218516104281173,67881968315165,9071711 19,03916/118,11318/66,159351 174,3032112,308331152,1074134,508301149,3417131,273

3611 76,3341110,279291153,6138132,999351179,252217363

RR: relative risk; CI: confidence interval; Hx: history.* Allergic reaction sufficiently severe to warrant the attention of amedical doctor.

TABLE. Relative Risks of Fatal Pancreas Cancer AmongAdventists for Selected Exposures as Derived From CoxsProportional Hazards Regression ModelsMultivariate-adjustedExposure (highllow) predicted RR 95%CI

Model I *Current meat,poultry or fish 2.26 0.72-7.12Current eggs 3.42 0.72- 16.26Current coffee 2.21 0.61-7.99Model 2tVegetarian protein

Beans, lentils andRaisins, dates, dryCurrent use of meat

products 0.15 0.03-0.890OO3-0.24eas 0.03

fruit 0.19 0.04-0.86poultry or fish 1 oo 0.28-3.6 I

* Variables simultaneously included in the proportional hazardsmodel include age, sex, smoking status, and the primary exposurevariable of interest ( ie.,meat, eggsor coffee consumption).tVariables simultaneously included i n the proporti onal hazardsmodel include age, sex, smoking status, meat, poultry, or fish and theprimary exposure variable of interest ( i x,egetarian products, beans,or dried fruit).etary), outcome data, and interpretation of analytic pro-cedures must be discussed.First, as mentioned, only 58% of all Adventist house-holds initially contacted by the Adventist Health Studyreturned the Census Questionnaire. This rather low re-sponse rate may reflect a certain selection bias in regardto those individuals under study in this particular inves-tigation of pancreas cancer. However, this low responserate was attributable to poor response rates among non-whites and Hispanics in the total study population(average response rates among these ethnic groups wasapproximately 35%)and for this reason, the entire studypopulation was divided into non-Hispanic whites andall others. The response rate to the Census Question-naire mailing among non-Hispanic whites was in excessof 75%,which lessens the likelihood that severe selectionbias influences the results reported here since the analy-sis was restricted to non-Hispanic whites in the studypopulation.Secondly, the quality of the dietary data should beaddressed. The original dietary data were gathered byafood frequency questionnaire which was self-adminis-tered in 1976. In 1984, a sample(N =623)of the non-Hispanic white componentof the total study populationwas recontacted and asked to recall the frequency ofconsumption of 35 of the foods initially reported in1976. Also current (1984) habits were elicited. Analysesfrom this substudy indicate, in general, a high degree ofcorrelation (most correlation coefficients2 0.6) betweenthese 35 foods as initially recorded and as recalled 8


6/8

No. 12 DIET,MEDICAL ISTORY,ND PANCREASANCER ISK - Mills et al. 2583years later (personal communication, K. Linsted) indi-cating a reasonable degree of validity in the dietary dataused in this analysis.A third concern is reliance on death certificates fordeaths presumably due to pancreas cancer. Previouscase-control studies,10 have noted that the percentageof pathologically confirmed cases of this disease is ap-proximately60%to 70%. Indeed the histologic confir-mation rate for pancreas cancer in the Third NationalCancer Survey was only 70%.Yet a review of deathcertificates from the Third National Cancer Survey withan underlying cause of pancreatic cancer death com-pared with hospital diagnoses for2500cases of pancreascancer showed that the confirmation rate for pancreascancer was 89%,45 ndicating that death certificate re-porting is consistent with in-hospital diagnoses, al-though those diagnoses may be based on clinical or sur-gical observation rather than histologically confirmedstudies. A review of the medical records of 40pancreascancer cases reported upon here which was done as partof the incidence portion of the AHS indicates that 28(70%)of the cases were histopathologically confirmed.The other cases for which information is available weredetected by autopsy (one case), cytology (one case),grossobservation (one case), clinical impression only (onecase) and other means (eight cases).The small number of cases available for analysis inthis study and the number of analytic comparisons thatwere completed also deserve comment. When dealingwith uncommon outcomes such as pancreas cancer,there is always insufficient material for in-depth analysiseven when, as in this case, almost200,000person yearsat risk have been accrued after 6 years of follow-up.Since little is known concerning the cause of pancreascancer, only age, sex, and cigarette smoking were con-sidered as covariates n most of the analyses. Each of theother exposure variables considered were identified onana-prioribasis after carefully reviewing both the exper-imental literature, as well as the few epidemiologica n-vestigations in man. No aposteriori comparisons weremade. However, in view of the multiple comparisonspresented, the following results should be interpretedcautiously.In this study the use of animal products, includingmeat, poultry, and fish as wel as eggs was associatedwith suggestive, although nonsignificant ncreases n riskof fatal pancreatic cancer. Substantial protective associa-tions, however, were noted between the use of meatsubstitutes(ie.,vegetarian protein products such as glu-ten, soy, or nuts) and pancreas cancer risk. Similarly,increasing consumption of beans, lentils, and peas aswel as dried fruit (e.g.,raisins and dates) showed signifi-cant protective relationships to risk. In previous ecologiccomparisons, total fat intake has been associated with

increased pancreas cancer risk,4 as have eggs, animalprotein, and sugar.5 n another cohort study of diet andpancreas cancer: daily intake of meat also was related toa higher risk of cancer of the pancreas. Recent case-control studies of diet and pancreatic cancer have notedincreased risks to be associated with frequent consump-tion of beef and butter and, in particular, with fre-quent use of fried or grilled meat. We noted no associa-tion between the use of margarine or butter and pan-creas cancer risk.In animal models, dietary unsaturated fats have beenshown to enhance pancreatic carcinogenesis during thepostinitiation stage.14 Similarly, raw soya flour enhancesthe growth of carcinogen-induced pancreatic tumors,more so than high levels of dietary unsaturated fat.I4Although the action of soya flour seems to operate viathe tripsin inhibitors contained in raw soy, the action ofthe inhibitors seems to be destroyed by heating. Also,the raw soya flour effect seems to be species-specific, andexperiments in higher primates indicate that raw soyaflour does not act as a promoter of pancreatic carcino-genesis. Vegetarian Adventists consume copious quan-tities of soy products (in place of meat), and our atten-tion was drawn to the possible deleterious effect thatthese substances might have on human health. In thesedata, however, the consumption of vegetable proteinproducts, including soy products, was associated withdecreased risk from pancreas cancer. This may be due tothe protease inhibitors found in soy as wel as in beans,lentils, and peas, which also were associated with de-creased risk. We conclude that the deleterious effect ofeating soybeans observed in animal studies is directlyrelated to species specificity, because rats, chicks, miceand guinea pigs are affected by the ingestion of rawsoybean mealz0but dogs, swine, and calves are not. Onestudy noted that although soy-based protein diets con-taining trypsin inhibitor caused pancreatic hypertrophyin the rat, there was no such effect in primates even after5 years of feeding such a diet.2These findings, and ours,are consistent with previous investigations of the poten-tial for trypsin inhibitors to adversely affect the humanpancreasz2

In this study there was a suggestive, though nonsignifi-cant, inverse relationship between the consumption offresh citrus fruit, fresh winter fruit (apples, bananas,pears), and pancreas cancer risk. Previous case-controlstudies of pancreas cancer noted similar relationshipswith fresh fruit consumption,-11 suggesting that theascorbic acid content of citrus fruit, in particular, mayplay a protective role in pancreatic cancer. However, wenoted no protective relationship between daily con-sumptionofcooked green vegetables or green salads andpancreas cancer risk.Daily consumption of regular coffee showed a sugges-


7/8

2584 CANCERune15 1988 Vol. 61tive, although nonsignificant, relationship to pancreascancer risk. Past consumption, however, showed a sug-gestive protective relationship. Numerous other studieshave examined this relationship with few studies sup-porting a role for coffee consumption in the etiology ofpancreas ~an cer , ~,~~nd most not supporting such a

Among Adventists, coffee drinking may represent anadherence to an entire constellation of factors inconsis-tent with church teachings or recommendations (such asmeat and cigarette use), which may be more meaningfulin relation to pancreas cancer risk. Current use of ciga-rettes in this study is associated with highly significantrisk estimates for pancreatic cancer risk which is consis-tent with several previous investigation ^ ^ however,only 3.7% of the incidence population reported currentsmoking, and 56.4% of these smokers are non-SDA.We evaluated prior medical and surgical history sinceprevious work in pancreas cancer suggests that diabetesmay increase riskI2 as does prior abdominal ~urgery,~','~whereas tonsillectomy' and allergies" may be associatedwith decreased risk of this disease. In this prospectivestudy, in which subjects were free of pancreas cancerwhen they completed the Lifestyle Questionnaire in1976, self-reported diabetes is associated with a signifi-cant elevation in risk for subsequent pancreas cancer.This estimate(RR =3.76) is as large as that observed inearlier st~di es'~,~~,~'nd is highly significant. Similarly,gastric surgery and, in particular, previous gastrectomyare associated with increased risk of pancreas cancer. Inthis study, a history of surgery for gastric or duodenalulcer was associated with a significantly elevated riskeven after controlling for cigarette smoking. History ofappendectomy also was associated with an increasedrisk, although this finding was not statistically signifi-cant. Some workers have suggested that the gastric hy-poacidity after gastric surgery results in bacterial over-growth, resulting in production of circulating carcino-gens, in particular, N-nitroso corn pound^ , ^Conversely, a history of tonsillectomy recently has beenshown to be associated with a decreased risk of pancreascancer' in a case-control study using both hospital andpopulation controls. In this study, a history of tonsillec-tomy was associated with a nonsignificant reduction inrisk. Other investigations have noted this protective re-l ati~nshi p,~~.~~nd it has been noted that the tonsils area source of antigenic stimulus.Finally, we examined allergies in relation to subse-quent cancer risk. With the exception of a history ofreaction to chemicals, each allergic disorder examinedwas associated with decreased risk estimates for pancre-atic cancer, although none were statistically differentfrom the null value. Allergy on the questionnaire used inthis investigation referred specifically to those allergic

ro]e9.10,24-32

reactions severe enough to warrant the attention of amedical doctor.We considered the possibility of multicollinearity inthese data, because Adventists who choose to digressfrom church teaching by consuming meat or other ani-mal products may also smoke and, conversely, thosewith close adherence to church teachings would con-sume vegetarian protein products rather than meat andwould not smoke. Meat and other animal product con-sumption would, therefore, be positively correlated withsmoking and negatively correlated with the consump-tion of vegetarian protein products, beans, and driedfruit. Multicollinearity was evaluated using variousmechanisms. Bivariate correlation coefficients were cal-culated between all exposure variables. None exceeded0.4in absolute value. Next, stepwise regression was donein order to calculate the multiple R2 values (ie., thepercentage of the variance in one variable explained bythe remaining variables) associated with each of the ex-posure variables. Again the magnitude of theR2valuesnever exceeded0.5. This implies that variance inflationfactors were within acceptable limits. Therefore, al-though there was reason to suspect that severe multicol-linearity may be present in the data on anapriori basis,in fact this did not seem to beso.This prospective study of diet and subsequent fatalpancreas cancer risk among Adventists indicates thatthe strength of the protective association between theregular consumption of vegetarian protein products (in-cluding soy products) as well as beans, lentils, and peasappears to be stronger than the increase in risk asso-ciated with consuming meat or other animal productspreviously re~orted.~.~~hese results, based on prospec-tively gathered data in a population characterized byunique dietary characteristics, are not hindered by valid-ity concerns commonly encountered in case-controltypesof investigations and should be considered in fu-ture studies of chemoprevention of human cancer.

REFERENCES1. Silverberg E, Lubera J . Cancer statistics, 1986. Ca 1986;2. DeVesaS, Silverman D. Cancer incidence and mortality trendsin the United States, 1935-74. J Natl Cancer Inst 1978; 60545-571.3. Axtell L ,AsireA , Myers M, eds. Cancer Patient Survival. ReportNo. 5. Washington D.C.: US. Gov't Printing Otfice, 1977. (DHEWPublication no. (NIH)77-992).4. Segi M et al.Cancer mortality of selected sitesin 24 countries,No. 5 1964-65, Sendai: Japanese Deptof Public Health, Tohokii Uni-versity School of Medicine, 1969.5. Armstrong B, Doll R. Environmental factors and cancer inci-dence and mortality in different countries with special reference todietary practices. Int J Cancer 1975; 15617-631.6. HirayamaT. Smoking in relation to death ratesof 265,118 menand women in Japan. (1972): A report on five years of follow-up.Presented at the American Cancer Society's 14th Cancer Seminar,Clearwater Beach, Florida, 1972.

36( 1):9-25.


8/8

No. 12 DIET,MEDICAL ISTORY,ND PANCREAS CA NCER RISK - Millsel d. 25857. MacMahon B, Y en S, Trichopoulos D, Warren K , Nardi G.Coffee and cancer of the pancreas. N Engl J Med 1981; 304:630-633.8. Durbec J P, Chevillotte G, B idart J M , Berthezene P, Sarles H.Diet, tobacco and risk of cancer of the pancreas: A case-control study.Br J Cancer 1983; 47:463-470.9. Gold EB, Gordis L, Diener M D et al. Diet and other risk factorsfor cancer of the pancreas. Cancer 1985; 55:460-467.10. Mack TM , Y u MC, Hanisch R, Henderson B. Pancreas cancerand smoking, beverage consumption, and past medical history. J NatlCancer Inst 1986; 76:49-60.11. Norell SE, Ahlbom A, Erwald R et a/. Diet and pancreaticcancer: A case-control study. AmJ Epidemiol 1986; 124:894-902.12. Kessler I. Cancer mortality among diabetics. J Natl Cancer Inst13. Troll W , Frenkel K , Wiesner R . Protease inhibitors: Their roleas modifiers of carcinogenic processes. I n: Friedman M ., ed. Nutri-tional and Toxicological Significance of Enzyme Inhibitors in Foods.New York: Plenum Publishing Corp, 1986; 153-65.14. Roebuck B, K apli ta P, Edwards B, Praissman M. Effects ofdietary fats and soybean protein of Azaserine-induced pancreatic carci-nogenesis and plasma cholecystokinin in the rat. Cancer Res 1987;15. M onson RR . A nalysis of relative survival and proportionalmortality. ComputBiomed Res 1974; 7:325-332.16. Baker R, Nelder J . The GL IM System: Release 3. 77. UsersGuide. Oxford: Numerical A lgorithms Group, 1985; 45-48.17. Hopkins A. B M DP Statistical Software Manual. Berkeley, CA :University of California Press, 1985; 576-594.18. K leinbaum D, K upper L, Morgenstern H . Epidemiologic Re-search: Principles and Quantitative Methods. Belmont, CA : LifetimeLearning Publications, 1982; 447-456.19. McGuinness E, Morgan R, L evison D, Hopwood D,WormsleyK . Interaction of asaserine and raw soya flour on the rat pancreas.Scand J Gastroenterol 1981; 16:49-56.20. Schneeman B, Gallaher D. Pancreatic response to dietary tryp-sin inhibitor: V ariations among species. Fed Proc 44(5): 1496 (Abstract# 6346), 1985.21. Harwood J , Ausman L, Ki ng N et al.Effect of long term feedingof soy-based diets on the pancreas of Cebus monkeys. Fed Proc44(5): 1496 (Abstract#6349), 1985.22. L iener I . Significance for humans of biologically active factorsin soybeans and other legumes. J A m Oil Chemist Society 1979;23. Nomura A, Stemmerman G, Heilbrun L . Coffee and pancreatic24. Severson R, Davis S, Pollisar L . Smoking, coffee and cancer of25. Wynder E, Hall N , Polansky. Epidemiology of coffee and cancer26. Goldstein N . No association found between coffee and cancer of27. Jick H, Dinan B. Coffee and pancreatic cancer (Letter). Lancet

1970; 44~673-686.

47: 1333-1338.

56: 121-129.cancer (Letter). Lancet 1981; 2:415.the pancreas (Letter). Br Med J 1982; 285:214.of the pancreas. Cancer Res 1983; 43:3900-3906.the pancreas (Letter). N Engl J M ed 1982; 306:997.1981; 2:92.

28. Kessler I. Coffee and cancer of the pancreas (Letter). N Engl JMed 1981; 304:1605.29. K inlen L, McPherson K. Pancreas cancer and coffee and teaconsumption: A case-control study. Br J Cancer 1984; 49:93-96.30. Whittemore A, Paffenbarger R, Anderson K, Halpern J . Earlyprecursors of pancreatic cancer in college men. J Chronic Dis 1983;31. Heuch I, Kvale G, J acobsen B. Useof alcohol, tobacco andcoffee and risk of pancreatic cancer. Br J Cancer 1983; 48537-643.32. Snowdon DA, Phillips RL . Coffee consumption and risk of fatalcancers. Am Public Health 1984; 74:820-823.33. Best EW . A Canadian study of smoking and health. Ottawa:Department of National Health and Welfare, 1966; 65-86.34. Cederlof R, Friberg L, Hrubec Z, Lorich U. The relationship ofsmoking and some social co-variables to mortality and cancer morbid-ity: A ten year follow-up in a probability sample of 55,000Swedishsubjects, age 18-69, part 111. Stockholm: Karolinska Institute, 1975.35. Doll R, Pet0 R. M ortality in relation to smoking: 20 years'observation on male B ritish doctors. Br Med J 1976; 2:1525-1536.36. Hammond EC. Smoking i n relation to the death rates of onemill ion men and women. In: Haenszel W, ed. Epidemiological Ap-proaches to the Study of Cancer and Other Chronic Diseases. USPubli c Health Service Natl Cancer Inst Monograph 19, 1966; 126-204.37. Hirayama T.Changing patterns of cancer in J apan with specialreference to the decrease in stomach cancer mortality. In: Hiatt HH,Watson JD, Winston J A, eds. Origins of Human Cancer, vol. 4. ColdSpring Harbor, New Y ork Cold Spring Harbor Laboratory 1977;55-75.38. K ahn HA . Report of 8-'h years of observation. In: Haenszel W,ed. Epidemiological Approaches to the Study of Cancer and OtherChronic Diseases. US Publi c Health Service. National Cancer InstituteMonograph 19, 1966; 1-125.39. Caygill C, Hill M , Hall N, K irkham J , Northfield T. Gastricsurgery as a risk factor in human carcinogenesis (Abstr). Dig Dis Sci1986; (Suppl) 31(1):220.40. Wynder EL , Mabuchi K , Maruchi N. Fortner J G. Epidemiol-ogy of cancer of the pancreas. J Natl Cancer Inst 1973; 50:645-667.4 1. Ragozzino M, M elton L, Chu C. Subsequent cancer risk in theincidence cohort of Rochester, M innesota residents with diabetes mel-litus. J Chronic Dis 1982; 35: 13- 19.42. L und E. Gastric cancer after gastric surgery: An increasingproblem in Norway (L etter). Lancet 1983; 1:973.43. L in R, K essler I. A multifactorial model for pancreatic cancer inman. JAMA 1981; 245:147-152.44. Cassimos C, Sklarunu-Zurukzoglu S, Catrin D. T he frequencyof tonsillectomy and appendectomy in cancer patients. Cancer 1973;45. Percy C, Stanek E, Gloeckler L. Accuracy of cancer death certif-icates and its effect on cancer mortality statistics.AmJ Public Health46. Ishii K , Nakamura K, Ozaki I, Y amada N, Takeuchi T. Epide-miological problems of pancreas cancer. J pn J Clin Med 1968;

36:25 1-256.

32:1374-1379.

1981; 71~242-250.

26: 1839-1842.

jurnal ca pnkreas

Documents