Dietary Patterns Among Men of Japanese Ancestry in Hawaii'

[CANCER RESEARCH 35, 3259-3264,November 1975]

Dietary Patterns among Men of Japanese Ancestry in Hawaii'

Jean H. Hankin,2 Abraham Nomura, and George G. Rhoads

Japan-Hawaii Cancer Study, National Cancer Institute and School ofPublic Health, University of Hawaii, Honolulu, Hawaii 96822

Summary total fat in Hawaii versus40% in Japan. However, the total fat intake in Hawaii Japanese (33%) was still below the The dietary patterns of 6663 men of Japanese ancestry, 42% reported for the average American (2). living in Hawaii and participating in the Honolulu Heart Only limited information on the use of specific food items and Japan-Hawaii Cancer Studies, were analyzed according was included in Tillotson's paper. The Hawaii Japanese ate to country of birth and boyhood education. Approximately substantially more bread, butter, margarine, and cheese 80% of the men were born and educated in Hawaii (Nisei); than men living in Japan, and slightly more meat products. the others were either born in Japan (Issei) or traveled'- to Consumption of fish, tofu3 (soybean curd), and rice was Japan for 5 or more years of boyhood education (Kibei). considerably less frequent in Hawaii than in Japan, al Twenty-four-hr diet recalls, obtained at the first cycle of though rice was still eaten two or more times daily by examinations (1965 to 1968), revealed that the Nisei one-half of the Hawaii Japanese men. Both the nutrient consumed significantly greater intakes of total and animal and food item data suggested that the Hawaii Japanese protein, total and saturated fat, and cholesterol than the men were largely, but not completely, Westernized with re Issei and Kibei. Values for weight, height, skinfold thick spect to their eating habits. ness, and serum cholesterol were in the same direction, and These findings support the concept that migrant popula the differences were statistically significant. tions usually change their diet and other traditional habits Food frequency questionnaires at the 1st and 3rd exami toward the prevailing customs of the host country. Such nations covered a 6-year interval. At both time periods, the changes are of interest since it is now clear that they Issei and Kibei ate Japanese foxls more frequently and in accompany, and probably cause, changes in the incidence of greater quantities than the Nisei. In general, the Nisei specific types of cancer and other chronic diseases. It is consumed more Western foods. Both food frequency ques important to determine whether these changes occur in a tionnaires included 6 identical items: coffee, milk, green tea, uniform or heterogeneous fashion in a given migrant rice, tofu (soybean curd), and tsukudani (preserved seaweed population, because variation provides an opportunity to paste). The frequent and infrequent consumers were simi link specific dietary or cultural changes to specific diseases. larly characterized over the 6-year period. We have looked for heterogeneity in the Hawaii Japanese The findings suggest that the country of birth and by comparing the consumption of specific food items and education has lasting effects on adult eating patterns. The nutrients among men who might be expected to have observed heterogeneity for specific food items and nutrients retained their Japanese habits with those who might not. between the Nisei and Issei-Kibei men augers well for For the 1st group, we identified men who were born or attempts to relate such items to chronic diseases such as educated in Japan; for the 2nd group, we used men born and cancer. educated in Hawaii.

Introduction Materials and Methods

Like other migrant populations, persons of Japanese Approximately 8000 men of Japanese ancestry, born ancestry living in Hawaii have retained some traditional between 1900 and 1919, were examined from 1965 to 1968 food habits and accepted others from their host country. during the 1st cycle of the Honolulu Heart Study. The Tillotson et a!. compared the diets ofJapanese men, 45 to 69 identification and recruitment of this cohort has been years of age, living in Japan and in Hawaii (10). They found describedpreviously (12). More than 6600men returned for that the Hawaii Japanese consumed about twice as much their 3rd round of examinations, which was carried out fat and substantially less carbohydrate than the indigenous under the combined auspices of the Honolulu Heart Study Japanese. Furthermore, 35% ofthe carbohydrate consumed by the Hawaii Japanese was in the form of sugar compared to 20% in Japan: Saturated fat comprised 65 to 70% of the 3 Description of Japanese food items: tofu, soybean curd: tsukudani, preserved seaweedpaste; soba, buckwheat noodles; udon or saimin, 1 Presented at the Conference on Nutrition in the Causation of Cancer, noodles with broth; miso soup, soup with soybean paste; kamabuko, May 19to 22, 1975,Key Biscayne,Fla. Researchsupportedby Contract steamed fish cake; sashimi, fillet of selected raw fish; hakusaizuke, pickled NOl CP 33216 from the National Cancer Institute, NIH. Chinese cabbage; takuwan, Japanese pickled turnips; and ume, Japanese 2 Presenter. pickled plums.

NOVEMBER 1975 3259

and the Japan-Hawaii Cancer Study approximately 6 years item. A description of this method and of its validation is later. described elsewhere (6). For this paper, we computed the On the basis of their birthplace and residential pattern, we individual estimated meandaily intake for 21of theseitems separated this cohort into 3 distinct groups which might be which were considered to be predominantly Japanese or expected to show different degrees of acculturation to the Western. Twelve items were common to both cultures and Western pattern of living. The Issei are Japanese persons were omitted. The frequency questionnaires of the 1st and born in Japan who migrated to Hawaii. Nisei are those of 3rd examinations included some identical items, permitting Japanese ancestry who were not born in Japan; nearly all comparisons of eating habits over a 6-year period of such men were born in Hawaii. Kibei (as defined for this residence in Hawaii. study) are Nisei who traveled to Japan before the age of 11 Information on body height, weight, subscapular skin years and stayed there 5 or more years, usually for their folds, and serum cholesterol was taken from the 1st elementary or secondary education. They currently reside in examination. Descriptions of the methods used for these Hawaii. measurements have been published previously (8). We utilized dietary data of the 6663 persons who participated in both the 1st and 3rd cycles of these Results examinations. During the 1st examination, dietitians used standard methods to obtain individual recalls of foods and The distribution of Nisei, Issei, and Kibei according to amounts consumed during the previous 24 hr ( 10). Food age at time of 1st examination is shown in Table 1. The models and serving utensils were available to illustrate numbers in the younger age groups were strikingly greater portion sizes, and standard food composition tables were for Nisei than Issei or Kibei, and approximately 80% of the used to compute the individual intakes of calories and total study participants were Nisei. nutrients. We selected the data on calorie, total and animal The specific food items of the 2 questionnaires were protein, total and saturated fat, total carbohydrate, sucrose, separated according to primary association with a Western starch, and cholesterol intakes for this study (10). or Japanese diet, and age-adjusted comparisons were made An additional dietary questionnaire was administered at between the Nisei and Issei and between the Nisei and that time to ascertain the usual frequencies of eating 26 Kibei. The Nisei-Issei comparisons excluded persons in the selected items. The frequencies were classified as: (a) almost 45 to 49 and 50 to 54 age groups because of the small never; (b) less than twice a week; (c) 2 to 4 times a week; (d) numbers of Issei under age 55. All 5 age groups were almost daily; and (e) more than once a day. Sixteen of these included in the Nisei-Kibei comparisons. items could be classified as typically Japanese or Western, Table 2 shows the percentage of frequent consumers of and these were tabulated for this analysis. For each item, Western and Japanese items at the 1st examination. In the subjects were classified as frequent or infrequent general, a greater proportion of Nisei frequently consumed consumers. The definition of a frequent consumer was based Western foods than did Issei or Kibei. Conversely, Issei and on the distribution of responses to the particular item. For particularly Kibei ate more Japanese items than the Nisei. commonly consumed foods, such as bread, butter, and milk, The differences for Japanese items were in general more a frequent consumer was defined as one who ingested the impressive than for Western items, and it was notable that item more than once a day. For foods occasionally con milk and ice cream were equally popular in all 3 groups. sumed, such as tofu, miso soup, and non, the definition was The age-adjusted quantitative intakes obtained in the 3rd applied to those who ate the item 2 or more times a week. examination among the 3 groups are presented in Table 3. During the 3rd cycle, trained bilingual interviewers With a few exceptions, such as celery, the trends for the 3 administered a 7-day recall questionnaire to obtain esti groups were similar to findings in Table 2. These data mated frequency of intake, as well as amount of 33 food suggested that Issei and Kibei have more traditionally items, suggested as possibly carcinogenic or protective of Japanese dietary patterns than the Nisei. The similarities gastrointestinal cancer in published papers (3, 5, 9, 11, 13, among the Issei and Kibei led us to combine these 2 groups 14). Portion sizes were estimated from 5- x 7-inch colored in subsequent analyses. photographs of small, medium, and large servings of each Table 4 reveals consistent differences in body measure

Table I Distribution of men by birthplace, residential pattern, and age

examinationGroupâ€•45â€”49 No. of persons by age at1st

yr65â€”69yrTotalNisei1500205510395391215254Issei1833III375179716Kibei1273011578721693yr50-54 yr55-59 yr60â€”64

a Nisei are persons of Japanese lineage not born in Japan and generally born in Hawaii. Issei are persons born in Japan who migrated to Hawaii. Kibei are Nisei who traveled to Japan before the age of I I andstayedthereSor moreyears,usuallyfor their elementaryor secondaryeducation.

3260 CANCER RESEARCH VOL. 35

Table 2 food items at the 1st examination tended to be similarly Percentage offrequent consumersof Western and Japanesefood items characterized 6 years later. among Nisei, Issei, and Kibei at 1st examination

values@'%Age-adjusted Discussion eatingfoodeating% itemsatitemsfood It is of interest that the country of birth and of education age55-69yr45â€”69yrFood ageat appears to have a lasting effect on adult eating patterns among the Japanese in Hawaii. These effects seem even stronger among the Kibei than among the Issei. This is not KibeiWesternMeat25.7itemsNisei IsseiNisei surprising, because the Kibei returned to their country of heritage during a formative period of their development and 28.9bHam,bacon,sausage17.5 20.533.3 lived under the direct influence ofJapanese culture. Further 16@3bButter, 17.622.2 more, the parents of Kibeis probably had closer ties with 4.6Bread33.7margarine,cheese8.3 5.67. 1 26.8bIcecream11.0 27.933.4 their native country than the parents of Nisei children who 9.4Coffee41.4 11.29.4 did not return to Japan. 40.8bMilk8.2 32.3b45.9 The existence of long-term dietary patterns is supported 9.0JapaneseRice46.7 8.87.0 by the reproducibility of dietary frequency data over a 6-year period, utilizing 2 different methods of data collec 61.2bTofu22.1 [email protected] tion. If chronic diseases, such as stomach cancer, have their 29.3â€•Tsukudani8. 41.4a17.2 onset in early life (5), characterization of current diet (as an .4â€•Greentea9.1 1 9.76.9 II indicator of prior dietary patterns) may provide useful 13.7bSoba38.9 15.3â€•6.4 information on etiology. For other chronic diseases, such as 39.7Udon10.3 38.335.1 14.3Misosoup14.5 12.411.5 colorectal cancer, which appear to be largely determined in 24.1â€•Fish7.3 26.9â€•13.0 later life (4), the usefulness of adult dietary studies is even 10.5â€•Nori,otherseaweeds18.6 I3.4b5.4 27.5â€•18.2 30.1Â° Table 3 a Age adjusted in 5-year age groups by the direct method. Mean daily intakes of Western and Japanesefood items among Nisei, b Differs significantly from Nisei at 0.01 level. Issei, and Kibei at 3rd examination

valuesaAmountAmountconsumedconsumed(g)at(g)atageAge-adjusted ments between the Nisei and Issei-Kibei. Values for weight, height, skinfold thickness, and serum cholesterol were significantly greater among the Nisei than the Issei-Kibei men. The 24-hr dietary recall data were generally in the yrFooditemsNisei 55-69 yrage 45-69 same direction. The greater intakes of total and animal protein, total and saturated fat, sucrose, and cholesterol KibeiWesternBeef191 IsseiNisei among Nisei than among Issei or Kibei groups were expected, due to the association of these nutrients with Western foods. Differences were statistically significant 196Â°Wieners28 167Â°220 33Spam 3434 at the 0.01 level for all variables except total carbohydrate 26Viennaorporksausage20or luncheonmeat18 2224 and cholesterol intakes. Total carbohydrate intakes did not 22Cornonthecob21 12Â°23 differ between the groups, although the percentage of total 14Â°Celery30 1522 calories as carbohydrate was slightly higher in the Issei 40â€•Coffee2336 4029 2633Milk650 21722723 Kibei than in the Nisei (p < 0.01). Eggs are used frequently 679JapaneseRice1439 680609 in traditionally Japanese diets and may account for the similarity in cholesterol intakes. Charts I to 6 illustrate the age-specific differences between the 2 groups for weight, 1737Â°Tofu148 1558ISIS 215Â°136 skinfold thickness, serum cholesterol, total calories, satu 201Â°Tsukudani2 3Greentea820 32 rated fat, and sucrose, respectively. The age-specific differ 1040â€•Saiminorudon332 1370â€•763 ences are consistent in direction for all of these variables. 352Â°Japanesesoup250 412Â°292 The 1st and 3rd examination dietary questionnaires 389Â°Kamabuko15 420'247 included 6 identical food items: coffee, milk, green tea, rice, 21Â°Sashimi57 19IS 83Â°Dried 77Â°52 tofu, and tsukudani. Although the methods of quantitating 22Driedfish, as mackerel26 2722 frequencies differed in the 2 questionnaires, the stability of 1Hakusaizuke42cuttlefishI 31 the individual food patterns can be observed. The mean 55Â°Takuwan12 5039 frequency of intake of each of these items at the 3rd 15Ume6 IlII examination is shown according to the frequency reported 96 10Â° by the same individuals at the 1st examination. Charts 7 and a Age adjusted in 5-year age groups by the direct method. 8 show that the frequent and infrequent consumers of the 6 b Differs significantly from Nisei at 0.01 level.

NOVEMBER 1975 3261

J. H. Hankin eta!.

@ 0@oâ€˜i'sâ€”(@4 00 t'.. â€˜1@Q@ @@@ .=2 ,,@$,@If, â€˜F@ 68@

66 2â€”@ @ @o0 â€˜@i@'.oo @ @o iâ€•.@oâ€˜r@â€˜0 2 @ c@ 64

@ 62 @ 2-@ ,@ 0 r@ 0@0@ @ o,__ @@ ,@. ,@. [email protected]@, C,) .a 60 V

,U @ 00 - 00 â€˜F@ @@@ â€˜@-e @.r@i â€˜r@m r-i 58 @ â€˜0â€˜0â€˜0 @â€˜0 @@ r- r-@r@r-i r-@r-@r-i r.4 e.i eI r.i r@4 r@i

45â€”4950-5455-59 60-64 65-69 @ r@râ€”@ â€˜0 r- r@@a AGE @ nte@_.@@ @r@;._:r@:@r.i â€” 0 â€˜0 (-.1 O@ @ (/)â€˜- nr@,r-1r.1e@, f@, @,r..ar.4 Chart 1. Comparison of body weightsof Nisei and Issei-Kibei men by â€˜O agegroups at 1stexamination. Valuesare shownas mean Â±S.E.

@ .@. r-@â€˜0 @ o,_.__ @ â€˜- a@o'oor-@ooo 00 r@- â€˜0t-.. ,@3

18 @u.@ @@@ .@ â€˜,@F-. @@Ã³r'io@@ @ <@. â€˜0â€˜0â€˜0 â€˜F@ I? 0 E

@ 0' â€˜0 r@ I 16 @ _ .@ ee â€” f@ â€˜,@â€˜I.0' @ L@2'@@' oa'a'ooooa' 0' 0' 00 (@-.

@ $5

.@ ,@ U @@@ .@ r'@ r'@ r-i C 0@ 00 00 0 â€˜1 @@ 0 OI@t@OO ,,@F-. 0' 0 â€˜4 @ â€” r.4 @4 0 @@ (@ @@-4r@[email protected] EN â€”

.â€˜ .@ â€˜3

C ,E @ c@.@ri o.'eoo.0.ri;@1 @ C @2ZZ 45-49 50-54 55-59 60-64 65-69 @@@@ c/) r-1r@r@a r@i (@â€˜ir-@ @4r.i AGE E Chart 2. Comparison of subscapularskinfold thicknessesof Nisei and Issei-Kibei men by age groups at 1st examination. Values are shown as t@, mean Â±SE. @ ,@ .0 .@ E â€˜r@ @o . o o@ 0@. â€˜I.râ€”@r @-.@2E 0@Ã³@O@OV,Ã³ @ U @â€˜-â€˜

U ii E @ â€¢i3cj â€˜0â€˜0â€˜0â€œ@â€œ@â€˜0

.@ >â€˜0 @ -.@ @o'.'1o.o.f@ 0 @O2â€”r@ ,@. â€” 0 00 00 â€˜@. @oâ€˜0â€˜0â€˜(@â€œ@â€˜0 I 00@- UZ @ @3V@ 0' C' 0' ,@0 @ 0 0 @ ,@. ,,@ @,O@O ,@. ,,@ @â€˜0'0

@ .!

@M , .@ 45-49 50-54 55-59 60-64 65-69 VU VU 0. ec@ ee AGE

(@2 @i; @n Chart 3. Comparison of serum cholesterol levels of Nisei and lssei .!@ l-@ ?;1 1- Kibei men by age groups at 1st examination. Values are shown as mean Â± z SE.

3262 CANCER RESEARCH VOL. 35

, , I I I I I I I I

2600 55

2400 50

2200 â€˜45 .@ C 2000@

$800 0I

$600 30 V

01@@ I I II 45-49 50-54 55-59 60:64 65-65 25 AGE

Chart 4. Comparison of total caloric intakes of Nisei and Issei-Kibei I I I 1@ men by age groups at 1stexamination. Values are shown as mean Â±S.E. - 45-49 50-54 55-59 60-64 65-69 AGE Chart 6. Comparisonofdietary sucroseintakesofNisei and Issei-Kibei men by age groups at 1stexamination. Values are shown as mean Â±SE. 40@

@ 35

0 .@ 0 @ 30

@ 25 0

=@ @ 20@ I-

$5 @1â€• 1' 45-49 50-54 55-@ 60-64 65-65 AGE Chart 5. Comparison of saturated fat intakes of Nisei and Issei-Kibei menby agegroupsat 1stexamination.ValuesareshownasmeanÂ±SE.

more apparent. In general, dietary differences between the Nisei and Issei-Kibei groups were greater for the intakes of particular foods than for nutrients. This may reflect uneven rates of acculturation to different specific food items. Thus, the

Nisei eat more meat and beef than the lssei-Kibei, but the â€˜ 2 3 4 5 ALMOST <2 TIMU 2-4 TIME ALMOST MO@ ThAN overall difference in animal protein and fat intakes is NEVEN P(R @EK @NN@K @ILY ON@CA DiV diminished by the absence of a similar difference in milk FREQUENCYCATEGORYATFIRST EXAM. and ice cream consumption. The heterogeneity found for Chart 7. Mean weekly frequenciesof intake of coffee, milk, tofu, and specific food items supports the usefulness of characterizing tsukudani at 3rd examination of Hawaii Japanese men according to the diet in this way. frequencies reported by the same persons at 1st examination, 6 years Various studies on diet a'nd cancer have hypothesized that previously. Values are shown as mean Â±SE. either specific nutrient intakes or food items with carcino genic potential may be associated with the development of cancer cases accumulate. Furthermore, if diet does play a neoplasms (1, 5, 7, 13, 14). Because data are available on role in oncogenesis, we may eventually observe significant both aspects of diet in this prospective study, the question differences in cancer incidence between the Nisei and can be investigated in both ways, as sufficient numbers of Issei-Kibei study populations in Hawaii.

NOVEMBER 1975 3263

2. Friend, B. Changes in Nutrients in the U.S. Diet Caused by Alterations in Food Intake Patterns. In: Proceedings. The Changing Food Supply in America Conference,Arlington, Va., May 22, 1974, pp. 1-13. Washington, D. C.: Food and Drug Administration, U. S. Dept. of Health, Education,and Welfare, 1975. â€˜5 3. Graham, S., Schotz, W., and Martino, P. Alimentary Factors in the RICE Epidemiology of Gastric Cancer. Cancer, 30: 927-938, 1972. I 4. Haenszel,W., and Kurihara, M. Studies of JapaneseMigrants. I. I- Mortality from Cancer and Other Diseases among Japanese in the United States. J. NatI. Cancer Inst., 40: 43â€”68,1968. 5. Haenszel, W., Kurihara, M., Segi, M., and Lee, R. K. C. Stomach z CanceramongJapaneseinHawaii. J. NatI. CancerInst., 49@969-988, 1972. 6. Hankin, J. H., Rhoads,G. G., and Glober, G. A. A Dietary Method for an Epidemiological Study ofGastrointestinal Cancer. Am. J. Clin. Nutr., 28: 1055-1060, 1975. 7. Hems, G. Epidemiological Characteristics of Breast Cancer in Middle and Late Age. Brit. J. Cancer,24: 226-234, 1970. 8. Kagan, A., Harris, B. R., Winkelstein, W., Johnson,K. G., Kato, H., Syme,S. L., Rhoads,G. G., Gay, M. 1., Nichaman, M. Z., Hamilton, H. B., and Tillotson, J. Epidemiologic Studies of Coronary Heart Disease and Stroke in Japanese Men Living in Japan, Hawaii and California: Demographic, Physical, Dietary and Biochemical Charac@ I 2 3 4 teristics. J. Chronic Diseases, 27: 345-364, 1974. @ ALMOST ONCE OR LE@ 2 TIMES 3 TItiV$ RIVER A DAY PERDi MOREPERDAY 9. Thorsteinsson, T. Polycyclic Hydrocarbons in Commercially and FREQUENCY CATEGORYAT F$RST EXAM. Home-Smoked Food in Iceland. Cancer, 23: 455-457, 1969. Chart 8. Mean weekly frequencies of intake of rice and green tea at 3rd 10. Tillotson, J. L., Kato, H., Nichaman, M. Z., Miller, D. C., Gay, M. examination of Hawaii Japanese men according to the frequencies L., Johnson, K. G., and Rhoads, G. G. Epidemiology of Coronary reportedby the samepersonsat IStexamination,6 yearspreviously.Values Heart Diseaseand Stroke in JapaneseMen Living in Japan, Hawaii, are shownas mean Â±SE. and California: Methodology for Comparison of Diet. Am. J. Clin. Nutr.,26:177â€”184,1973. 11. Wolff, I. A., and Wasserman, A. E. Nitrates, Nitrites, and Nitrosa Acknowledgments mines. Science, 177: 15-19, July 7, 1972. 12. Worth, R. M., and Kagan, A. Ascertainment of Men of Japanese The dietary data at the 1st examination were collected under the Ancestry in Hawaii through World War II Selective Service Registra supervision ofJeanne Tillotson ofthe National Heart and Lung Institute. tion. J. Chronic Diseases, 23. 389-397, 1970. 13. Wynder, E. L., Kajitani, T., Ishikawa, S., Dodo, H., and Takano, A. References Environmental Factors of Cancer of the Colon and Rectum. I. JapaneseEpidemiologicalData. Cancer,23: 1210-1220,1969. 1. Drasar, B. S., and Irving, D. Environmental Factors and Cancer of the 14. Wynder, E. L., Kmet, J., Dungal, N., andSegi, M. An Epidemiologi Colon and Breast. Brit. J. Cancer,27: 167-172, 1973. cal Investigation ofGastric Cancer. Cancer, 16: 1461â€”1496,1963.

3264 CANCER RESEARCH VOL. 35

Jean H. Hankin, Abraham Nomura and George G. Rhoads

Cancer Res 1975;35:3259-3264.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/35/11_Part_2/3259

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/35/11_Part_2/3259. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.