Diabetes Mellitus and Its Vascular Complications in ... - Diabetes Care

D

iabetes Mellitus and Its Vascular Complications in Japanese Migrants on the Island of Hawaii

RYOSO KAWATE, MICHIO YAMAKIDO, YUKIO NISHIMOTO, PETER H. BENNETT, RICHARD F. HAMMAN, AND WILLIAM C. KNOWLER

Japanese migrants and their offspring on the island of Hawaii and Japanese living in Hiroshima were examined for diabetes mellitus and its vascular complications. The same methods and investigators were used in both locations. Death certificates of Japanese and Caucasians dying on the island during the past 26 yr were analyzed. Diabetes, defined as a venous serum glucose concentration of at least 200 mg/dl 2 h after a 50-g oral glucose load, was significantly more common in the Hawaiian Japanese than in the Hiroshima Japanese subjects. This suggests that diabetes is more prevalent in Japanese in Hawaii than in Japan, although lack of knowledge about the total population of Japanese migrants in Hawaii makes this generalization uncertain. The proportion of deaths attributed to diabetes was much higher in Japanese migrants and their offspring in Hawaii than in Japan. During the 1950s, the proportional death rate from diabetes was about half as large in Japanese Hawaiians as in Caucasian Hawaiians, but it increased to become 1.6 times the Caucasian rate during the 1970s. A nutritional study revealed that the total caloric intake was similar in Japanese in Hawaii and Hiroshima, although the estimated level of physical activity was less in the Hawaiian subjects. Consumption of animal fat and simple carbohydrates (sucrose and fructose) were at least twice as high in Hawaiian as in Hiroshima Japanese. Conversely, Hiroshima Japanese consumed about twice the amount of complex carbohydrate as the Hawaiian Japanese. These observations support the hypothesis that a high fat, high simple carbohydrate, low complex carbohydrate diet and/or reduced levels of physical activity increase risk of diabetes. The proportion of deaths attributed to ischemic heart disease was higher in both diabetic and nondiabetic Japanese Hawaiians than in diabetic subjects in Japan. The rates were similar for Japanese and Caucasians in Hawaii. There was no evidence of an environmental influence on the development of microangiopathy (retinopathy) in diabetes, as the prevalence of diabetic retinopathy (stratified for diabetes duration) was similar in Japanese subjects in Hawaii and in Japan, and it was similar to previous reports from England. On the other hand, diabetes alone did not appear to account for the greater prevalence of macroangiopathy in Hawaiian Japanese than in Hiroshima. Thus environmental factors, possibly including diet, appear to be involved in the development of macrovascular complications of diabetes, DIABETES CARE 2.- 161-170, MARCH-APRIL 1979.

M

any pure Japanese migrants and their offspring

mellitus and the development of diabetic vascular compli-

live in the State of Hawaii. Because of great environmental differences, especially in dietary habits, between Hawaii and Japan, we studied the influence of environmental factors on diabetes

cations by investigating Japanese in Hawaii and those in Japan, A survey of diabetes and complications in Japanese in Hawaii was conducted in 1973 and 1976. A similar survey

DIABETES CARE, VOL. 2 NO. 2, MARCH-APRIL 1979

161

DM AND VASCULAR COMPLICATIONS IN JAPANESE IN HAWAII/RYOSO KAWATE AND ASSOCIATES

was conducted in Hiroshima in 1975 by the same investigators using the same methods. Death certificates were analyzed for Japanese and Caucasians on the island of Hawaii during the 26 yr from 1952 through 1977. This paper includes a comparison of nutrient intakes and physical activity levels as well as prevalence of diabetes mellitus and the frequency of its macrovascular complications between these two Japanese groups. BACKGROUND OF JAPANESE MIGRANTS AND THEIR OFFSPRING

The island of Hawaii is the largest (4020 square miles) of the Hawaiian Islands. The main industries are sugar cane, stock-farming, fishing, and the growing of coffee, macademia nuts, fruits, vegetables, and orchids. From the first Japanese migrations to Hawaii in 1868 until 1924, when migration was prohibited by law, 149,602 Japanese immigrated. Since immigration was permitted again in 1953, it is estimated that more than 20,000 Japanese have immigrated. According to the State census in I9601 the population on the island was 61,332 of which 44% were Japanese. Due to the subsequent increase in Caucasians, the population became 76,600 in 1976,2 and the proportion of Japanese was estimated to be approximately 30% of the total. Initially, Japanese migrants were engaged in heavy labor at sugar plantations. Japanese people were forced to stay in distressing circumstances during the Second World War. After the war, however, their living standard improved rapidly, and Japanese offspring have become active in politics, education, business, and the professions, with a resulting decline in their levels of physical activity. Physical activity levels were reduced in agricultural workers also, due to modern mechanization. Recently, many third-generation Japanese have intermarried with other races, but many Japanese over 40 yr of age have married only among those of Japanese ancestry. The majority of the first generation who migrated before 1924 have died, and the population of the fourth generation has been recently increasing. The major language in the first generation is Japanese, but members of subsequent generations speak English as their mother tongue, and their living style is Americanized. However, most of those who participated in our clinical survey were fluent in Japanese. SUBJECTS

The ancestors of most Japanese inhabitants of the island of Hawaii migrated from Hiroshima Prefecture, Japan, where many of their relatives still live. The subjects for this study were pure Japanese inhabitants of the island of Hawaii who were migrants or descendents of migrants from Hiroshima Prefecture. 18% were first generation (born in Japan), 71% were second generation, and 11% were third generation. Voluntary cooperation was obtained with the assistance of 162

the Hiroshima Kenjin-kai, an association of immigrants from Hiroshima Prefecture and their descendents in Hilo and Kona districts of the island. Diabetes prevalence in Japanese Hawaiians was determined in the 500 subjects examined in 1976. For comparison, diabetes prevalence was determined in 430 inhabitants of the three towns of Okimi, Chiyoda, and Oasa in Hiroshima Prefecture, Japan, in 1975. The major occupation in these towns was agriculture. In both countries, subjects were aged 40 through 96 yr. In addition, we examined 57 pairs of relatives, matched by sex and age to within 10 yr, with one member of each pair living in Hawaii and one in Hiroshima Prefecture. Most pairs were siblings; some were first cousins. Clinical characteristics related to vascular complications were compared in 106 diabetics who were examined in Hawaii in the surveys in 1973 and 1976, and in 351 diabetics examined during a survey in Hiroshima City from 1975 through 1977. All of these diabetic subjects were 40 yr of age or over. The prevalence of diabetic retinopathy was determined in 51 Japanese in Hawaii who were being treated with insulin or oral hypoglycemic agents. This was compared with the retinopathy prevalence in 696 diabetics attending the diabetes clinic of Hiroshima University. METHODS

T

he surveys of diabetes prevalence in Hawaii and Hiroshima were conducted by the same methods by the same investigators. Blood specimens were obtained fasting, 1 h and 2 h after a 50-g glucose load, in the morning after an overnight fast. Serum was separated within 1 h of venesection and frozen. Glucose and other chemistries were determined in the same laboratory in Hiroshima for both Hawaii and Hiroshima samples. Serum glucose was determined by the AutoAnalyzer. Diabetes was defined as a 2-h postload venous serum glucose concentration of at least 200 mg/dl. Serum cholesterol and triglycerides were determined in the fasting specimens by an enzyme method. Ischemic heart disease (IHD) was defined by electrocardiographic findings of either one or a combination of 1-1, IV-1 ~ 3, and V-l ~ 4 classified by the Minnesota Code. Both ocular fundi were examined through dilated pupils with a direct ophthalmoscope. Diabetic retinopathy was defined by the presence of microaneurysms or retinal hemorrhages. Proteinuria was determined by a dipstick (at least one plus) on a urine sample obtained at fasting. The dietary history was taken by two well-trained dietitians. Since the dietary history in this survey was intended to determine an average dietary intake over a long period, the method of Owada et al. 3 was employed to obtain specific information in regard to average amount and


DM AND VASCULAR COMPLICATIONS IN JAPANESE IN HAWAII /RYOSO KAWATE AND ASSOCIATES

frequency of the dietary intake. In Hawaii, the method was slightly modified to account for locally available food. Better results were observed by this method in 147 cases as compared with another method of 24-h recall of the intake on the preceding day.4

TABLE 1 Age and sex specific prevalence rates of diabetes* Hiroshima

Hawaii No.

r

i

Diabetes

or suujects

No.

%

25 41 71 75 212

1 5 12 12 30

4.0 12.2 16.9 16.0 14.2

No.

r i or SUD*

jects

Diabetes No.

%

RESULTS

Male (age in yr)

Frequency of Diabetes Mellitus Prevalence rates of diabetes. The results of the 1973 survey

40-49 50-59 60-69 70-96

26 37 50 75

1 1

3 5

3.8

2.7 6.0 6.7 5.3

in Hawaii were reported previously.5 The diabetic type of 188 10 All ages glucose tolerance test was more prevalent in Hawaii than in Hiroshima. The diabetic type was defined by the criteria of the Diagnostic Criteria Committee of the Japanese Diabetic Female (age in yr) 2 4.2 0 48 34 0 40-49 Society;6 i.e. as venous serum glucose levels above 185 mg/dl 29 2 66 9 13.6 6.9 50-59 at 1 h and above 150 mg/dl at 2 h after the glucose 85 6 13 12.9 101 7.1 60-69 challenge. In the present report, diabetes was defined as a 13 8 11.0 13.8 73 94 70-96 2-h postload serum glucose concentration at least 200 mg/dl, 242 21 32 11.1 All ages 288 8.7 regardless of the presence or absence of a previous diagnosis of diabetes. As shown in Table 1, the prevalence of diabetes Total (age in yr) 60 was higher in Hawaiian than in Hiroshima subjects. Of 500 3 73 1 40-49 4.1 1.7 66 3 13.1 4.5 50-59 107 14 Hawaiian subjects, 12.4% were diabetic compared with 7.2% 25 172 14.5 135 9 60-69 6.7 of 430 Hiroshima subjects. After stratification by age and sex, 20 70-96 148 169 18 13.5 10.7 this difference was statistically significant (P < 0.01), with 31 62 430 7.2 All ages 500 12.4 age-sex adjusted prevalence rate in Hawaii being 1.79 times that in Hiroshima. In the analysis of matched pairs of the Age-sex adjustedt 12.3 6.9 relatives living in Hawaii and Hiroshima, diabetes was defined as a 2-h glucose concentration at least 200 mg/dl or Diabetes prevalence was significantly higher in the Hawaiian than in the treatment with either insulin or oral hypoglycemic agents. Hiroshima subjects; summarizing on all age and sex strata,37 Xi2 = 6.82, Diabetes prevalence was 11/57 (19%) in Hawaii and 6/57 P < 0.01. (11%) in Hiroshima. There were 13 pairs discordant for * 2-h postload serum glucose at least 200 mg/dl, regardless of treatment for diabetes; nine in which the Hawaiian member and four in diabetes. which the Hiroshima member had diabetes. This difference t Direct standardization using the combined age and sex specific populations was in the same direction as the major study but was not of the Hawaii and Hiroshima samples as a standard. statistically significant. Table 2 shows diabetes prevalence stratified by age, sex, and obesity. After controlling for obesity, the difference Autopsy Cases published by the Japanese Society of Patholbetween locations persisted. The age-sex-obesity adjusted ogy, and the report10 on autopsy cases in Kyoto University. prevalence rate in Hawaii was 1.74 times that in Hiroshima The proportional mortality attributed to diabetes was 7.2% (P < 0.05). Obesity did not have a consistent relationship in Caucasians compared with 2.7% in Japanese in Hawaii to diabetes prevalence as seen in Table 2 and clearly did during 1950-1959. During 1960-1969, however, the races not account for the differences in diabetes prevalence. were similar, with diabetes attributed mortality of 5.7% in Mortality. With the cooperation of the Public Health Japanese and 6.1% in Caucasians. In 1970-1977, the rate in Office, Department of Health, in Hawaii County, death Hawaiian Japanese was 9.9%, exceeding the rate of 5.9% in certificates of pure Japanese and Caucasians were reviewed Caucasians on the island. The proportion of deaths atfor the 26 yr from 1952 through 1977. Since the racial tributed to diabetes in Hawaiian Japanese was obviously composition of the population on the island was unknown, high compared with that reported in Japan, as shown in mortality rates by race could not be compared. Instead, Table 3. The analysis of the deaths was based on death proportional mortality rates were computed as the number of certificates in which diabetes was not always listed even when deaths ascribed to diabetes divided by the total deaths. The it had been present, so their interpretation is limited. Hawaiian rates were compared with those reported in Japan; Nevertheless, the death certificate data are consistent with rates were calculated by the authors using the National the survey data showing a higher diabetes prevalence in Japanese Vital Statistics,8-9 Annual Reports of Pathological Hawaiian Japanese than in Japanese in Japan. DIABETES CARE, VOL. 2 NO. 2, MARCH-APRIL 1979

163


TABLE 2 Age, sex, and obesity specific prevalence rates of diabetes Hiroshima

Hawaii No. C

Male 40-59 yr BMlt < 2.0

2.0-2.5 >2.5 60-96 yr BM1 < 2.0 2.0-2.5 >2.5 Female 40-59 yr BM1 < 2.0 2.0-2.5 >2.5 60-96 yr BM1 < 2.0 2.0-2.5 >2.5

Diabetes

No.

%

jects

Diabetes

l

or suD' jects

No.

No.

%

6 35 25

0

0.0

11

1

2

5.7

4

16.0

47 5

1 0

20 88 38

3 13 8

15.0 14.8 21.1

52 63 10

3 4 1

5.8 6.3 10.0

29 58 27

0 6 5

0.0 10.3 18.5

17 36 10

0 2 0

0.0 5.6 0.0

35 104 35

3 9 9

Age-sex-obesity adjusted§

11.6

51 94 34

8.7 25.7

7 10 2

12.0

9.1 2.1 0.0

13.7 10.6 5.9

6.9

Diabetes prevalence was significantly higher in the Hawaiian than in the Hiroshiman subjects; summarizing over all age, sex, and obesity strata,37 X,2 = 4.89, P < 0.05. * 2-h postload serum glucose at least 200 mg/dl, regardless of treatment for diabetes. t BMI is the body mass index, or body weight (grams) divided by the square of height (centimeters). § Direct standardization using the combined age-, sex-, and obesity-specific populations of the Hawaii and Hiroshima samples as a standard.

Frequency of Vascular Complications Clinical observation. Clinical characteristics related to vascular disease were compared between 106 Hawaiian and 351 Hiroshima diabetic patients, aged at least 40 yr. As shown in Table 4, hypercholesterolemia, hypertriglyceridemia, and hypertension were more frequent in Hawaiian diabetic subjects than in Hiroshima diabetic subjects. The frequency of IHD was 27.4% in Hawaiian diabetic subjects and 20.5% in Hiroshima diabetic subjects, showing no significant difference. However, 9.4% of Hawaiian diabetic subjects had received medical care for IHD, significantly more than the 1.1% of Hiroshima diabetic subjects. There was no difference in the frequency of proteinuria between these two groups. Retinal examinations in 51 Hawaiian diabetic subjects were compared with examinations of patients attending the diabetes clinic at the Department of Internal Medicine, Hiroshima University. Stratified by duration of diabetes, retihopathy prevalence (Table 5) was similar in Japanese in Hawaii and Hiroshima and in English patients studied by Burditt et al. u Vascular diseases in deceased cases ivith diabetes. Vascular

diseases were analyzed from death certificates of diabetic subjects, both Japanese and Caucasian, on the island, and the results were compared with those reported by the Joslin Clinic12 and in Japan.13'14 IHD and cerebrovascular accident (CVA) were classified according to the International Classification of Disease. Diabetic nephropathy was defined as the presence of uremia or renal failure without documentation of urinary tract infection or malignant neoplasia of the urinary tract. Statistics in Japan were taken from Goto et al., 13 who reviewed clinical cases collected from 14 clinics in Japan during the period between 1952 and 1972, and from Hirata et al., 14 who collected information on 1885 diabetics dying at nonprivate hospitals throughout Japan during 1968-1970, by means of mail questionnaire. We also tabulated for comparative

TABLE 3 Proportion of death attributed to diabetes in Caucasians and Japanese on the Island of Hawaii and in Japan (per cent in parentheses) Japan Hawaii Years

Caucasians

Japanese

Vital statistics

Annual report (autopsy)

Okamotot (autopsy)

1952-1959 1960-1969 1970-1977

37/516 (7.2) 49/798 (6.1) 42/714 (5.9)

42/1529 (2.7) 95/1662 (5.7) 115/1164 (9.9)

2191/693089 (0.3) 5115/700237 (0.7) 9032/702281 (1.3)

97/19356 (0.5)§ 2232/174221 (1.3) 2515/143287 (1.8)"

12/1414 (0.8) 91/3954 (2.3)

* The proportion in 1955, 1965, and 1975. t Reference 10. § The proportion in 1958 and 1959. " The proportion between 1970 and 1975.

164



TABLE 5 Frequency of diabetic retinopathy in terms of duration of diabetes mellitus (percent in parentheses)

TABLE 4 Clinical findings in diabetic subjects of Japanese origin on the Island of Hawaii and in diabetic subjects detected in a survey in Hiroshima City Clinical findings

Hawaii

Hiroshima

Number of subjects Male: female Age in years (mean ± SD) Hypercholesterolemia* (%) Hypertriglyceridemiat (%) Hypertension§ (%) IHD (by ECG) (%) Treatment for IHD (%) Proteinuria (%)

106 43:63 67.0 ± 9.6 32.3 53.8 39.6 27.4 9.4 13.2

351 139:212 60.5 ± 10.6 21.9" 32.511

10-19 yr

20 yr

(4.0)

7/14 (50.0)

8/12

(66.7)

137/585 (23.4)

54/94 (57.4)

12/17 (70.6)

828/2775(29.8)

482/881 (54.7)

184/250(73.6)

Being treated for diabetes with insulin or oral hypoglycemic agents (regardless of serum glucose concentration). t Attending diabetes clinic at Hiroshima University. § Reference 11.

only 12.4% of diabetic patients at the Joslin Clinic. There were only small differences in the frequency of renal diseases in these three populations. In contrast to the similarities between proportional mortality rates due to vascular disease in Caucasians and Japanese in Hawaii and the Joslin Clinic population, vascular disease appeared to be less common in diabetic subjects in Japan, where its proportional mortality was only 47.3-51.0%. CVA was noted in 18.3-19.1% of the diabetic subjects in Japan, similar to the Hawaiian rates. In contrast, proportional mortality rates attributed to IHD were much lower in Japanese diabetic subjects, at 9.7-13.5% than the rate of 32.7% in the Hawaiian Japanese diabetic population. The rate of IHD in diabetic subjects in Japan was even lower than the rate of 20.3% in 4104 nondiabetic Japanese in Hawaii. The proportion of deaths attributed to renal failure was 10.6% in Hawaiian Japanese diabetic subjects, whereas it was 16.1-16.4% of clinically observed cases and 14-4% of the autopsy cases in Japan.

TABLE 6 Frequency of vascular disease in deceased diabetic subjects (percent)

Total Brain IHD Other cardiac Renal

Hawaiit

Joslin Clinic clinical diagnosis (Marks et al.*), 4097 cases

Caucasians, 128 cases

Japanese, 251 cases

Clinical diagnosis (Goto et a 1. §), 1917 cases

77.1 12.4 39.8 13.8 9.1

75.8 18.0 32.8 15.6 8.6

74.5 21.9

47.3 18.8

32.7}

i1 i

8.8 / 10.6

16.4

Japan clinical diagnosis (Hirata et al."), 1885 cases

Autopsyt cases, 4421 cases

51.0 18.3 9.7 6.9 16.1

48.8 19.1 13.5 1.8 14.4

* Reference 12. t Cases in Caucasians and Japanese on the Island of Hawaii and the autopsy cases in Japan were at least 40 yr of age. § Reference 13. 11 Reference 14.


165


TABLE 7 Mean nutrient intake (male) Ages (yr) 50-59

40-49 Nutrient No. of subjects Total energy (cal) Total protein (g) Animal origin (g) Vegetable origin (g) Total fat (g) Animal origin (g) Vegetable origin (g) Total carbohydrate (g) Simple (g) Complex (g)

Hawaii

Hiroshima

Hawaii

70+

60-69 Hiroshima

Hawaii

All ages

Hiroshima

Hawaii

Hiroshima

Hawaii

Hiroshima

23

26

44

37

70

50

73

75

210

188

2413 99 70 29 78 48 30 261 88 173

2495 81 41 40 49 22 27 381 33 348

2395 98 69 29 82

2638 83 42 41 53 21 32 397 35 362

2256 96 68 28 77 46 31 273 101 172

2282 76 37 39 42 19 23 347 38 309

1890 77 49 28 57 30 27 253 88 165

1971 68 34 34 35 16 19 315 36 279

2175 91 62 29 72 42 30 265 94 171

2258 74

47 35 284 97 187

Environmental Differences between Hawaii and Hiroshima Nutrient intakes. The Japanese migrants have maintained some Japanese diet habits by eating rice and fish cake or soy bean cake (tofu). Although many Japanese on the island of Hawaii eat rice at least once a day, the proportion of the total energy derived from rice was significantly smaller in Hawaii than in Japan.16 Younger people tend to eat food cooked with oil (more western style), whereas the older generation eats boiled food (more Japanese style) frequently. The results of the dietary survey including

37 37 43 19

24 349 36 313

average daily nutrient intake are summarized in Tables 7 and 8. The results of nutrient intake in more than 10,000 Japanese in Hiroshima, Nagasaki, Honolulu, and San Francisco investigated by Tillotson et al.17 during 1965-1968 are shown in Table 9. The average daily total energy intake was 2175 cal in Hawaiian male Japanese, 2258 cal in Hiroshima men, 1718 cal in Hawaiian Japanese women, and 1930 cal in Hiroshima women, showing little difference between Hawaiian and Hiroshima subjects. These estimates are similar to the results of Tillotson et al.17 Total calorie consumption was

TABLE 8 Mean nutrient intake (female) Ages (yr) 40-49 Nutrient No. of subjects Total energy (cal) Total protein (g) Animal origin (g) Vegetable origin (g) Total fat (g) Animal origin (g) Vegetable origin (g) Total carbohydrate (g) Simple (g) Complex (g)

166

Hawaii

60-69

50-59 Hiroshima

Hawaii

Hiroshima

Hawaii

All ages

70+ Hiroshima

Hawaii

Hiroshima

Hawaii

Hiroshima

51

34

67

29

92

84

65

94

275

241

1851 84 61 23

2221 78 41 37 53 20 33 345 35 310

1742 80

2090 67 32 35

1744 71

1968 70 35 35 41 17 24 322 34 288

1554 66 42 24 46 24 22 216 76 140

1742 61 31 30

1718

1930 67 34 33 41 17 24 314 38 276

74 42 32 210 81 129

54 26 62 34 28 222 83 139

44 18 26 346 31 315

46 25 52 25 27 219 85 134


35 16 19 286 42 244

75 50 25 57 30 27 218 82 136

DM AND VASCULAR COMPLICATIONS IN JAPANESE IN HAWAII /RYOSO KAWATE AND ASSOCIATES

TABLE 9 Mean nutrient intakes by 24-h recall at examination: men of Japanese ancestry living in Japan (J), Hawaii (H), and California (C) (Tillotson*) Age (yr) i

No. of subjects Total energy (cal) Percent! energy Total protein Total fat Total carbohydrate Alcohol Saturated fatty acids Polyunsaturated fatty acids Animal protein (%)§ Vegetable protein (%)§ Saturated fat (%)§ Unsaturated fat (%)§ Simple carbohydrate (%)§ Complex carbohydrate (%)§ Dietary cholesterol (mg) Estimated sodium (g)

45-49

W-44

J

C

J

138 2378

79 2358

333 2249

14. 1 16. 5 61..0 10..2

16.0 39.3 42.7 2.4 15.2

14.5 15.8 62.0 9.0

55-59

50-54

65-69

60-64

J

H

C

J

H

C

J

H

C

J

H

C

1832 85 2475 2354

437 2260

2790 2355

38 2153

449 2152

1593 2241

39 2321

481 2055

1337 2024

15 2183

437 1978

451 1816

11 1916

16.6 34.8 44.9 3.8 12.8

15.8 38.8 43.3 2.6 14.4

14.5 15.7 61.8 9.5

16.7 34.0 45.3 4.1 12.5

17.2 39.8 41.2 2.0 15.0

14.1 15.3 62.5 9.6

16.8 32.6 47.1 3.5 12.0

16.6 34.6 46.0 3.2 12.8

14.1 14.6 64.6 7.8

16.7 31.6 48.6 3.2 11.5

16.2 34.6 48.7 0.6 12.4

16.8 29.5 50.2 3.6 10.8

15.4 34.5 46.4 3.6 12.9

H

C

14.5 14.1 64.7 7.7

49..7 50..3 42..2 57. 8

5.8 74.2 25.8 71.6 28.4

51.4 48.6 39.9 60.1

6.3 74.5 25.5 68.6 31.4

6.3 74.8 25.2 69.7 30.3

49.2 50.8 41.2 58.8

6.1 74.1 25.9 69.1 30.9

6.1 74.9 25.1 68.7 31.3

49.6 50.4 41.6 58.4

5.9 72.7 27.3 68.5 31.5

6.0 71.8 28.2 67.5 32.5

48.0 52.0 41.3 58.7

5.8 70.1 29.9 67.0 33.0

6.0 69.3 30.7 66.8 33.2

48.1 51.9 41.0 59.0

5.3 68.7 31.3 67.1 32.9

5.5 73.5 26.5 77.5 22.5

18. 5

39.7

18.2

36.1

37.5

18.9

35.9

36.1

19.5

34.7

39.4

18.5

32.7

35.7

19.5

31.8

42.2

81. 5 499

60.2 526 3.3

81.8 503 4.9

63.9 580 3.1

62.5 519 3.1

81.1

64.0 568 3.0

63.9 592 3.1

80.5

65.3 529 2.8

60.6 543 3.3

81.5

67.3 502 2.6

64.3

80.5 418

68.2

57.8

445

4.4

2.3

576 2.3

4. 7

489 4.8

454 4.6

434 4.5

441 3.0

* Reference 17. t Percentages do not necessarily total 100 because of rounding. 8 Derived from intercenter food grouping system.

slightly higher in Hiroshima, but this can be accounted for fat. Thus vegetable fat was the major source in Hiroshima, by the difference in physical activity levels which will be and animal fat was the major source in Hawaii. described below. The average intake of carbohydrate was 265 and 349 g in The average intake of total protein was 91 g in Hawaiian Hawaiian and Hiroshima men, respectively. In women it was men compared with 74 g in Hiroshima men, and 75 g and 67 g 218 and 314 g, respectively. The proportion of energy derived in Hawaiian and Hiroshima women, respectively. The from carbohydrate was 48.7 and 61.8% in Hawaiian and proportion of total energy from protein was 16.7 and 13.1% in Hiroshima men, and 50.7 and 65.1% in Hawaiian and Hawaiian and Hiroshima men, respectively, and it was 17.5 Hiroshima women, respectively, similar to the results and 13.7% in Hawaiian and Hiroshima women, respec- of Tillotson et al.17 Although Hiroshima subjects contively. These proportions were similar to those of 16.6-16.8% sumed more carbohydrate than the Japanese Hawaiians; in Honolulu Japanese and 14.1-14.5% in Japanese in consumption of simple carbohydrates, including sugar Japan in all age groups reported by Tillotson et al.17 In and fructose, was more than twice as great in Hawaii both Hiroshima men and women, the amounts of protein of as in Hiroshima; it was 95 and 36 g in Hawaiian animal and vegetable origins were approximately equal, but and Hiroshima males, and 82 and 38 g in Hawaiian and the proportion of animal protein was 68 and 67% in Hiroshima women, respectively. Conversely, Hawaiian subHawaiian men and women, respectively, similar to the jects consumed only about half the amount of complex estimates of Tillotson et al.17 carbohydrates as the Hiroshima subjects. In summary, total energy consumption was similar in The total fat intake was 72 g in Hawaiian men compared with 43 g in Hiroshima men. In women, it was 57 g and Hawaii and Hiroshima, but the Hawaiian subjects con41 g, respectively. The Hawaiian Japanese and Hiroshima sumed about twice as much animal fat and simple carbohysubjects consumed similar amounts of vegetable fat, whereas drates and about half the quantity of complex carbohydrates. the Hawaiian Japanese consumed about twice as much animal Physical activity levels. Since many of the subjects were


167


retired, occupational history was divided into the preretirement and the present occupations. About half of the Hawaiian males had been engaged in agricultural or manual labor, and the rest of the male subjects were engaged in such occupations as managerial, technical, office, or sales work. Approximately 40% of Hawaiian women were housewives, 20% were laborers and about 30% were engaged in technical, sales, or service businesses. In the Hiroshima group, 87% of the male and 86% of the female subjects were engaged in agricultural or manual labor, and many of the older subjects had continued light work. The levels of physical activity, classified according to the criteria of occupational classification tabulated by the Japanese Ministry of Health and Welfare,18 are shown in Table 10, which indicates that the Hiroshima subjects were engaged in heavier work than the Hawaiian subjects, for both past and present activity.

DISCUSSION

A

comparative medical study was conducted in Japanese subjects over 40 yr of age living on the island of Hawaii and those in Hiroshima, Japan, using the same methodology in both locations. The prevalence of diabetes, defined by serum glucose levels above 200 mg/dl 2 h after a 50 g oral glucose challenge, was estimated as 12.4% in Hawaiian subjects and 7.2% in Hiroshima subjects, suggesting a higher prevalence in Hawaiian Japanese than in Hiroshima. A similar differ-

TABLE 10 Levels of physical activity at present and in the past Hawaii Male Physical activity Past Very light Light Moderate Heavy Unknown Total Present Very light Light Moderate Heavy Unknown Total

168

No.

27 64 80 2 2 175

86 45 42 0 2 175

Hiroshima Female

Male

Female

%

No.

%

No.

%

No.

%

15.4 36 .6 45 .7 1.1 1.1

38 173 40 0 3

15.0 68.1 15.7 0 1.2

8 18 36 126 0

4.3 9.6 19.1 67.0 0

2 42 196 1 0

0.8 17.4 81.3 0.4 0

100

49 .1 25 .7 24 .0 0 1.1 100

254

81 154 19 0 0 254

100

31.9 60.6 7.5 0 0 100

188

31 67 52 38 0 188

100

16.5 35.6 27.6 20.2 0 100

241

41 126

74 0 0 241

100

17.0 52.3 30.7 0 0 100

ence was observed in a comparison of 57 age- and sex-matched pairs of relatives; one member in each pair living on the island and another living in Hiroshima. These findings suggest a higher diabetes prevalence in Japanese in Hawaii than in Hiroshima, although this interpretation must be made with caution, as the target population of all Japanese migrants and their descendents was unknown and hence not randomly sampled. Further evidence for higher diabetes prevalence in Hawaii is provided by the proportional mortality rates attributed to diabetes, which were higher in Japanese on the island of Hawaii than among Japanese in Japan. Dietary differences between Hiroshima and Hawaii may account for some of the difference in diabetes prevalence. The role of diet in the development of diabetes has not been clearly defined. Some investigators proposed that a high fat and low carbohydrate diet causes precipitation of diabetes, whereas others propose that there is a deleterious effect of excessive sugar intake. Kageura,19 Hales et al., 20 and Yamada et al.21 reported that deterioration of glucose tolerance was caused by high fat and low carbohydrate intakes. Sakai22 described a decrease in the number of insulin receptors in the rat liver following a high fat diet. An extensive epidemiologic study by Himsworth23 compared nutrient intake and diabetes prevalence in many countries in the world and in both urban and rural areas. Nutrition and diabetes prevalence were compared over a 30-yr period in some countries. Nutritional changes were related to diabetes prevalence during the First World War. The investigator concluded that low carbohydrate and high fat diets were risk factors for diabetes. Cohen et al.24 compared recent Jewish migrants to Israel with Jewish migrants from Yemen who had lived in Israel at least 25 yr. There were no differences in the consumption of fat and protein, but daily sugar consumption was 6.6 ± 1.0 g in the newly migrated group and 63 ± 6 g in the previously migrated Jewish group. The authors speculated that this dietary difference might be responsible for the higher diabetes prevalence in the previously migrated group. Campbell23 emphasized the risk of excessive sugar intake from observations of Indians who migrated to Durban in Africa compared with Indians living in India, and from observations of Zulus living in urban and rural areas. Diabetes prevalence increased in Iceland when the consumption of fat and protein decreased and purified carbohydrate consumption increased.25 Schaeffer26 reported an increase in diabetes prevalence in Eskimos with increased sugar consumption. Despite these reports suggesting a relationship between diet and diabetes prevalence, several studies including those by Mouratoff et al., 27 Poon-King et al., 28 and Prior et al.29 were unable to demonstrate such an effect. Diabetes prevalence is extremely high in Pima Indians whose diet is high in carbohydrates, yet low in sucrose.30 Our survey revealed lower physical activity levels in



Hawaiian Japanese than in Hiroshima subjects. Total caloric intake was similar in these two groups, but the differences in nutrient composition were substantial. In Hawaiian Japanese, intakes of animal fat and simple carbohydrate were strikingly high whereas that of complex carbohydrate was quite low, consistent with our previous hypothesis that the high prevalence of diabetes in Hawaiian Japanese was due to the diet patterns of high animal fat, high simple carbohydrate, and low complex carbohydrate intakes as well as low physical activity levels.5 Diabetes prevalence appears to increase greatly during rapid progress from primitive to developed environments. For instance, American Indians,31'32 Polynesians,29'31'33 Nauruans,34 Micronesians,31 migrants from the Middle East or North Africa to Israel,33 and Indians in South Africa25 (all of whom have undergone recent environmental changes) have higher diabetes prevalence rates than Caucasians. Similarly, the Japanese in Hawaii were laborers and in poor economic condition before 1950. Since then, however, their living standards have improved rapidly. Now diabetes appears to be more prevalent than in Japanese in Hiroshima. Furthermore, as evidenced by death certification, this change apparently took place during the period since the Second World War when their living standards have been improving. Although many such populations who have undergone rapid environmental changes now have high diabetes prevalence rates, it is unclear exactly what dietary or other environmental change, if any, is common to all of them, and hence might be responsible for the increase in diabetes. The frequencies of the different types of vascular complications seen in Hawaiian Japanese were greatly different from those in Japanese in Japan. The frequency of diabetic retinopathy stratified by duration of diabetes was similar in Hawaii to that observed in Hiroshima University Clinic (the authors' diabetes clinic) and that reported in England.11 This suggests that the development of microvascular complications was not much influenced by environmental factors such as diet and physical activity, but was mainly influenced by diabetes mellitus per se. On the other hand, the proportional mortality attributed to IHD in Hawaiian Japanese diabetic subjects was considerably higher than that in Japanese diabetic. subjects in Japan and was as high as that in Caucasian Hawaiians. This suggests that, beyond the effect of diabetes alone, environmental factors influence the development of IHD and is consistent with the finding of less IHD in Pima Indians than in Caucasians, despite the extremely high diabetic prevalence in Pimas.36 Thus, environmental factors, such as diet, appear to influence the development not only of diabetes but of macrovascular complications such as IHD. In contrast, microvascular complications, such as retinopathy, are strongly related to duration of diabetes but not to any environmental factors yet identified.

ACKNOWLEDGMENTS: We are indebted to the members of the

Hiroshima Kenjin-Kai organizations of Hilo and Kona, Hawaii, for their participation in this study. Mr. J. M. Hirano and Mr. M. Inaba were especially helpful in ensuring their cooperation. Dr. F. Ishibashi, Dr. K. Nakano, Ms. F. Murakami, and Ms. Y. Ebitani was responsible for much of the data collection. The study was supported in part by the Overseas Research Grant provided by the Japanese Ministry of Education. From the Department of Internal Medicine, Hiroshima University School of Medicine, Hiroshima, Japan. P. H. Bennett's and W. C. Knowler's present address is National Institute of Arthritis, Metabolism, and Digestive Disease, Phoenix, Arizona. R. F. Hamman's present address is Department of Preventive Medicine, University of Colorado, Denver. REFERENCES 1

Matayoshi, J.: Statistical observation on mortality among the Japanese on the Island of Hawaii. Tohoku Igaku Zasshi 78: 62—162, 1969. 2 Department of Health, State of Hawaii: Statistical Report 1976, p. 7. 3 Owada, K., Date, C , Tanaka, H., Ueda, U., and Tsue, H.: A trial simple method of diet survey. J. Clin. Nutr. (Jpn.) 45: 343-351, 1976. 4 Ebitani, Y., Kawate, R., and Nishimoto, Y.: Comparative study on the method of epidemiological survey on the diet. Transactions on the 32nd Japanese Congress on Food and Nutrition. 1978, p. 12. 5 Kawate, R., Miyanishi, M., Yamakido, M., and Nishimoto, Y.: Preliminary studies of the prevalence and mortality of diabetes mellitus in Japanese in Japan and on the Island of Hawaii. In Advances in Metabolic Disorders. Vol. 9. Levine, R., Luft, R., Miller, M., and Bennett, P. H., Eds. New York, Academic Press, 1978, pp. 200-224. 6 Kuzuya, N., Abe, M., Ueda, H., Kuzuya, K., Kuzuya, T., et al.: Report of diagnostic criteria for diabetes mellitus in glucose tolerance test. (Recommendation for the criteria for glucose tolerance test in the diagnosis of diabetes mellitus). J. Japan Diabetic Soc. 13: 1-7, 1970. 7 Minowa, S., Takahashi, H., Mayuzumi, N., and Miyashita, H.: Studies on the standard body weight in adults. Supplement: Figures of calculation on weight. Jpn. Med. J. 1988: 24-29, 1962. 8 Health and Welfare Statistics Association: Kosei no Shihyo 3(8): 89, 1956; 13(11): 176, 1966; 23(9): 342, 1976. 9 Health and Welfare Statistics and Information Department, Minister's Secretariat: Vital Statistics 1975, Japan. Vol. 1. Health and Welfare Assoc. Tokyo, p. 116, 1975. 10 Okamoto, K., Hazama, F., and Yamasaki, Y.: Pathology of diabetes mellitus in Japan. In Diabetes Mellitus in Asia. Tsuji, S., and Wada, M., Eds. Internatl. Congr. Ser. No. 221, Amsterdam, Excerpta Med. Found., 1971, pp. 106-119. 11 Burditt, A. F., Caird, F. I., and Draper, G. J.: The natural history of diabetic retinopathy. Q. J. Med. 37: 303-317, 1968. 12 Marks, H. H., and Krall, L. P.: Onset, course, prognosis


169


and mortality in diabetes mellitus. In Joslin's Diabetes Mellitus. Marble, A., White, P., Bradley, R. F., and Krall, L. P., Eds. 7th ed., Philadelphia, Lea and Febiger, 1971, pp. 225. 13 Goto, Y., Sato, S., and Masuda, M.: Causes of death in diabetics and its vascular complications. Sogorinsho 22: 779—786, 1973. 14 Hirata, Y., Mihara, T., Sakasegawa, K., et al.: Statistical studies of 1,885 death cases of diabetes in Japan. The relation between causes of death and therapeutic methods of diabetes. J. Jpn. Diabetic Soc. 16: 290-299, 1973. 15 Health and Welfare Statistics Association: Kosei no Shihyo 24(9): 356, 1977. 16 Kawate, R.: Diabetes mellitus observed in Japanese on the Island of Hawaii. In 1978 Diabetology. Kosaka, K., Ed. Tokyo, Shindan to Chiryo Co. Ltd., 1978, pp. 88-108. 17 Tillotson, J. L., Kato, H., Nichaman, M. I., et al.: Epidemiology of coronary heart disease and stroke in Japanese men living in Japan, Hawaii and California; Methodology for comparison of diet. Am. Clin. Nutr. 26: 117-124, 1973. 18 Matsushima, S.: In Current Status of Physical Strength in Japan. Matsushima, S., Ed. Tokyo, Daiichi Houki Shuppan Co. Ltd., 1970, pp. 171. 19 Kageura, J.: Treatment of diabetes mellitus. J. Jpn. Soc. Intern. Med. 40: 443-461, 1951. 20 Hales, C. N., and Randle, P. J.: Effects of low-carbohydrate diet and diabetes mellitus on plasma concentrations of glucose, nonesterified fatty acid, and insulin during oral glucose-tolerance tests. Lancet I: 790-794, 1963. 21 Yamada, K., and Katsumata, K.: Studies on the impaired carbohydrate metabolism induced by high fat diet. J. Jpn. Diabetic Soc. 9; 112-117, 1966. 22 Sakai, I.: Studies o n insulin receptor. II: T h e effect of fasting and high fat diet o n insulin receptors of guinea pigs a n d rats. Med. J. of Hiroshima University 1979. In press. 23 Himsworth, H . P . : Diet a n d incidence of diabetes mellitus.

Clin. Sci. 2: 117-148, 1935-1936. 24

Cohen, A. M., Bavly, S., and Poznanski, R.: Change of diet of Yemenite Jews in relation to diabetes and ischemic heart disease. Lancet 2: 1399-1401, 1961. 25 Campbell, G . D . : Diabetes in Asians a n d Africans in a n d around Durban. S. Afr. Med. J. 37: 1 1 9 5 - 1 2 0 8 , 1963.

170

26

Schaeffer, O . : W h e n t h e Eskimo comes to town. Nutrition

Today 6: 8-16, 1971. 27 Mouratoff, G. J., Caroll, N . V., and Scott, E. M.: Diabetes mellitus in Eskimos. J . A . M . A . 199: 9 6 1 - 9 6 6 , 1967. 28 Poon-King, T . , Henry, M. V., and Rampersad, F.: Prevalence and natural history of diabetes in Trinidad. Lancet 1: 1 5 5 - 1 6 0 , 1968. 29 Prior, A . M . , Beaglehole, R., Davidson, F., a n d Salmond, C. E.: T h e relationships of diabetes, blood lipids, a n d uric acid levels in Polynesians. In Advances in Metabolic Disorders, Vol. 9. Levin, R., Luft, R., Miller, M . , and Bennett, P. H . , Eds., New York, Academic Press, 1978, pp. 2 4 2 - 2 6 1 . 30 Reid, J. M . , Fullmer, S. D . , Pettigrew, K. D . , et al.: N u trient intake of Pima Indian women: relationships to diabetes mellitus a n d gallbladder disease. A m . J. Clin. Nutr. 24: 1281 — 1289, 1971. 31 West, K. M . : Diabetes in American Indians and other native populations of the new world. Diabetes 2 3 : 8 4 1 - 8 5 5 , 1974. 32 Bennett, P. H., Burch, T. A., and Miller, M.: Diabetes mellitus in American (Pima) Indians. Lancet 2: 125-128, 1971. 33 Sloan, N. R.: Ethnic distribution of diabetes mellitus in Hawaii. J.A.M.A. 183: 419-424, 1963. 34 Zimmet, P., and Taft, P.: The high prevalence of diabetes mellitus in Nauru, a central pacific island. In Advances in Metabolic Disorders. Vol. 9. Levine, R., Luft, R., Miller, M., and Bennett, P. H., Eds. New York, Academic Press, 1978, pp. 225-241. 35 Medalie, J. H., Herman, J. B., Goldbourt, U., and Papier, C. M.: Variations in incidence of diabetes among 10,000 adult Israeli males and the factors related to their development. In Advances in Metabolic Disorders, Vol. 9. Levine, R., Luft, R., Miller, M., and Bennett, P. H., Eds. New York, Academic Press, 1978, pp. 93-110. 36

Ingelfinger, J. A . ,

Bennett,

P. H . , L i e b o w , I. M . , et a l . :

Coronary heart disease in the Pima Indians: electrocardiographic findings and postmortem evidence of myocardial infarction in a population with a high prevalence of diabetes mellitus. Diabetes 25: 561-565, 1976. 37 Mantel, N., and Haenszel, W.: Statistical aspects of the analysis of data from retrospective studies of disease. J. Natl. Cancer Inst. 22: 719-748, 1959.
