Click for next page ( 373


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 372
16 Cancer incidence and Mortality A number of authors (e.g., Doll, 1967, 1977; Higginson, 1969; Higginson and Muir, 1979; Wynder and Gori, 1977) have pointed to the large differences in cancer incidence and mortality that exist between countries and the likelihood that these differences are largely due to environmental factors. Here the term environment is used in its widest sense, encompassing all factors external to humans, as distinct from differences that are attributable to man's genetic makeup. Higginson (1969) calculated the proportion of cancers that were theoretically pre- ventable and suggested that approximately 90% of all cancers in humans are influenced by exogenous factors. This observation has stimulated attempts to distinguish between occupation and way of life (Fox and Adelstein, 1978) and to identify those components linked to lifestyle (Higginson and Muir, 1979; Miller, 1981--specifically, to examine diet as a component of lifestyle (Miller et al., 1980; Wynder and Gori, 1977) This chapter reviews those aspects of descriptive epidemiology that indicate the importance of dietary factors in explaining differences in cancer incidence and mortality among various population groups. In gen- eral, incidence data are used when available, since they more directly relate to etiology, being uninfluenced by changes in survival due to advances in the management and treatment of cancer. In the absence of incidence data, mortality data have been used. As pointed out by Doll and Peto (1981), mortality data have certain advantages as well. Their interpretation is less often complicated by changes in diagnostic prac- tices or cancer registration. The diet is generally associated with cancers of the gastrointestinal tract (i.e., esophagus, stomach, colon, rectum, pancreas, and liver) and cancers of some sex-hormone-responsive sites (i.e., breast, prostate, endometrium, and ovary). There is also evidence that diet is associated to some degree with cancers of the respiratory system and bladder. Inci- dence data for 1973-1977, as compiled by the National Cancer Institute Surveillance, Epidemiology, and End Results (NCI-SEER) Program (Young et al., 1981), indicate that cancer of the stomach, colon, breast, bladder, and prostate comprise m39% of the cancers in males and m43% of cancers in females. Thus, these presumably diet-sensitive sites account for approximately 40% of the cancers in both sexes (Young et al., 1981~. If we add the incidence of lung cancer in males, which may be influenced by diet as well as by smoking, the total for cancer in males reaches 60%. (;E:OGRAPHICAL DIFFERENCES RELATED TO ETHNICITY In the United States, incidence data for whites and blacks have been compared. Table 16-1 provides such comparisons for gastrointestinal 372 16-1

OCR for page 372
Cancer Incidence and Mortality 373 TABLE 16-1 Cumulative Incidence Rate (Age 0-74) per 100 Persons for Malignant Neoplasms in All Areas of the United States (Excluding Puerto Rico), 1973-1977a Whites Blacks Site Males Females Males Females Esophagus 0.5 0.2 1.7 0.4 Stomach 1.1 0.4 2.1 0.9 Colon 3.1 2.7 3.2 2.7 Rectum and rectosigmoid 1.8 1.0 1.4 1.0 Liver 0.2 0.1 0.5 0.2 Gallbladder O.1 O.1 O.1 O.1 Pancreas 1.1 O. 7 1.7 1.1 Lung 7.7 2.3 11.5 2.4 Breast O.1 8.2 0.1 7.0 Corpus uteri 3.1 1.4 Ovary 1.4 0.9 Prostate 5.2 9.6 Bladder 2.4 0.6 1.2 0.5 Kidney 0.9 0.4 0.8 0.4 aData from Young et al., 16-2 1981.

OCR for page 372
374 DIET, NUTRITION, AND CANCER cancers and for some other sites (Young et al., 1981~. The measure used (i.e., cumulative incidence to age 74) is an approximation of the life- time risk of developing cancer of that site in the absence of death from other causes (Day, 1976~. Rates are higher for blacks of both sexes for cancers of the esophagus, stomach, liver, and pancreas, and there is a substantial excess of prostate cancer in black males. However, the inci- dence of bladder cancer is lower in blacks. A number of cancer registries provide incidence data for different racial groups (Waterhouse et al., 1976~. For example, the age-adjusted rates for stomach and large bowel cancer among Chinese living in the San Francisco Bay area are similar to those for whites, whereas rates for prostate and bladder cancer are lower. In Hawaii, the rates for stomach cancer are higher for Hawaiians and Japanese of both sexes and for Chinese and Filipino females than the corresponding rates for Caucasians (Table 16-2) (Young et al., 1981~. The differences in the rates of cancer among racial and ethnic groups do not necessarily have a ready interpretation. Some, such as the similarity of rates among the various ethnic groups (Henderson et _., in press) suggest the possibility of a genetic contribution to cancer at some sites. However, other explanations should also be considered. Most of the differences would indicate that cultural and, thus, environmental factors are involved in the etiology of cancers at the various sites. Among these, dietary factors are likely to be of critical importance. CHANGES SUBSEQUENT TO MIGRATION The hypothesis that differences in cancer incidence and mortality among racial or ethnic groups are due largely to cultural rather than genetic factors receives considerable support from data on groups that have migrated (Haenszel, 1961; Kmet, 1970~. In general, the incidence of cancer in migrant groups is similar to that of the country of origin or is intermediate between that of the country of origin and the host coun- try. After one or more generations, it becomes the same as that of the host country. These trends have been well documented for gastric, colon, and breast cancer in Japanese who migrated to Hawaii and to the western continental United States and Canada (Buell, 1973; Kolonel et al., 1980~; Eastern Europeans who migrated to the United States and Canada (Kmet, 1970~; Icelanders who migrated to Manitoba, Canada (Choi et al., 1971~; and Southern Europeans who migrated to Australia (McMichael et al., 1980~. The changes seem to be most rapid for colon cancer and somewhat less rapid for stomach cancer. They are slowest for breast cancer, requiring more than one, or possibly two, generations for a full effect to be mani- fested. These different rates of change may reflect differences in the stage of life during which factors exert an effect, or they may reflect the differences in the action of carcinogenic initiators or promoting agents. 16-3

OCR for page 372
Cancer Incidence and Mortality 375 TABLE 16-2 Average Annual Age-Adjusted Incidence (per 100,000) in Different Ethnic Groups in Hawaii, 1973-1977a Race Site Sex Hawaiian Caucasian Chinese Filipino Japanese Stomach M 51.4 15.6 14.6 13.3 47.3 F 23.9 7.0 9.4 7.3 19.9 Colon M 20.2 33.8 39.3 24.3 36.5 F 17.5 24.4 28.2 13.2 24.4 Rectum and M 15.9 17.6 20.6 16.5 28.2 rectosigmoid F 10.9 10.2 11.0 10.0 11.0 Liver M 12.6 2.9 10.3 13.3 7.5 F 9.7 1.9 3.6 2.1 2.8 Gallbladder M 2.3 -- 3.3 2.4 1.4 F 1.5 0.9 2.2 1.0 2.2 Pancreas M 12.9 11.4 11.5 8.6 11.7 F 9.2 8.7 7.6 1.8 6.1 Breast F 104.3 99.9 64.1 29.2 51.3 Corpus uteri F 40.4 41.5 33.4 17.3 22.4 Ovary F 11.8 10.7 8.1 5.4 8.0 Prostate ~ 66.3 86.7 40.0 46.9 54.1 Kidney and M 5.8 13.5 2.9 4.7 6.1 pelvis F 5.5 3.0 4.6 3.8 2.5 Bladder M 6.8 31.3 9.9 10.8 13.7 F 8.3 5.2 1.5 5.6 4.8 aData from Young et al., 1981. 16-4

OCR for page 372
376 DIET, NUTRITION, AND CANCER When interpreting the effects of migration, it is important to recog- nize that migrants are not necessarily representative of the population from which they were derived. They may have led more active lives or have a socioeconomic background that is not typical for the general population in their country of origin, or they may have come from a specific region with atypical characteristics. Furthermore, migrant groups may retain some cultural habits, possibly including different dietary practices, while in their host country, even after several generations. MacDonald (1966) compared the diet of 65 Japanese men, age 50 or more who had migrated to Canada 35 to 55 years before, with 65 non-Japanese control subjects. The Japanese ate more fish and rice than did the control group and ingested less beef, potatoes, bread, milk, and cereals other than rice. Fruit and vegetable consumption was similar in the two groups. Hence, there may be differences in cancer incidence among groups from different countries for some time after they migrated. In the United States, breast cancer in postmenopausal women has been linked to German ethnicity (Blot _ al., 1977b). Mortality from colon and rectal cancer is elevated in counties in the United States where there are large popu- lations with Irish, German, or Czechoslovakian descent (Blot et al., 1976~. Mortality from renal cancer is also elevated in counties where a large percentage of the population is of German, Scandinavian, or, especially, Russian descent (Blot and Fraumeni, 1979~. Where differences have been explored in detail, the findings support the operation of environmental factors, as distinct from those of genetic origin, even when "genetic isolates" are considered (Gaudette et al., 1978; Martin _ al., 1980~. CHANGING TIME TRENDS IN INCIDENCE AND MORTALITY - For various reasons, cancer incidence and mortality in the United States are reported neither completely nor accurately. Generally speak- ing, although reporting of incidence has improved during the past 25 years, reporting is probably more accurate and complete for mortality. There have also been recent improvements in the reporting of categories of cancer. Therefore, long-term trends must be interpreted cautiously, and even recent trends are the subject of controversy. Because of population growth and changes in its age distribution, the annual number of cancer cases is steadily increasing in the United States. For 1981, approximately 400,000 cancer deaths are projected, and about 800,000 new cases expected to occur. Cancer is the second most frequent cause of death (about 20% of total deaths). Against a background of drastic environmental modification and rapid technologi- cal change, some have suggested that a vast cancer "epidemic" may have begun. 'Rowever, since more people are living longer and cancer rates 16-5

OCR for page 372
Cancer Incidence and Mortality 377 are higher among older people, adjustment for age corrects the impres- sion that cancer rates are being driven upward dramatically by an in- creasing environmental threat. In fact, examination of recent age- adjusted rates indicates that there has been little change overall, although there are some increases and decreases in cancer incidence at certain sites. Apart from the smoking-associated cancers, stomach cancer, and cer- vical cancer, the incidence of and mortality from cancer at nearly all sites have remained remarkably stable for the last 30 to 40 years (Devesa and Silverman, 1978; Miller, 1980~. There has been a slight increase in the rates of nonrespiratory cancers for white males age 75 or more (but, if anything, a decrease at ages 55-59 and 40-44~. For other age groups, the rates have been stable. In females, there has been an increase in the 74-84 age group, but decreases in the age groups 85 or more, 70-74, and 40-54, and a relative stability for other age groups. The data indicate that rates for the youngest age groups (especially for women) appear to be falling. This is a critical observation because the youngest age groups would have been expected to show the effect first if the environment were becoming increasingly carcinogenic. There has been an increased incidence of breast cancer, especially in postmenopausal women, but there has been little or no change in mortality from cancer at this site (Barclay et al., 1975; Cutler et al., 1971; Fabia et al., 1977; Grace et al., 1977~. This finding is probably due to a com- bination of factors. For example, improved cancer registration has pro- vided a more accurate reflection of cancer incidence. Furthermore, in- creased awareness by the general population and physicians and intensi- fied pathological examination of resected tissue have led to early detec- tion and, thus, slightly more successful treatment. In Iceland, the age-specific curves for breast cancer have shifted from higher rates for premenopausal women (similar to the pattern in Japan in this century) to higher rates for postmenopausal women (similar to the pattern in North American women) (Bjarnason et al., 1974~. How- ever, an analysis of the data by birth cohort demonstrated that the shapes of the age-specific curves were the same and that the incidence at each age increased in successive birth cohorts. This suggests that the inci- dence of breast cancer may have increased because more recent birth co- horts were exposed to environmental factors that increase the risk of breast cancer. The incidence of endometrial cancer increased remarkably during the 1970's, then fell sharply beginning in 1976, especially on the west coast of the United States. There now seems to be little doubt that these changes are due to the excessive use and then partial withdrawal of con- jugated estrogens at the time of menopause (Jick et al., 1980~. Despite the dramatic changes in incidence, mortality has remained stable. 16-6

OCR for page 372
378 DIET, NUTRITION, AND CANCER Both incidence of and mortality from stomach cancer have declined steadily in Western countries during the past 30 to 40 years (Devesa and Silverman, 1978; Miller, 1980~. More recently, they have started to decline in Japan (Hirayama, 1977~. These trends and stabilities are reflected in international incidence data, apart from minor fluctuations due to changing registration practices (Stukonis, 1978~. Recently, controversy has arisen over whether the general incidence of cancer has been increasing significantly over the past several dec- ades, in particular during the past 5 years. Resolution of this contro- versy is complicated by the inadequacies in the data and by the different trends for specific cancer sites. If lung and skin cancer are excluded (the former because most of it is attributed to cigarette smoking, the latter because it is easily curable and poorly reported), the remaining aggregated cancer incidence appears to be roughly stable up to 1971. Differences of interpretation have arisen from the comparison of SEER data (collected after 1973) with the earlier data because they represent different population samples and different methods of data collection. Thus, Pollack and Harm (1980) suggested that cancer incidence is rising in the United States, even for cancers not associated with smoking. Their report was based on data from the third National Cancer Survey (1959-1971) and the SEER Program for 1973-1976. Although the authors took pains to exclude methodological reasons for the increase and to regard it as real, it still seems possible that most of the ob- served increases are artifactual, caused by more efficient registration following the change from the earlier methods of a one-time survey to those of a permanent registration system. The matter should be resolved in the next few years as SEER data continue to accumulate. INTERS ITE CORRELATIONS OF INC IDENCE . Burkitt (1971) pointed out that similar frequencies for different diseases might imply common etiological factors. Using data from the third National Cancer Survey (Cutler and Young, 1975), Winkelstein et al. (1977) studied the geographic variation in the occurrence of cancers at sites common to both sexes and at five sex-specific sites. They found strong correlations among the incidence rates for cancer at three gastro- intestinal sites, i.e., cancers of the colon and rectum were directly cor- related with each other, and inversely correlated with stomach cancer. In addition, there was a strong direct correlation between colorectal cancer and bladder cancer in both men and women. There was also a strong direct correlation among cancer of the breast, corpus uteri, and ovary in women. These two groups of interrelated sites were also correlated with one another. Berg (1975) pointed out that international incidence rates for the hormone-dependent cancers (i.e., breast, corpus uteri, and ovarian can- cers in females and testis and prostate cancers in males) were closely 16-7

OCR for page 372
Cancer Incidence and Mortality 379 correlated with the rates for large bowel cancer. Table 16-3 shows the extent to which the the cancer incidence rates for certain sites are intercorrelated. The lowest correlation between males and females was found for esophageal cancer; the highest, for colon cancer. Significant direct correlations within the same sex occur for colon and rectal cancer and for both sites combined with pancreatic cancer. There are significant inverse correlations for colon and stomach cancer. In men, esophageal cancer is directly correlated with liver cancer; in women, liver cancer is inversely correlated with colon and rectal cancer. Colon and, to a lesser extent, rectal and pancreatic cancers are directly correlated with breast, endometrial, and ovarian cancers in women. ASSOCIATIONS WITH SOCIOECONOMIC STATUS Using data for 1950 to 1969, Blot et al. (1977a) analyzed mortality in the United States by county. This study has provided a number of opportunities to evaluate mortality rates by socioeconomic status, which may reflect differences in diet. Thus, mortality from colon and rectal cancer has been associated with higher income and education levels (Blot et al., 1976~; mortality from breast cancer in postmenopausal women has been linked to higher socioeconomic status (Blot et al., 1977a); and mortality from renal cancer has been directly correlated in both sexes with higher socioeconomic status (Blot and Fraumeni, 1979~. In the United Kingdom, the Registar General's Decennial Supplement for England and Wales (Registrar General's Office for England and Wales, 1978) and a report by Fox and Adelstein (1978) have shed further light on the relative importance of socioeconomic status and occupational factors. Standardization for social class has shown that nearly all significantly high standardized mortality ratios for occupational groups are reduced to nonsignificance after social class is considered. Fox and Adelstein (1978) have calculated that occupation accounts for approximately 12% of the variation in cancer mortality between social class groups, and that lifestyle accounts for 88%. The corresponding proportions for all causes of death are 18% and 82%, respectively. Teppo et al. (1980), using data from the Finnish Cancer Registry, reported an association between social class and cancer at a number of sites. They found inverse correlations between social class and cancer of the lip and stomach in males and direct correlations between social class and colon cancer in both sexes and breast and lung cancer in females. After examining the incidence of breast cancer by province and selected municipalities in Finland, Hakama et al. (1979) reported an association with taxable income. They concluded that factors reflected by the standard of living and fertility might act independently. 16-8

OCR for page 372
380 cn a v o s~ v ~D ~ V ~ o o o a o tn C) 1 p4 s" C~ o ~ o ~ ~ ~ ~ C~ ~ ~ o c~ ~ ~ ~ O O r~ ~ ~ ~ ~D e - O O O O O O O O O O O 1 1 C~ O ~ ~ ~ ~ ~ ~ ~ . ~ ~ ~ C~ C~ ~ ~ ~ ~ ~ ~ ~ O ~ - O O O O O O O O O O O ' ' 1~ ' ' Io ~ U~ oo o ~ C~ ~ o o o o 1 1 = C~ oo C~ . - a' 0 0 0 0 0 1 1 1 ~ ~ o - 0 O ~ cr~ o o o o C~ 1 1 o a, a a ~: o a, ~ ^ o U ~ o to v S~ ~ JJ u ~ oo ~ ~ o ~ ~ o o o 1 1 l ! l s V U~ o ~ - o o o u C~ cn a, U' C~ s cn ~o ~ Ct V . - C~ o o CO U ~ C~ ~ o o 1 o oo o ~ ~ o o o o 1 1 C~ oo ~ oo o o C~ ~ o o o o o o 1 1 1 o V o C~ o C~ ~ C~ o o o o 1 1 1 ~ ~ C~ o 1 O O ,_ C~ ~ oo ~ o o o 1 ~ ~ U~ o o o 1 C~ o' ~ ~o o o o 1 o C~ ~ o C~ ~ ~ ~ o o o o o 1 1 1 1 o~ U~ o C~ C~ o o o o 1 1 1 s~ s" a) ~ C' ~ U S~ o~ V P~ s~ s~ C' a' o o u o ;: C~ ~o ~g o ,D ~o JJ C~ ~ 0 C) A o V U C) . . o C~ 0 a cr~ . e a' - 0 0 0 . 0 V cn ^ _d U a a) ~ U U 00 0 ~ ~ O e ~ ~ C~ cn c~ a) o 3 ~ 3 o ^ o 3 ~ - U ~D O "a ~ 0 V o ^ cn a U SJ C~ o o o U I\

OCR for page 372
1 Cancer Incidence and Mortality 381 Urban-rural differences in cancer incidence and mortality have long been recognized as indicators of the effect of socioeconomic status and lifestyle. However, only the Norwegian Cancer Registry routinely reports its data in this form, since reports are grouped according to the offi- cial administrative boundaries (Waterhouse et al., 1976~. In urban areas, there is a marked excess of esophageal and liver cancer in males and of lung cancer in both sexes and a moderate excess of colon and pancreatic cancer in males, breast cancer in females, and bladder and other urinary tract cancers in both sexes. In rural areas, there is a marked excess of lip cancer in males. RELIGIOUS PRACTICES AND CANCER INCIDENCE AND MORTALITY Religious groups whose lifestyles and dietary habits differ from those of the general population have been a fruitful source for assessing the possible effect of dietary variables. Phillips and colleagues have studied Seventh-Day Adventists (SDA's), a religious group with approximately 600,000 members in North America (Phillips, 1975; Phillips et al., 1980b). SDA's abstain from smoking and drinking, and approximately 50% of them follow a lacto-ovovegetarian diet. In earlier studies, Phillips (1975) suggested that the lacto- ovovegetarian diet may protect against colon cancer; the evidence for breast cancer was less clear. These studies compared cancer mortality among California SDA's with mortality data for the general California population. In more recent analyses, Phillips et al. (1980a,b) used as a non-SDA control group Californians who enrolled in a concurrent pro- spective study conducted among the general population by the American Cancer Society. This study showed that the risk of dying from colorectal cancer and cancers related to smoking is clearly lower among SDA's than among non-SDA's of comparable age, sex, and socioeconomic status. The risk of dying from breast cancer was also reduced, but not significantly. This may be clarified in further analyses that take into account im- portant risk factors for breast cancer. But since a number of SDA's are adult converts, their risk of breast cancer may have been at least par- tially determined by exposures in early (adolescent or young adult) life, before they adopted SDA practices. However, the SDA age-specific mor- tality curve for breast cancer is consistent with those of other low-- risk populations, in which postmenopausal women have been observed to have a lower risk for cancer at this site (Phillips et al., 1980b). The Mormons comprise another religious group whose lifestyle differs markedly from that of the general U.S. population. For at least 80 years, these members of the Church of Jesus Christ of Latter-Day Saints have proscribed the use of alcohol, tobacco, coffee, and tea in all forms to help ensure good health (Lyon et al., 1980~. In addition, they recom- mend a well-balanced diet, especiSEly the use of grains, fruits, and vegetables, and moderate consumption of meat (Enstrom, 1980~. In a study 16-10

OCR for page 372
382 DIET, NUTRITION, AND CANCER based on data from the Utah Cancer Registry, Lyon et al. (1976) reported that the incidence of cancers associated with cigarette smoking; breast, uterine, cervical, and ovarian cancer in females; and stomach cancer in males was much lower in Mormons than in non-Mormons. Colon cancer was also significantly lower in females, but not in males. In an update of these analyses (Lyon et al., 1980), the same reductions were observed, but for colon and rectal cancer they had become significant for both sexes. In states adjacent to Utah (Idaho and Wyoming), mortality from smoking-associated cancers and from cancers of the rectum and breast is almost identical to that of Utah (Rawson, 1980~. Thus, there may be some general environmental variable peculiar to this entire area affecting cancer incidence and mortality. In India, cancer incidence differs among religious groups, especially between the Parsi and Hindu communities of Bombay (Jussawalla, 1976~. In the Parsi community, the rates of colon, rectal, and breast cancer are substantially greater than those in the Hindu population, although they are not as high as those in Western countries. There are many differences other than dietary factors between religious groups and the general population. Although attempts have been made to control for the nondietary differences (e.g., Phillips et al., 1980b), studies of groups will probably never be as effective as direct evaluations of the effect of dietary factors and other variables in individuals. Nevertheless, these studies of subgroups of the popula- tion have value because they indicate which hypotheses should be evalu- ated more directly by other means. CORRELATIONS OF INCIDENCE AND MORTALITY WITH DIETARY AND OTHER VARIABLES In several studies, dietary and other variables have been found to be strongly correlated with geographical differences in the incidence of and mortality from cancer at a number of sites (Armstrong and Doll, 1975; Carroll, 1975; Knox, 1977~. Armstrong and Doll (1975) correlated inci- dence rates for cancer at 27 sites in 23 countries and mortality rates for cancer at 14 sites in 32 countries with a wide range of dietary and other variables. They reported strong correlations between dietary vari- ables and cancer at several sites, especially meat and fat intake with cancers of the colorectum, breast, corpus uteri, and ovary. Direct cor- relations with dietary variables were also found for cancer of the small intestine (sugar), pancreas (eggs, animal protein, and fat), and ovary and bladder (fats and oils); inverse associations were reported for gas- tric cancer (meat, animal protein, and fat) and cervical cancer (total protein and fruit). Many of the dietary variables were strongly inter- correlated, especially fat and protein of animal origin, and were also correlated with gross national product. Carroll (1975) observed a strong correlation between per capita intake of dietary fat and age-adjusted mortality from breast cancer. The correlation was strongest for total 16-11

OCR for page 372
Cancer Incidience and Mortality 383 fat, almost as strong for animal fat, but almost completely absent for vegetable fat. Knox (1977) suggested that associations between alcohol intake and cancer of the mouth and larynx, between total fat intake and cancer of the large intestine and breast, and between beer intake and cancer of the rectum were causal. In Japan, Hirayama (1977) found that the intakes of fat and pork were associated with mortality from breast cancer in 12 different prefectures. After having reviewed data for cancer and fat intake, Enig et al. (1979) retracted their original suggestion that cancer was correlated with the intake of total fat and vegetable fat, but not with animal fat. One disadvantage of this type of study, especially in relation to breast and gastric cancer, is that current dietary factors are usually correlated with current information on incidence and mortality, whereas a more appropriate time relationship might be established by taking dietary information recorded some 20 or 30 years ago and correlating it with current incidence or mortality rates (Miller et al., 1980~. By conducting personal interviews with 4,137 subjects to determine their usual weekly food consumption, Kolonel et al. (1981) determined the average daily consumption of several components of fat in the diets of the five main ethnic groups in Hawaii. The intake of total fat correlated with the ethnic-specific incidence rates of breast cancer in Hawaii, but not with colon or prostate cancer incidence. There was no correlation between cholesterol consumption and incidence of colon cancer. One possible difficulty with this study is that many of the means of dietary intake for the different ethnic groups were rather close. Furthermore, current diet as measured in this study may not be relevant to current incidence, especially for sites where incidence is changing. Evidence associating fiber intake with certain cancer sites has also been contradictory. Drasar and Irving (1973) were unable to find any association of breast and colon cancer incidence with per capita fiber intake using data from 37 countries, although they demonstrated a high correlation with fat and animal protein intake. However, the Interna- tional Agency for Research on Cancer Intestinal Microecology Group (1977) found that differences in dietary fiber in Scandinavian populations appeared to correlate well with incidence of and mortality from colon cancer. Dietary correlation studies have produced evidence for groups, rather than for individuals. Although the variation in the intake of nutrients is great internationally, it is not necessarily extensive for groups within countries. The variation of incidence and mortality within coun- tries may also be low. Furthermore, the dietary information is generally derived from food disappearance data (i.e., per capita intake), and not necessarily from individual food consumption data. Hence, lack of cor- relation, either nationally or internationally, does not suffice to dis- prove a hypothesis. Conversely, one should not rely too heavily on ob- served correlations in case they are confounded by some factor that could not be studied or has not yet been identified (Stavraky, 1976~. 16-12

OCR for page 372
384 DIET, NUTRITION,AND CANCER ASSOCIATIONS WITH OTHER DISEASES Burkitt (1971) noted an association between a number of chronic dis- eases and "affluence." lIe suggested that such "diseases of affluence" might have a common etiology and that dietary fiber may play a role in protecting against such illnesses as cancer of the colon and other sites, coronary heart disease, diverticulitis, etc. There has been a recent decline in mortality from ischemic heart disease in the United States, Australia, and Canada, but not in the United Kingdom (Dwyer and Hetzel, 1980~. This decline has been associated by some investigators with in- creased consumption of polyunsaturated fats and reduced cigarette smok- ing. We may therefore question why there has not been any indication of a reduction in cancers associated with dietary components, especially fat. Since the amount of vegetable fat in the U.S. diet has been in- creasing (Page and Friend, 1978), one possible explanation could be that such fat has an adverse influence on colorectal and breast cancer inci- dence (Enig et al., 1978~. An alternative explanation might be that the substitution of polyunsaturated fats for saturated fats may help to re- duce cardiovascular diseases, but for this substitution to have an effect on fat-associated cancers, a concurrent reduction in total fat consump- tion may also be necessary (Miller et al., 1980~. It is also possible that any effect of fat on cancer incidence would take longer to appear than its effect on cardiovascular disease; however, in view of the rapid changes in colorectal cancer rates after migration, this appears to be an unlikely explanation for cancer at this site. Finally, dietary changes may not have had any influence on changes in cardiovascular disease. SUMMARY The incidence of cancers differs greatly among countries and to a limited extent within countries. Studies of migrants (e.g., Japanese who migrated to the United States) suggest environmental causes for these differences. In affluent countries, stomach cancer rates have fallen, rates for intestinal and breast cancer are stable, and pancreatic cancer rates, which have increased, are now descreasing for males. The inci- dence of cancer of the colon, rectum, breast, corpus uteri, ovary, and prostate are directly correlated with each other, but cancer of the colon or rectum is inversely correlated with stomach cancer. Mortality from colorectal and breast cancer is directly associated with socioeconomic status, and stomach cancer is inversely associated. In males, incidence of cancer of the esophagus, liver, colon, pancreas, and lung is higher in urban areas. Seventh-Day Adventists and Mormons have a low risk for colon and breast cancer. Internationally, the intake of fat has been directly correlated with cancer of the breast, colon, rectum, pancreas, corpus uteri, and ovary. 16-13

OCR for page 372
Cancer Incidence and Mortality 385 REFERENCES Armstrong, B., and R. Doll. 1975. Environmental factors and cancer incidence and mortality in different countries, with special refer- ence to dietary practices. Int. J. Cancer 15:617-631. Barclay, T. H. C., M. M. Black, B. F. Hankey, and S. J. Cutler. 1975. The increasing incidence of breast cancer in Saskatchewan. Pp. 282-288 in P. Bucalossi, U. Veronesi, and N. Cascinelli, eds. Proceedings of the XI International Cancer Congress, Volume 3. Excerpta Medica, Amsterdam. Berg, J. W. 1975. Can nutrition explain the pattern of international epidemiology of hormone-dependent cancers? Cancer Res. 35:3345- 3350. Bjarnason, O., N. Day, G. Snaedal, and H. Tulinius. 1974. The effect of year of birth on the breast cancer age-incidence curve in Ice- land. Int. J. Cancer 13:689-696. Blot, W. J., and J. F. Fraumeni, Jr. 1979. Geographic patterns of renal cancer in the United States. J. Natl. Cancer Inst. 63: 363-366. Blot, W. J., J. F. Fraumeni, Jr., B. J. Stone, and F. W. McKay. 1976. Geographic patterns of large bowel cancer in the United States. J. Natl. Cancer Inst. 57:1225-1231. Blot, W. J., T. J. Mason, R. Hoover, and J. F. Fraumeni, Jr. 1977a. Cancer by county: Etiologic implications. Pp. 21-30 in H. H. Hiatt, J. D. Watson, and J. A. Winsten, eds. Origins of Human Cancer. Book A, Incidence of Cancer in Humans. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. Blot, W. J., J. F. Fraumeni, Jr., and B. J. Stone. 1977b. Geographic patterns of breast cancer in the United States. J. Natl. Cancer Inst. 59:1407-1411. Buell, P. 1973. Changing incidence of breast cancer in Japanese- American women. J. Natl. Cancer Inst. 51:1479-1483. Burkitt, D. P. 1971. Guest editorial: Some neglected leads to cancer causation. J. Natl. Cancer Inst. 47:913-919. Carroll, K. K. 1975. Experimental evidence of dietary factors and hormone-dependent cancers. Cancer Res. 35:3374-3383. 16-14

OCR for page 372
386 DIET, NUTRITION, AND CANCER Choi, N. W., D. W. Entwistle, W. Michaluk, and N. Nelson. 1971. Gastric cancer in Icelanders in Manitoba. Isr. J. Med. Sci. 7:1500-1508. Cutler, S. J., and J. L. Young, Jr., eds. 1975. Third National Cancer Survey: Incidence data. Nat. Cancer Inst. Mbnogr. 41:1-454. Cutler, S. J., B. Christine, and T. H. C. Barclay. 1971. Increasing incidence and decreasing mortality rates for breast cancer. Cancer 28:1376-1380. Day, N. E. 1976. A new measure of age standardized incidence, the cumulative rate. Pp. 443-445 in J. Waterhouse, C. Muir, P. Correa, and J. Powell, eds. Cancer Incidence in Five Continents, Volume III. IARC Scientific Publications No. 15. International Agency for Research on Cancer, Lyon, France. Devesa, S. S., and D. T. Silverman. 1978. Cancer incidence and mor- tality trends in the United States: 1935-74. J. Natl. Cancer Inst. 60:545-571. Doll, R. 1967. Prevention of Cancer. Pointers from Epidemiology. The Nuffield Provincial Hospitals Trust, London. 143 pp. Doll, R. 1977. Strategy for detection of cancer hazards to man. Nature 265:589-596. Doll, R., and R. Peto. 1981. The causes of cancer: Quantitative estimates of avoidable risks of cancer in the United States today. J. Natl. Cancer Inst. 66:1191-1308. Drasar, B. S., and D. Irving. 1973. Environmental factors and cancer of the colon and breast. Br. J. Cancer 27:167-172. Dwyer, T., and B. S. Hetzel. 1980. A comparison of trends of coronary heart disease mortality in Australia, USA and England and Wales with reference to three major risk factors--hypertension, cigarette smoking and diet. Int. J. Epidemiol. 9:65-71. Enig, M. G., R. J. Munn, and M. Keeney. 1978. Dietary fat and cancer trends--a critique. Fed. Proc. Fed. Am. Soc. Exp. Biol. 37:2215- 2220. Enig, M. G., R. J. Munn, and M. Keeney. 1979. Response to letters. Fed. Proc. Fed. Am. Soc. Exp. Biol. 38:2437-2439. Enstrom, J. E. 1980. Health and dietary practices and cancer mortality among California Monmons. Pp. 69-90 in J. Cairns, J. L. Lyon, and M. Skolnick, eds. Cancer Incidence in Defined Populations. Banbury Report 4. Cold Spring Harbor Laboratory, Cold Spring Harbor, N. Y. 16-15

OCR for page 372
Cancer Incidence and Mortality 387 Fabia, J., P.-M. Bernard, and G. Hill. 1977. Recent time-trends of age-specific death rates for breast cancer: Quebec and other provinces, 1965 through 1974. Can. Med. Assoc. J. 116:1135-1138. Fox, A. J., and A. M. Adelstein. 1978. Occupational mortality: Work or way of life. J. Epidemiol. Community Health 32:73-78. Gaudette, L. A., T. M. Holmes, L. M. Laing, K. Morgan, and M. G. A. Grace. 1978. Cancer incidence in a religious isolate of Alberta, Canada, 1953-74. J. Natl. Cancer Inst. 60:1233-1238. Grace, M., L. A. Gaudette, and P. E. Burns. 1977. The increasing incidence of breast cancer in Alberta, 1953-1973. Cancer 40:358-363. Haenszel, W. 1961. Cancer mortality among the foreign-born in the United States. J. Natl. Cancer Inst. 26:37-132. Hakama, M., I. Soini, E. Kuosma, M. Lehtonen, and A. Aromea. 1979. Breast cancer incidence: Geographical correlations in Finland. Int. J. Epidemiol. 8:33-40. Henderson, B., L. Kolonel, and F. Foster. In press. Cancer in the Polynesians. Natl. Cancer Inst. Mongr. Series. Higginson, J. 1969. Present trends in cancer epidemiology. Pp. 40-75 in J. F. Morgan, ed. Proceedings of the Eighth Canadian Cancer Research Conference. Pergamon Press, Oxford, London, Edinburgh, New York, Toronto, Paris, and Frankfurt. Higginson, J., and C. S. Muir. 1979. Guest editorial: Environmental carcinogenesis: Misconceptions and limitations to cancer control. J. Natl. Cancer Inst. 63:1291-1298. Hirayama, T. 1977. Changing patterns of cancer in Japan with special reference to the decrease in stomach cancer mortality. Pp. 55-75 in H. H. Hiatt, J. D. Watson, and J. A. Winsten, eds. Origins of Human Cancer. Book A, Incidence of Cancer in Humans. Cold Spring Harbor Laboratory, Cold Spring Harbor, N. Y. International Agency for Research on Cancer Intestinal Microecology Group. 1977. Dietary fibre, transit-time, faecal bacteria, steroids, and colon cancer in two Scandinavian populations. Lancet 2:207-211. Jick, H., A. M. Walker, and K. J. Rothman. 1980. The epidemic of endometrial cancer: A commentary. Am. J. Public Health 70:264- 267. 16-16

OCR for page 372
388 DIET, NUTRITION, AND CANCER Jussawalla, D. J. 197 6. The problem of cancer in India: An epidemio- logical assessment. Gann Monogr. Cancer Res. 18:265-273. ~net, J. 1970. The role of migrant population in studies of selected cancer sites: A review. J. Chronic Dis. 23:305-324. Knox, E. G. 197 7. Foods and diseases. Br. J. Prev. Soc. Med. 31: 71-80. Kolonel, L. N., M. W. Hinds, and J. H. Hankin. 1980. Cancer patterns among migrant and native-born Japanese in Hawaii in relation to smoking, drinking and dietary habits. Pp. 327-340 in H. V. Gelboin, B. MacMahon, T. Matsushima, T. Sugimura, S. Takayama, and H. Takebe, eds. Genetic and Environmental Factors in Experimental and Human Cancer. Japan Scientific Societies Press, Tokyo. Kolonel, L. N., J. H. Hankin, A. M. Nomura, and S. Y. Chu. 1981. Dietary fat intake and cancer incidence among five ethnic groups in Hawaii. Cancer Res. 41: 3727-3728. Lyon, J. L., M. R. Klauber, J. W. Gardner, and C. R. Smart. 1976. Cancer incidence in Mormons and non-Mormons in Utah, 1966-1970. N. Engl. J. Med. 294 :129-133. Lyon, J. L., J. W. Gardner, and D. W. West. 1980. Cancer risk and life-style: Cancer among Mormons from 1967-1975. Pp. 3-27 in J. Cairns, J. L. Lyon, and M. Skolnick, eds. Cancer Incidence in Defined Populations. Banbury Report 4. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. MacDonald, W. C. 1966. Gastric cancer among the Japanese of British Columbia: Dietary studies. Pp. 451-459 in R. W. Bigg, C. P. Leblond, R. L. Noble, R. J. Rossiter, R. M. Taylor, and A. C. Wallace, eds. Proceedings of the Sixth Canadian Cancer Research Conference. Pergamon Press, London, Edinburgh, New York, Toronto, Paris, and Frankfurt. Martin, A. 0., J. K. Dunn, and B. Smalley. 1980. Use of a genea- logically linked data base in the analysis of cancer in a human isolate. Pp. 235-251 in J. Cairns, J. L. Lyon, and M. Skolnick, eds. Cancer Incidence in Defined Populations. Banbury Report 4. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. McMichael, A. J., M. G. McCall, J. M. Hartshorne, and T. L. Woodings. 1980. Patterns of gastro-intestinal cancer in European migrants to Australia: The role of dietary change. Int. J. Cancer 25:431-437. 16-17

OCR for page 372
Cancer Incidence and Mortality 389 Miller, A. B. 1980. The epidemiology of malignant disease. A basis for public policy. Health Community Informatics 6:283-294. Miller, A. B. 1981. Epidemiology of gastrointestinal cancer. Compr. Therapy 7:53-58. Miller, A. B., G. B. Gori, S. Graham, T. Hirayama, M. Kunze, B. S. Reddy, and J. H. Weisburger. 1980. Nutrition and cancer. Prev. Med. 9:189-196. Page, L., and B. Friend. 1978. The changing United States diet. Bioscience 28:192-197. Phillips, R. L. 1975. Role of life-style and dietary habits in risk of cancer among Seventh-Day Adventists. Cancer Res. 35:3513-3522. Phillips, R. L., J. W. Kuzma, and T. M. Lotz. 1980a. Cancer mortality among comparable members versus nonmembers of the Seventh-Day Adventist Church. Pp. 93-102 in J. Cairns, J. L. Lyon, and M. Skolnick, eds. Cancer Incidence in Defined Populations. Banbury Report 4. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. Phillips, R. L., J. W. Kuzma, W. L. Beeson, and T. Lotz. 1980b. Influ- ence of selection versus lifestyle on risk of fatal cancer and car- diovascular disease among Seventh-Day Adventists. Am. J. Epidemiol. 112:296-314. Pollack, E. S., and J. W. Harm. 1980. Trends in cancer incidence and mortality in the United States, 1969-1976. J. Natl. Cancer Inst. 64 :1091-1103. Rawson, R. W. 1980. The total environment in the epidemiology of neoplastic disease: The obvious ain't necessarily so. Pp. 109-119 in J. Cairns, J. L. Lyon, and M. Skolnick, eds. Cancer Incidence in Defined Populations. Banbury Report 4. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. Registrar General's Office for England and Wales. 1978. Occupational Mortality: The Registrar General's Decennial Supplement for England and Wales, 1970-72, XVIII. Series D S 1. HMSO No. [O 11 69064483. Her Majesty's Stationery Office, London. Stavraky, K. M. 1976. The role of ecologic analysis in studies of the etiology of disease: A discussion with reference to large bowel cancer. J. Chronic Dis. 29:435044. Stukonis, M. K. 1978. Cancer Incidence Cumulative Rates--International Comparison Based on Data from Cancer Incidence in Five Contients . 16-18

OCR for page 372
390 DIET, NUTRITION, AND CANCER IARC Technical Report No. 78/002. International Agency for Research on Cancer, Lyon, France. 54 pp. Teppo, L., E. Pukkala, M. Hakama, T. Hakulinen, A. Herva, and E. Saxen. 1980. Way of life and cancer incidence in Finland. A municipality- based ecological analysis. Scand. J. Sac. Med. Suppl. 19:1-84. Waterhouse, J., C. Muir, P. Correa, and J. Powell, eds. 1976. Cancer Incidence in Five Continents, Volume III. IARC Scientific Publications No. 15. International Agency for Research on Cancer, Lyon, France. 584 pp. Winkelstein, W., Jr., S. T. Sacks, V. L. Ernster, and S. Selvin. 1977. Correlations of incidence rates for selected cancers in the nine areas of the Third National Cancer Survey. Am. J. Epidemiol. 105:407-419. Wynder, E. L., and G. B. Gori. 1977. Contribution of the environment to cancer incidence: An epidemiologic exercise. J. Natl. Cancer Inst. 58:825-832. Young, J. L., Jr., C. L. Percy, and A. J. Asire, eds. 1981. Surveil- lance, epidemiology, and end results: Incidence and mortality data, 1973-1977. Natl. Cancer Inst. Monogr. 57:1-1081. 16-19