Caloric Restriction, the Traditional
Okinawan Diet, and Healthy Aging
The Diet of the World’s Longest-Lived People
and Its Potential Impact on Morbidity
and Life Span
BRADLEY J. WILLCOX,a,bD. CRAIG WILLCOX,a,cHIDEMI TODORIKI,d
AKIRA FUJIYOSHI,bKATSUHIKO YANO,aQIMEI HE,a
J. DAVID CURB,a,bAND MAKOTO SUZUKIe
aPacific Health Research Institute, Honolulu, Hawaii 96813, USA
bDepartments of Geriatric Medicine and Medicine, John A. Burns School of
Medicine, 1356 Lusitana Street, 7F, Honolulu, Hawaii 96813
cCollege of Nursing, Okinawa Prefectural University, 1-24-1 Yogi, Naha,
Okinawa, Japan 902-0076
dDepartment of Environmental and Preventive Medicine, Faculty of Medicine,
University of the Ryukyus, 207 Uehara, Nishihara-cho, Okinawa,
eFaculty of Medicine, University of the Ryukyus, 207 Uehara, Nishihara,
Okinawa, Japan 903-0215 and Okinawa Research Center for Longevity Science,
1-27-8 Ahacha, Urasoe, Okinawa, Japan 901–2114
ABSTRACT: Long-term caloric restriction (CR) is a robust means of re-
ducing age-related diseases and extending life span in multiple species,
but the effects in humans are unknown. The low caloric intake, long life
expectancy, and the high prevalence of centenarians in Okinawa have
been used as an argument to support the CR hypothesis in humans.
However, no long-term, epidemiologic analysis has been conducted on
traditional dietary patterns, energy balance, and potential CR pheno-
types for the specific cohort of Okinawans who are purported to have
had a calorically restricted diet. Nor has this cohort’s subsequent mor-
tality experience been rigorously studied. Therefore, we investigated six
decades of archived population data on the elderly cohort of Okinawans
(aged 65-plus) for evidence of CR. Analyses included traditional diet
composition, energy intake, energy expenditure, anthropometry, plasma
DHEA, mortality from age-related diseases, and current survival pat-
terns. Findings include low caloric intake and negative energy balance
at younger ages, little weight gain with age, life-long low BMI, relatively
Address for correspondence: Bradley J. Willcox, M.D., Pacific Health Research Institute, 846 South
Hotel Street, Suite 301, Honolulu, HI 96813. Voice: 808-524-4411; fax: 808-524-5559.
Ann. N.Y. Acad. Sci. 1114: 434–455 (2007). C
2007 New York Academy of Sciences.
WILLCOX et al. 435
high plasma DHEA levels at older ages, low risk for mortality from age-
related diseases, and survival patterns consistent with extended mean
and maximum life span. This study lends epidemiologic support for phe-
notypic benefits of CR in humans and is consistent with the well-known
literature on animals with regard to CR phenotypes and healthy aging.
KEYWORDS: caloric restriction; longevity;Okinawa; DHEA; biomarker;
human; healthy aging; life span
Energy restriction, also known as caloric restriction (CR), is the most robust
and reproducible means of reducing age-related diseases and extending life
span in short-lived animals, but the effects in humans are unknown.1–7Prelim-
inary data from ongoing research with CR on long-lived nonhuman primates
(rhesus and squirrel monkeys) also suggest potential health benefits—although
it will take some years before final results become available and firmer con-
clusions can be reached regarding aging and life span.8,9
Some epidemiologic and short-term human studies support CR-related
health benefits.9–13 However, a role for CR in human aging is difficult to
ascertain since human life span makes long-term investigations impractical
and there are no universally accepted biomarkers to measure the rate of hu-
man aging.14 Whether CR (without malnutrition) affects human aging may be
among the most significant unanswered questions in modern biogerontology.9
Yet we have few available human populations or studies that can address this
question. Most human populations who have experienced low calorie intake
have suffered from high mortality due to infectious diseases and malnutrition.
Studies of human CR volunteers are currently under way with promising
early results,12,15,16 but will be unable to address longevity as an outcome.
Only one long-term (>30 years) epidemiologic study has linked CR to human
longevity.17 This 36-year follow-up study reported a weak trend for lower
all-cause mortality in healthy never-smoking Japanese–American men whose
caloric intake was 15% lower than the cohort average, suggesting that those
who maintained a modestly low energy intake (mean kcal/day 1882, range
1705–2061) in mid-life had the lowest late-life mortality risk. There was higher
risk for mortality when caloric intake dropped below 50% of the group mean.
These data are consistent with previous findings in animals, but much more
study is required of healthy human populations with low caloric intake in order
to understand the effects of CR on human aging and life span.
Therefore, it is of significant interest that Kagawa18 reported low caloric
intake in the Okinawan population, relative to other Japanese, and hypothe-
sized that this may have been, at least in part, responsible for their healthy
longevity. Kagawa18 reported dietary information from the 1972 Japan Na-
tional Nutrition survey that suggested Okinawan adults consumed only 83%
436 ANNALS OF THE NEW YORK ACADEMY OF SCIENCES
of the Japan average caloric intake. This report followed an earlier report by
Hokama et al.19 that Okinawan school children consumed only 62% of the
calories of other Japanese school children in the early 1960s. Kagawa18 also
presented anthropometric and morbidity data from selected small samples of
older Okinawans that were consistent with CR and reported markedly lower
risk for age-associated diseases in middle-aged Okinawans (60–64 year olds)
than in other Japanese. A later study by Chan et al. also reported dietary and
phenotypic data in Okinawan septuagenarians and centenarians consistent with
There are several weaknesses with these previous reports. One, to our knowl-
edge no population-based dietary information has been reported in a peer-
reviewed journal on Okinawan adults before the 1972 National Nutrition Sur-
vey. Since the Japanese lifestyle underwent radical changes from the 1950s,18
including changes in food choices, caloric intake, and energy expenditure, it is
unlikely that the 1972 Japan National Nutrition Survey reflects the traditional
CR diet that may be implicated in Okinawan longevity. Two, Okinawans are
smaller than other Japanese (and Americans) and likely require fewer calo-
ries. Therefore, accounting for their unique energy requirements is necessary
before concluding that Okinawans were calorically restricted. For example,
Okinawans currently consume 8% fewer calories than other Japanese,21 yet
they now have the largest body mass index (BMI) in Japan.22 Three, since
the Okinawan mortality advantage has all but disappeared except in older co-
horts (aged 65-plus),22,23 it would be informative to have a more detailed,
population-based epidemiologic analysis of the traditional diet, energy intake,
energy expenditure, phenotype, and the subsequent mortality experience of
this older cohort. These data might help answer the question as to whether
Okinawans were truly calorically restricted and to what degree, the phenotypic
consequences, and the current consequences for age-related mortality and life
span. Importantly, an in-depth epidemiologic analysis of this older cohort using
longer-term population data might also provide significant new information
on the potential human effects of CR.
Fortunately, detailed population surveys of the traditional Okinawan diet and
anthropometry were conducted in 1949.24 Smaller, more limited population
surveys were conducted at irregular intervals thereafter by U.S. and Okinawan
postwar administrations.25 These surveys were archived and have not been
systematically analyzed. Therefore, we investigated these data in detail and
asked three questions: (1) were older Okinawans truly calorically restricted
across the life course? (2) is there longer-term anthropometric evidence to
support this hypothesis? and (3) is there evidence for slower aging in this
To address question (1), we assessed population food-consumption patterns,
energy intake, and energy expenditure for older Okinawans (aged 70s) at two
different times—when they were young adults (aged in their 30s) and when they
were middle-aged (aged 50s). To answer question (2), we analyzed archived
WILLCOX et al. 437
data on height and weight, and calculated BMI for multiple time points in the
Okinawan adult population coincident with whole-population energy balance.
To answer question (3), whether delayed aging has occurred in this cohort,
we took three approaches. First, since the adrenal steroid dehydroepiandros-
terone (DHEA) has been purported to satisfy the main criteria for a candidate
biomarker of aging in nonhuman primates,26 and is of hypothetical utility as a
biomarker of the rate of human aging,27 we measured plasma levels in a sample
of older Okinawans (aged 70s). Second, we analyzed late-life survival patterns
for potential evidence of extended average and maximum life span. Third, we
calculated age-adjusted mortality patterns standardized to the world standard
population for age-related diseases. Finally, we compared selected findings to
published data from non-CR Americans and Japanese of similar chronological
age, in whom differences in diet, energy balance, anthropometry, DHEA, age-
related mortality and late-life survival should be robust, if older Okinawans
truly experienced long-term CR.
Study Materials and Measurements
Population Dietary Intake and Energy Expenditure
Archived dietary data are used to estimate energy balance and potential
CR status in Okinawan septuagenarians at younger ages. Okinawan data were
derived from the Office of the Civil Administrator of the Ryukyu Islands
(Okinawa) for the year 1949, when the current septuagenarians were aged
approximately 30 years,24 and approximately every 5–10 years thereafter by
the Okinawa prefectural government.25 Studies relied on 3-day food records
of usual dietary intake to estimate caloric intake in conjunction with dietitian
For the U.S. population, data collected with comparable methodology for the
same birth cohort of Americans (circa 1915–1925 birth cohort) is extremely
limited. The closest data set appears to be the NHANES I data set, which used
24-hour dietary recall methods to estimate dietary intake in adult subjects (aged
20–74 years) in 1971–1974.28 We could find no comparative data set before
the 1970s for the U.S. population.
Archived data of anthropometric measurements (height and weight) and
demographic data (age, occupation) in the Okinawan population and in the
NHANES population were used for estimation of energy expenditure using the
Harris–Benedict equation.29 This equation estimates basal or resting metabolic
rate (BMR). An activity factor was then used to calculate additional calories
burned according to the following levels of activity:
Sedentary =BMR ×1.2 (little or no exercise, desk job)
438 ANNALS OF THE NEW YORK ACADEMY OF SCIENCES
Lightly active =BMR ×1.375 (light exercise/sports 1–3 days/wk)
Moderately active =BMR ×1.55 (moderate exercise/sports 3–5 days/wk)
Very active =BMR ×1.725 (hard exercise/sports 6–7 days/wk or physical
Extremely active =BMR ×1.9 (hard daily exercise/sports and physical
Energy balance was estimated in the Okinawans and Americans by subtract-
ing estimated energy expenditure from caloric intake measured in the dietary
surveys. BMI was also calculated by body weight (kg)/height (m)2and used
as an additional, longer-term marker of energy balance in the Okinawan and
DHEA Measurement in Study Subjects
All subjects from Okinawa were healthy, community-dwelling men and
women aged approximately 75 years (n=54 septuagenarians; 29 males, mean
age 74.5 ±0.7 years and 25 females, mean age 74.7 ±0.6 years) in 1988, the
time of DHEA measurement,30 and were born into the cohort from which en-
ergy balance data were derived. Septuagenarians were selected from subjects
who were attending their annual physical exam and were recruited as part of the
annual Okinawa Centenarian Study, a population-based study of Okinawans
over the age of 100 and selected other elderly controls that begun in 1976.31
The reference population for DHEA levels in Americans consisted of
healthy, community-dwelling American septuagenarians from the Rancho
Bernardo Study (n=991 septuagenarians; 534 men, mean age 68.6 ±9.0
years; 457 women, mean age 72.1 ±8.0years) who had plasma DHEA mea-
sured in 1984–1987.32
DHEA Measurement Protocol
Similar protocols were used in both the U.S. reference population from the
Rancho Bernardo Study and the septuagenarian subjects from the Okinawa
Centenarian Study. Specifically, nonfasting venous blood samples were drawn
between 8 AM and 4 PM, separated and stored at −20 to −80 ◦C for up to 8
months. Plasma was assayed for DHEA using a solid-phase 125I RIA (ra-
dioimmunoassay) (Okinawa samples: SRL Laboratories, Tokyo, Japan; Ran-
cho Bernardo samples: Rancho Bernardo Study Laboratory, San Diego, CA).
The inter- and intra-assay coefficients of variations averaged 5.2% and 10.0%
in the Okinawan samples and 6.7% and 6.1% in the Rancho Bernardo samples,
WILLCOX et al. 439
Statistical Analysis of DHEA Levels
Septuagenarian DHEA levels were compared between the study population
in Okinawa and the U.S. reference population in Rancho Bernardo. Two sample
unpaired Student t-tests with unequal variances were used to compare the
Okinawans to the U.S. reference population. A two-tailed P-value of 0.05 was
considered a statistically significant difference.
Mortality and Survival Comparisons
The late-life survival experience (mortality rates) for Okinawan, Japanese,
and American populations was modeled according to the SAS procedure
LIFEREG, which fits parametric models to failure-time data.33 Period life-
table data from the U.S. National Center for Health Statistics34 and the Japan
Ministry of Health and Welfare35 for the calendar year 1995 were used for
constructing the survival models. Cause-specific mortality for particular age-
related diseases was also calculated for the calendar year 1995 from data
obtained from the World Health Statistics Annual36 and Okinawa Prefectural
Government, Division of Statistics37 and age-standardized to the World Stan-
To assess whether there was evidence for CR in elderly Okinawans and if
so, for what period of time, we analyzed long-term trends in whole-population
caloric intake and energy balance for the years 1949–1998. These data demon-
strate that the Okinawan population appeared to be in a relative “energy deficit”
consistent with CR until the 1960s, eating approximately 10.9% fewer calo-
ries than would normally be recommended for maintenance of body weight,
according to the Harris–Benedict equation (FIG. 1A).29 Consistent with adap-
tation to a long-term energy deficit, the BMI of adult Okinawans remained
stable at a very lean level of approximately 21 kg/m2until the 1960s. During
the 1960s the Okinawan adult BMI began to rise (FIG. 1A). This was coincident
with a shift to consistently positive energy balance.
To further clarify whether these population data indeed support a CR state for
Okinawans before 1960, analysis of body weight by age strata was performed
for the year 1949, when the Okinawan population appeared to be under CR
conditions. Studies in nonhuman primates show that there is consistently lower
body weight at all ages and relatively small weight gain beyond adulthood
in male CR monkeys in contrast to those with ad libitum access to food.8
Comparisons between the effects of CR on body weight in nonhuman primates
(FIG. 1B) and Okinawan men in 1949 (FIG. 1C) show several similarities. There
440 ANNALS OF THE NEW YORK ACADEMY OF SCIENCES
FIGURE 1. Population energy intake, energy expenditure, and anthropometric data for
adult Okinawans for the years 1949–1998. (A) Data standardized for 50-year-old Okinawans
from 1949–1998 show that Okinawans were in negative energy balance (CR) of approxi-
mately 11% until the 1960s. As Okinawans transitioned to positive energy balance, BMI
began to increase. This was due to a combination of increased energy intake and decreased
energy expenditure. BMI peaked at the height of positive energy balance in the mid 1990s.
(B) Body weight in CR nonhuman primates is markedly lower across all age strata versus
non-CR controls.8(C) Anthropometric data from the year 1949 in Okinawa while under
CR conditions demonstrate that body weight is markedly lower across all age strata versus
1972, when energy balance had shifted to non-CR conditions.
WILLCOX et al. 441
was a marked difference in body weight at all ages and relatively small change
in body weight across age strata in both groups after peak body weight is
achieved at adult ages. In Okinawans during 1949, peak body weight was
reached at approximately age 30 years and appeared relatively stable until
elderly ages, when it began to decline.
To quantify the degree of CR in the Okinawan population, energy balance
calculations are presented in FIGURE 2A. Okinawan whole-population data for
the year 1949, when the septuagenarian cohort was aged approximately 30,
show an energy intake of 1785kcal/day and energy requirement estimated at
2003 kcal/day. This energy deficit is quantified in FIGURE 2B as a negative
energy balance of −218 kcal/day (10.9%) and is consistent with the very lean
adult population BMI of 21.2, thus supporting a CR phenotype. Population data
collected in 1972 demonstrate a positive energy balance of 212 calories per
day and a corresponding BMI of 23.3, suggesting that the CR phenomenon
was largely over for the Okinawan population sometime in the 1960s. For
comparative purposes, caloric intake and energy expenditure calculations for
the U.S. population using data collected during the years 1971–1974 in the
NHANES I study,28 demonstrate a positive energy balance of 239 calories per
day, consistent with a higher population BMI of 25.6 (FIG.2B).
Since the CR paradigm rests upon “undernutrition”without malnutrition
we analyzed whether the traditional Okinawan diet was of sufficient nutri-
tional quality to ensure that widespread nutritional deficiency did not occur
and to generate hypotheses as to the impact that other food components might
have had on health status (e.g., protein intake, antioxidant vitamins). TABLE 1
presents whole-population dietary intake for the year 1949.24 Notable is the
high intake of vegetables, particularly sweet potatoes and soy, as a principal
protein. TABLE 2 presents micronutrient data that suggest that the diet was ade-
quate in most micronutrients. Notable is that intake of the antioxidant vitamins
C and E, as well as folate and vitamin B6were very high at 289%, 190%, 295%
and 221% of recommended intake, respectively, whereas vitamins D, B2, and
B12 were quite low at 2%, 45%, and 27% of recommended intake, respectively.
TABLE 3 presents physical examination data from the same subjects from
whom dietary information was collected in 1949. Notable is the relatively high
prevalence of cheilosis (dry, cracked lips and mouth) at 10.7% of the popula-
tion. This is consistent with the low consumption of vitamin B2(riboflavin)
reported in TABLE 2. Notable as well is the relatively high prevalence of delayed
menstruation and deficient lactation, consistent with low caloric intake and/or
low body fat levels in women.38,39
If indeed a CR phenomenon occurred in elderly Okinawans at younger
ages, it is theoretically possible that there might be biomarker evidence of CR-
linked delayed physiological aging, as suggested by animal data.8Therefore, we
measured DHEA in 54 Okinawan septuagenarians, who would have undergone
CR until at least middle age, according to the previous population data. As
442 ANNALS OF THE NEW YORK ACADEMY OF SCIENCES
FIGURE 2. Daily energy balance and BMI in Okinawans and Americans (kcal/day).
(A) Caloric expenditure for various levels of activity is based on reported occupation and
activity levels from NHANES I (U.S. National Center for Health Statistics 1978), the Office
of the Civil Administrator of the Ryukyu Islands (1949), the U.S. Department of the Office
of the Civil Administrator of the Ryukyu Islands (1949), and the National Nutrition Survey,
Japan Ministry of Health, Labor and Welfare (1972). BMR=basal metabolic rate (based
on sex, height, body weight at age 50, Harris–Benedict equation).29 Energy balance shifted
from negative to positive in Okinawa from 1949 to 1972, supporting an early-life CR
phenotype for older Okinawans. By the 1970s, population data from the U.S. and Okinawa
standardized to 50 year olds show that both Americans and Okinawans were in positive
energy balance. (B) Population data from the U.S. and Okinawa show that Okinawans were
in negative energy balance of approximately −218 kcal/day in 1949. Both Okinawans and
Americans were in positive energy balance in the 1970s. Americans had a positive energy
balance of approximately 239 kcal/day in the 1970s, while Okinawans were in positive
energy balance of approximately 212 kcal/day. This supports an energy balance shift of
approximately 400 kcal/day for Okinawans during these years.
WILLCOX et al. 443
TABLE 1. Traditional dietary intake of Okinawans and other Japanese circa 1950
Okinawa, 1949aJapan, 1950b
Total calories 1785c2068
Total weight (grams) 1262 1057
Caloric density (calories/gram) 1.4 2.0
Total protein in grams (% total calories) 39 (9) 68 (13)
Total carbohydrate in grams (% total calories) 382 (85) 409 (79)
Total fat in grams (% total calories) 12 (6) 18 (8)
Saturated fatty acid 3.7 4.7
Monounsaturated fatty acid 3.6 5.3
Polyunsaturated fatty acid 4.8 8.0
Total fiber (grams) 23 23
Food group Weight in grams (% total calories)
Rice 154 (12) 328 (54)
Wheat, barley, and other grains 38 (7) 153 (24)
Nuts, seeds <1(<1) <1(<1)
Sugars 3 (<1) 8 (1)
Oils 3 (2) 3 (1)
Legumes (e.g., soy and other beans) 71 (6) 55 (3)
Fish 15 (1) 62 (4)
Meat (including poultry) 3 (<1) 11 (<1)
Eggs 1 (<1) 7 (<1)
Dairy <1(<1) 8 (<1)
Sweet potatoes 849 (69) 66 (3)
Other potatoes 2 (<1) 47 (2)
Other vegetables 114 (3) 188 (1)
Fruitd<1(<1) 44 (1)
Seaweed 1 (<1) 3 (<1)
Pickled vegetables 0 (0) 42 (<1)
Foods: flavors & alcohol 7 (<1) 31 (2)
aData derived from analysis of U.S. National Archives, archived food records, 1949 and based on
survey of 2279 persons.
bJapan National Nutrition Survey, 1950.
cTotal daily caloric intake was originally reported as 1785 kcal/day in 1949. This was estimated
to be 17% less than government-recommended daily intake. Differences in assumptions regarding
particular foods, cooking methods, and choice of nutritional analysis programs result in a range of
1605 to 2012 kcal/day.
dPapaya and tomatoes were classified as vegetables.
predicted by the CR hypothesis, the mean value of DHEA for male septuagenar-
ians in Okinawa was significantly higher at 2.59 ng/mL (95% CI: 2.24–2.94)
versus 2.00 ng/mL (95% CI: 1.91–2.10) in non-CR American men (P<0.001)
(FIG. 3). Larger differences were seen in similarly aged women from the two
populations with levels of 3.03 ng/mL (95% CI: 2.48–3.58) found in Oki-
nawans versus 1.13 ng/mL (95% CI: 1.06–1.20) in Americans (P<0.001)
444 ANNALS OF THE NEW YORK ACADEMY OF SCIENCES
TABLE 2. Micronutrient sufficiency of the traditional Okinawan diet and Japanese diet
Okinawa, 1949bJapan, 1950c
Micronutrient Amount % RDAJaAmount % RDAJa
Vitamin A (RE) 602 110 337 62
Vitamin D (mcg)d0.4 2 7.9 31
Vitamin E (mg) 16.6 190 6.3 72
Vitamin K (mcg) 87.6 160 65.8 120
Vitamin B1: thiamin (mg) 1.4 137 1.1 113
Vitamin B2: riboflavin (mg) 0.5 45 0.5 47
Niacin (mg) 13.2 93 18.1 127
Vitamin B6: pyridoxine (mg) 3.0 221 1.6 118
Folate (mcg) 557.4 295 267.2 141
Vitamin B12: cobalamin (mcg)e0.6 27 4.0 176
Vitamin C (mg) 273.4 289 94.9 100
Calcium (mg)f505.3 82 325.5 53
Iron (mg) 11.6 109 11.0 103
Phosphorus (mg) 864.1 115 1191.2 159
Magnesium (mg) 396.1 151 327.4 125
Potassium (mg) 5199.6 272 2712.3 142
Zinc (mg) 6.2 62 10.6 107
Sodium (mg) 1133.0 113 2450.8 245
aMicronutrient requirements are from 6th Recommended Dietary Allowances for the Japanese
bCalculated from U.S. National Archive, archived food records, 1949.
cCalculated from the Japan National Nutrition Survey and the Statistics Record of the Ministry of
Agriculture, Forestry and Fisheries, Government of Japan, 1950.
dOkinawa is located at a subtropical (26.4◦) latitude and subjects would likely have manufactured
enough in vivo vitamin D from sunlight to meet RDAJ.70
ePeriodic festivals (approximately monthly) in which pork and other meats were consumed are not
accounted for in this analysis.
fOkinawan drinking water is high in calcium and other minerals, which are not accounted for in
In addition, if delayed aging occurred in Okinawans, there may be a right-
ward shift in the survival curve as seen in animal studies,3with increases in
both average life span (defined here as age at death for 50th percentile of popu-
lation) and maximum life span (defined here as age at death of 99th percentile
of survival). FIGURE 4 displays survival curves for Okinawan, Japanese, and
U.S. populations for the year 1995. These data show increases in both average
and maximum life span in the Okinawan population compared to Japanese
and American populations, consistent with CR. Average life span and max-
imum life span in the Okinawan, Japanese, and U.S. populations was 83.8
and 104.9 years, 82.3 and 101.1 years, and 78.9 and 101.3 years, respectively.
These data are based on a conservative survival model (LIFEREG Procedure)
and differences are even greater in maximum life span between Okinawans
and Japanese and/or Americans using a Weibull survival model (data not
WILLCOX et al. 445
TABLE 3. Nutritional deficiency symptoms in Okinawa, 1949
Urban Rural All Okinawa
Number of persons examined 797 1029 1826
Persons with no symptoms (%) 74.3 81.6 80.5
Persons with one or more symptoms (%) 25.7 18.4 19.5
Anemia 3.1 1.1 1.4
Hyperkeratosis 0.4 0.7 0.7
Xerophthalmia –0.7 0.6
Cheilosis 14.1 10.1 10.7
Glossitis 2.0 0.7 0.9
Loss of knee jerk 5.5 2.7 3.1
Edema 0.5 1.6 1.5
Chronic diarrhea 0.9 0.9 0.9
Bradycardia 0.8 1.8 1.6
Delayed menarchea4.7 10.2 9.4
Deficient lactationa5.6 20.0 17.8
aIn women only.
Finally, since CR induces profound reductions in risk for chronic disease and
increases the age at onset of chronic diseases,2we summarize age-adjusted
mortality data for specific age-related diseases in FIGURE 5. Coronary heart
disease, and forms of cancer, such as lymphoma, and cancer of the prostate,
breast, and colon are remarkably low in age-matched Okinawans versus other
Japanese and Americans. Since the forces of mortality act most strongly at
older ages, these mortality differences reflect mainly the mortality experience
of the older Okinawan cohort, which appears to have been subjected to mild
The older cohort of Okinawans (aged 65-plus) is remarkable in many ways.
Of particular interest is that they possess among the highest functional capacity
and the longest survival in Japan, the country with world’s longest-lived popu-
lation. Life expectancy at birth for the year 2000 was 86.0years for Okinawan
women and 77.6 years for Okinawan men, respectively. Life expectancy for
the septuagenarian cohort (life expectancy at age 65) is the highest in Japan,
and possibly the world, at 24.1 years for females and 18.5 years for males,
respectively.40 This compares to 22.5 years and 17.6 years for the same birth
cohort in mainland Japan and 19.3 years and 16.2 years for the corresponding
U.S. birth cohort of females and males, respectively.41
The question addressed by this article is whether or not CR is linked to this
phenomenon. Since this is an epidemiologic study, reporting mainly population
data, the results must be viewed as hypothesis-generating and not conclusive.
The question of whether CR or some other unknown factor, such as genetics,
446 ANNALS OF THE NEW YORK ACADEMY OF SCIENCES
FIGURE 3. Plasma DHEA levels (ng/mL) in elderly Okinawans and Americans.
DHEA was significantly higher in older Okinawan men (P<0.001) and Okinawan women
(P<0.001) versus older American men and women, respectively. yo =years old.
is responsible for the healthy longevity of the Okinawans are hypotheses that
require further testing. Nevertheless, this study is supportive of at least a partial
role for CR in Okinawan longevity and is significant for several reasons.
One, this is the first in-depth epidemiologic study of long-term CR in the
Okinawan population, which has been cited as a potential human example of
the effects of CR on aging and age-related disease.18 Two, this study is the
only study to report population-based dietary and phenotypic data over the
adult life span for the older cohort of Okinawans. Previous studies reported
on small selected samples of Okinawans, which are more subject to bias, and
reported data at a single point in time.18,20 Three, this study supports and
extends earlier observations from the 1970s regarding the potential impact
of CR on Okinawan health and longevity.18 Okinawans indeed appear to be
one of the few populations in the world that may have experienced mild long-
term CR without significant malnutrition, and this may be linked to their
exceptionally healthy survival. That this phenomenon appears to have occurred
in their natural environment also makes them a population of special interest.
Four, there is a dearth of research on factors that may affect the rate of
human aging and controversy on how exactly to measure this. While there are
no universally accepted biomarkers of aging,14 this study adds human data that
are consistent with the literature in animals with regard to the effects of CR on
DHEA levels.8,27 Furthermore, this is the first study that has shown extended
WILLCOX et al. 447
FIGURE 4. Survival model for Okinawans, Japanese, and American populations. Data
show extended life span (average, maximum) for Okinawans (83.8 years, 104.9years) com-
pared to other Japanese (82.3 years, 101.1 years), and Americans (78.9years, 101.3 years).
average and maximum life span in a human population that is potentially due
to CR, as indicated by a rightward shift of the survival curve for Okinawans.
While mortality due to age-related disease was much lower in the Okinawans
and helped lead to a higher average life span, the apparent increase in maximum
life span is suggestive of slower aging in the Okinawans. Finally, this study
population has among the longest life expectancy in the world, and includes
what may be the world’s highest prevalence of exceptionally aged individuals,
such as centenarians, and therefore is an important population for the study
of environmental and genetic factors that may predict or lead to exceptional
There are several strengths to this study. One, this is among the most com-
prehensive studies of dietary and phenotypic data related to CR in Okinawa
to date and the first to report population-based energy balance and pheno-
typic data in older Okinawan adults before the 1970s, when they were young
adults. This is important for establishing a link between CR and longevity
in Okinawa since it appears that only older Okinawans have a survival ad-
vantage compared to other Japanese.22 Two, a sample of fairly well-defined
subjects from the older cohort of Okinawans was available for biomarker
448 ANNALS OF THE NEW YORK ACADEMY OF SCIENCES
FIGURE 5. Mortality from age-associated diseases in Okinawans versus Americans.
Numbers represent age-adjusted mortality rate in deaths per hundred thousand persons
per year for 1995. Coding was according to ICD-9 codes; populations were age-adjusted to
World Standard Population. These data show markedly lower mortality risk from age-related
diseases in Okinawans versus other Japanese and Americans.
(DHEA) study, and both the Okinawan and American samples were se-
lected from specific populations of healthy septuagenarians in geographically
defined communities living in their natural environments. Three, while exact
dietary data are not known for these individuals, population data are other-
wise known for their birth cohorts, collected with good dietary methodology.
While these are only crude measurements of energy balance, the population
data support an energy deficit in Okinawans (but not Americans) that would
require physiological adaptations consistent with a CR phenotype.
Four, anthropometric data, such as BMI, are consistent with a large difference
in energy balance between the study populations and support a CR status for
Okinawan septuagenarians at younger ages. The study cohort to which these
Okinawans belong was relatively thin and appeared to have been calorically
restricted at least until middle age. Support for a CR phenotype for members
of this older cohort can also be derived from later-life measurements of their
BMI, which was 23.5 for males and 24.1 for females when aged approximately
75 years.20 When corrected for potential loss of height with aging43 the true
BMI for these subjects may be closer to 22.4 for males and 22.0 for females.
This low BMI, compared with an average BMI of 21.2 in 1949 at approximately
age 30, suggests minimal weight gain with age that is also consistent with CR.
WILLCOX et al. 449
There were several limitations to this study, mostly centered on the compar-
isons with the U.S. study population. However, this study is mainly concerned
with the cumulative evidence that CR occurred in the older Okinawan popula-
tion and whether there is evidence that CR-linked outcomes may have occurred,
including phenotypic outcomes, lower mortality from age-associated diseases,
and extended survival.
Limitations include the fact that population data are used to estimate whether
this older generation of Okinawans were calorically restricted and these data
are subject to measurement error. Clearly it is not possible to maintain an en-
ergy deficit in the long term and still be healthy. Nevertheless, the low BMI
from the Okinawan population until the 1960s is consistent with a long-term
adaptive response to limited energy availability and periodic energy deficits.
These energy deficits are supported by historical reports of periodic crop fail-
ures that occurred in Okinawa in the early 20th century and a long history of
marginal food supply.44 A CR status appears to have remained until at least
middle age for Okinawan septuagenarians since dietary and physical activity
patterns remained relatively consistent in Okinawa until the 1960s (FIG. 1A).
After the 1960s standards of living improved, physical activity declined, and
Westernization (increased fat, meat and bread) and Japanization (increased
polished Japanese or “Japonica”white rice) of the diet occurred, with large
shifts in energy balance.
Another weakness of this study centers on interpretation of the DHEA lev-
els. Since the populations were of different ethnicity and the Okinawans come
from an island population with restricted gene flow, there may be genetic or
other unique features that account, at least in part, for the differences in DHEA
levels and health status of the populations.45 However, we could find no pub-
lished data on DHEA in other Japanese populations, and the only prior reported
comparisons of Japanese and Americans involved dehyroepiandrosterone sul-
fate (DHEAS). These data demonstrate that Japanese men actually have lower
DHEAS levels than American men when younger.46 Therefore, persistently
higher life-long DHEA levels in Okinawans compared with Americans is un-
likely to account for the older-age differences in DHEA we found in this study.
More likely is a slower age-related decline in DHEA in the Okinawans, con-
sistent with CR.
In support of this, Nafziger et al.47 reported DHEA levels from a study
of community-dwelling white Americans that were not markedly different
from the values for the comparison group in this study (Rancho Bernardo
cohort). Plasma levels were reported as 1.6 ng/mL for septuagenarians of both
sexes—lower than Okinawans of similar age. In addition, in a separate study
of plasma estrogen levels in these same subjects, similar differences were
observed between Okinawan and U.S. populations, with significantly higher
levels of estrogen seen in Okinawans at older ages.48 Estrogen, as a downstream
byproduct of DHEA, usually reflects DHEA’s plasma level (i.e., high DHEA
means relatively high estrogen).
450 ANNALS OF THE NEW YORK ACADEMY OF SCIENCES
Finally, other factors might be responsible for the mortality advantages and
phenotypic differences between Okinawans, other Japanese, and Americans.
These include differences in dietary macronutrients, such as differences in
protein intake or amino acid composition of the diet, which has been linked to
longevity in rodents49,50 and fruit flies51 and to stroke risk in animals and hu-
mans,52,53 lower glycemic index carbohydrates in the Okinawans, which may
lower cardiovascular and cancer risk,54 higher consumption of flavonoids (e.g.,
soy foods and Okinawan sweet potatoes),55 which may stimulate sirtuin pro-
teins and have been implicated as CR mimetics,56 higher intake of antioxidant-
rich vegetables, lower intake of sodium and higher K/Na ratio (both implicated
in lower blood pressure),57 or multiple other dietary differences.
Chronic low caloric intake and chronic low protein and/or low intake of
particular amino acids (e.g., methionine) may also be chronic low-intensity
stressors and may have contributed to a phenomenon known as “hormesis,”
where low levels of otherwise damaging agents have positive, potentially life-
Other factors, such as genetic differences41,59,60 and psychosocial factors,61
are important survival factors that have not been accounted for in this study. In
particular, recent work suggests that Okinawans may possess familial survival
advantages that account, at least in part, for their remarkable longevity.42 On
the other hand, fuller expression of the potential health and longevity benefits
of CR in Okinawans would have been limited by the lack of a good public health
infrastructure until the 1960s, and resultant high death rates due to tuberculosis
and other infectious diseases.62 Furthermore, of the nutritional factors, only
CR has been found in animal studies capable of consistently increasing both
average and maximum life span, indicated by a rightward shift of the survival
Thus, the septuagenarian cohort of Okinawans appears to have experienced
CR of approximately 11% with respect to their estimated energy requirements,
until middle age. Older cohorts in Okinawa may have experienced CR for even
longer or at greater levels, consistent with their increased survival advantage
compared with that of other Japanese, although estimates of the degree of
CR should be viewed with caution since they were obtained from energy bal-
ance estimates for the whole Okinawan population and were not validated
by precise metabolic methods, such as indirect calorimetry or doubly labeled
The CR status of the older Okinawans appears to be due, in part, to a high
level of occupational energy demand from their main vocation as farmers. This
was coupled to low caloric intake from an energy-poor but nutrient-dense diet
rich in Okinawan sweet potatoes, other vegetables, legumes, and other foods
low in energy density (FIG. 1).24 The Okinawan sweet potato, with a caloric
density of 1.0 kcal/gram, has been the main carbohydrate of the Okinawan diet
from the 1600s until approximately 1960, accounting for more than 50% of
calories.64 The higher caloric density of the traditional Japanese diet is partly
WILLCOX et al. 451
due to their higher consumption (75% of calories) of high caloric density
Japonica (white) rice, with a caloric density of 1.5 kcal/gram.
The CR phenotype appears to be disappearing in Okinawa except among
older birth cohorts (aged 65-plus years). Calorie intake has been increasing and
activity levels decreasing since the 1960s with concomitant population-wide
increases in BMI, although this trend appears to have stabilized after reversion
of Okinawa from U.S. to Japanese governance in the 1970s (FIG. 1A). Loss of
the CR phenotype among younger Okinawans and increased weight gain in this
cohort is also associated with higher mortality from obesity-linked diseases.
This can be inferred from a mortality crossover, which has followed a BMI
crossover in Okinawans versus other Japanese.22 Concomitantly, there has been
a significant slowing of life expectancy gains over the last two decades. This
may, as is hypothesized in Americans,65 have an unfavorable impact upon the
health and longevity of subsequent generations.
Early-life energy deficits experienced by older Okinawans would have re-
quired compensatory responses in terms of more efficient use of energy66
since long-term energy deficits are unsustainable and inconsistent with health.
These adaptations may have included less body fat, lower blood sugar and
insulin levels, hormetic responses, and multiple changes in gene expression,
among other adaptations potentially linked to longer life span.3–6,8,67–69 Ani-
mal studies suggest that CR at any stage of the life cycle (early or later life)
may potentially result in mortality advantages versus non-CR controls.4
In conclusion, we observed low calorie intake coupled with high physical
activity levels that appear to have contributed to a CR phenotype in older Ok-
inawans. This phenotype includes a life-long low BMI, relatively high plasma
levels of DHEA at older ages, reduced mortality from age-associated diseases,
and extended average and maximum survival. While these conclusions are ten-
tative in nature, an adaptive response to early and mid-life energy restriction
in the older cohort of Okinawans may be implicated in their low morbidity
and exceptionally long survival. This is consistent with the well-known animal
literature that supports a beneficial effect of CR on BMI, age-related biomark-
ers, morbidity/mortality, and life span. More studies using the CR paradigm
(energy restriction without malnutrition) are required to validate the effects of
low energy intake and/or energy balance in human populations.
This work was supported by the National Institute on Aging (Grants RO3
AG021293-01, K08 AG22788-02 and R01 AG027060-01 to B.W.) and the
Japan Ministry of Health, Labor and Welfare. All authors contributed to
the design of the experiment, collection of data, analysis of data, and writ-
ing of the manuscript. We thank Sayaka Mitsuhashi and Joshua Spiegelman
for research assistance and manuscript preparation.
452 ANNALS OF THE NEW YORK ACADEMY OF SCIENCES
1. MCCAY, C.M., M.F. CROWELL & L.A. MAYNARD. 1935. The effect of retarded
growth upon the length of life span and upon the ultimate body size. Nutrition
2. MAEDA, H., C.A. GLEISER, E.J. MASORO,et al. 1985. Nutritional influences on
aging of Fischer 344 rats: II. Pathology. J. Gerontol. 40: 671–688.
3. WEINDRUCH, R., R.L. WALFORD,S.FLIGIEL,et al. 1986. The retardation of aging
in mice by dietary restriction: longevity, cancer, immunity and lifetime energy
intake. J. Nutr. 116: 641–654.
4. DHAHBI, J.M., H.J. KIM, P.L. MOTE,et al. 2004. Temporal linkage between the
phenotypic and genomic responses to caloric restriction. Proc. Natl. Acad. Sci.
USA 101: 5524–5529.
5. MASORO, E.J. 2005. Overview of caloric restriction and ageing. Mech. Ageing
Dev. 126: 913–922.
6. MASORO, E.J. 2006. Caloric restriction and aging: controversial issues. J. Gerontol.
A. Biol. Sci. Med. Sci. 61: 14–19.
7. DIRKS, A.J. & C. LEEUWENBURGH. 2006. Caloric restriction in humans: potential
pitfalls and health concerns. Mech. Ageing Dev. 127: 1–7.
8. MATTISON, J.A., M.A. LANE, G.S. ROTH,et al. 2003. Caloric restriction in rhesus
monkeys. Exp. Gerontol. 38: 35–46.
9. LANE, M.A., J.A. MATTISON, G.S. ROTH,et al. 2004. Effects of long-term diet
restriction on aging and longevity in primates remain uncertain. J. Gerontol. A.
Biol. Sci. Med. Sci. 59: 405–407.
10. LEE, I.M., S.N. BLAIR,D.B.ALLISON,et al. 2001. Epidemiological data on the
relationships of caloric intake, energy balance, and weight gain over the life span
with longevity and morbidity. J. Gerontol. A. Biol. Sci. Med. Sci. 56: 7–19.
11. WALFORD, R.L., D. MOCK,R.VERDERY,et al. 2002. Caloric restriction in biosphere
2: alterations in physiologic, hematologic, hor monal, and biochemical parameters
in humans restricted for a 2-year period. J. Gerontol. A. Biol. Sci. Med. Sci. 57:
12. HEILBRONN,L.K.&E.RAVUSSIN. 2003. Calorie restriction and aging: review of
the literature and implications for studies in humans. Am. J. Clin. Nutr. 78: 361–
13. FONTANA, L., T.E. MEYER,S.KLEIN,et al. 2004. Long-term calorie restriction is
highly effective in reducing the risk for atherosclerosis in humans. Proc. Natl.
Acad. Sci. USA 101: 6659–6663.
14. BUTLER, R.N., R. SPROTT,H.WAR NER,et al. 2004. Biomarkers of aging: from
primitive organisms to humans. J. Gerontol. A. Biol. Sci. Med. Sci. 59: B560–
15. MEYER, T.E., S.J. KOVACS, A.A. EHSANI,et al. 2006. Long-term caloric restriction
ameliorates the decline in diastolic function in humans. J. Am. Coll. Cardiol. 17:
16. FONTANA, L. & S. KLEIN. 2007. Aging, adiposity, and calorie restriction. JAMA
17. WILLCOX, B.J., K. YANO,R.CHEN,et al. 2004. How much should we eat? The
association between energy intake and mortality in a 36-year follow-up study of
Japanese American men. J. Gerontol. A. Biol. Sci. Med. Sci. 59: 789–795.
18. KAGAWA, Y. 1978. Impact of Westernization on the nutrition of Japanese: changes
in physique, cancer, longevity and centenarians. Prev. Med. 7: 205–217.
WILLCOX et al. 453
19. HOKAMA, T., H. ARAGAKI,H.SHO,et al. 1967. Nutrition survey of school children
in Okinawa. Sci. Bull. Coll. Agr. Univ. Ryukyus 14: 1–15.
20. CHAN, Y.C., M. SUZUKI &S.YAMAMOTO. 1997. Dietary, anthropometric, hemato-
logical and biochemical assessment of the nutritional status of centenarians and
elderly people in Okinawa, Japan. J. Am. Coll. Nutr. 16: 229–235.
21. JAPAN MINISTRY OF HEALTH,LABOR AND WELFARE. 2002. Results from Ministry
of Health, Labor and Welfare Nutrition Survey, 2000. Daiichi Publishers. Tokyo,
22. TODORIKI, H., D.C. WILLCOX &B.J.WILLCOX. 2004. The effects of post-war dietary
change on longevity and health in Okinawa. Okinawa J. Amer. Studies 1: 52–
23. WILLCOX, D.C. 2005. Okinawan longevity: where do we go from here? Nutr. Di-
etetics 8: 9–17.
24. U.S. DEPARTMENT OF THE OFFICEOFTHECIVIL ADMINISTRATOR OF THE RYUKYU
ISLANDS:RECORDS OF HEALTH,EDUCATION AND WELFARE. 1949. U.S. Oc-
cupation Headquarters, World War II. Record Group 260.12.5. National
Archives at College Park, 8601 Adelphi Road, College Park, MD 20740-6001.
25. DIVISION OF HEALTH,LABOR AND WELFARE,OKINAWA PREFECTURAL GOVERNMENT.
1976. Health and Welfare Dataset 1880–1976. Okinawa Prefectural Government.
26. LANE, M.A., D.K. INGRAM, S.S. BALL,et al. 1997. Dehydroepiandrosterone sulfate:
a biomarker of primate aging slowed by caloric restriction. J. Clin. Endocrinol.
Metab. 82: 2093–2096.
27. ROTH, G.S., M.A. LANE, D.K. INGRAM,et al. 2002. Biomarkers of caloric restriction
may predict longevity in humans. Science 297: 811.
28. NATIONAL CENTER FOR HEALTH STATISTICS. 1978. National Health and Nutrition
Examination Survey I Dataset. 1971–74. Vital and Health Statistics. U.S. Gov-
ernment Printing Office. Washington, DC.
29. FRANKENFIELD, D.C., E.R. MUTH & W.A. ROWE. 1998. The Harris-Benedict studies
of human basal metabolism: history and limitations. J. Am. Diet Assoc. 98: 439–
30. SUZUKI,M.&N.HIROSE. 1999. Endocrine function of centenarians. In Japanese
Centenarians. Medical Research for the Final Stages of Human Aging. H. Tauchi,
T. Sato & T. Watanabe, Eds.: 101–110. Aichi Medical University Press. Aichi,
31. SANABE, E., I. ASHITOMI &M.SUZUKI. 1977. Social and medical survey of cente-
narians. Okinawa J. Pub. Health 9: 98–106.
32. GREENDALE, G.A., S. EDELSTEIN &E.BARRETT-CONNOR. 1997. Endogenous sex
steroids and bone mineral density in older women and men: the Rancho Bernardo
Study. J. Bone. Miner. Res. 12: 1833–1843.
33. SAS INSTITUTE INC. 1999. The LIFEREG Procedure, SAS/STATUser’s Guide,
Version 8. SAS Institute Inc. Cary, NC.
34. NATIONAL CENTER FOR HEALTH STATISTICS. 1998. Vital Statistics of the United
States, 1995, preprint of Vol. II, Mortality, part A, section 6, life tables. National
Center for Health Statistics. Hyattsville, MD.
35. JAPAN HEALTH AND WELFARE STATISTICS ASSOCIATION. 1996. 1995 Annual Statis-
tical Report of National Health Conditions. Tokyo, Japan.
36. WORLD HEALTH ORGANIZATION. 1996. 1995 WorldHealth Statistics Annual. WHO.
37. OKINAWA PREFECTURAL GOVERNMENT,DIVISION OF STATISTICS. 1996. 39th Oki-
nawa Statistics Annual, 1995. Okinawa, Japan.
454 ANNALS OF THE NEW YORK ACADEMY OF SCIENCES
38. MEYER, F., J. MOISAN,D.MARCOUX,et al. 1990. Dietary and physical determinants
of menarche. Epidemiology 1: 377–381.
39. DEWEY, K.G. 1998. Effects of maternal caloric restriction and exercise during
lactation. J. Nutr. 128(2 Suppl): 3865–3895.
40. JAPAN MINISTRY OF HEALTH,LABOR AND WELFARE. 2000. Prefectural Life Tables.
Statistics and Information Department, Health and Welfare Statistics Associa-
tion. Tokyo, Japan.
41. CENTERS FOR DISEASE CONTROL AND PREVENTION,NATIONAL CENTER FOR HEALTH
STATISTICS. 2003. National Vital Statistics System. NCHS. Hyattsville, MD.
Available at http://www.cdc.gov/nchs/.
42. WILLCOX, B.J., D.C. WILLCOX,Q.HE,et al. 2006. Siblings of Okinawan centenar-
ians exhibit lifelong mortality advantages. J. Gerontol. A. Biol. Med. Sci. 61:
43. SORKIN, J.D., D.C. MULLER &R.ANDRES. 1999. Longitudinal change in height
of men and women: implications for interpretation of the body mass index: the
Baltimore Longitudinal Study of Aging. Am. J. Epidemiol. 150: 969–977.
44. KERR, G. 2000. Okinawa: the History of an Island People. Tuttle Publishing.
45. ROTTER, J.I., F.L. WONG, E.T. LIFRAK,et al. 1985. A genetic component to the
variation of dehydroepiandrosterone sulfate. Metabolism 34: 731–736.
46. LACROIX, A.Z., K. YANO & D.M. REED. 1992. Dehydroepiandrosterone sulfate,
incidence of myocardial infarction, and extent of atherosclerosis in men. Circu-
lation 86: 1529–1535.
47. NAFZIGER, A.N., S.J. BOW LI N, P.L. JENKINS,et al. 1998. Longitudinal changes in
dehydroepiandrosterone concentrations in men and women. J. Lab. Clin. Med.
48. WILLCOX, B.J., D.C. WILLCOX,M.SUZUKI,et al. 2000. Serum estrogen and long
term survival in Okinawan-Japanese men and women. J. Am. Geriatr. Soc. 48:
49. ZIMMERMAN, J.A., V. MALLOY,R.KRAJCIK,et al. 2003. Nutritional control of aging.
Exp. Gerontol. 38: 47–52.
50. SANZ, A., P. CARO &G.BARJA. 2004. Protein restriction without strong caloric re-
striction decreases mitochondrial oxygen radical production and oxidative DNA
damage in rat liver. J. Bioenerg. Biomembr. 36: 545–552.
51. MAIR, W., M.D. PIPER &L.PARTRIDGE. 2005. Calories do not explain extension
of life span by dietary restriction in Drosophila. PloS. Biol. 3: e223.
52. ABBOTT, R.D., J.D. CURB, B.L. RODRIGUEZ,et al. 1996. Effect of dietary calcium
and milk consumption on risk of thromboembolic stroke in older middle-aged
men: the Honolulu Heart Program. Stroke 27: 813–818.
AMORI, Y., S. MURAKAMI,K.IKEDA,et al. 2004. Fish and lifestyle-related dis-
ease prevention: experimental and epidemiological evidence for anti-atherogenic
potential of taurine. Clin. Exp. Pharmacol. Physiol. 31: S20–S23.
54. JENKINS, D.J., C.W. KENDALL,A.MARCHIE,et al. 2004. Too much sugar, too much
carbohydrate, or just too much? Am. J. Clin. Nutr. 79: 711–712.
55. WILLCOX, B.J., K. FUCHIGAMI, D.C. WILLCOX,et al. 1995. Isoflavone intake in
Japanese and Japanese-Canadians. Am. J. Clin. Nutr. 61: 901.
56. WOOD, J.G., B. ROGINA,S.LAVU,et al. 2004. Sirtuin activators mimic caloric
restriction and delay ageing in metazoans. Nature 430: 686–689.
57. BRAY, G.A., W.M. VOLLMER, F.M. SACKS,et al. 2004. A further subgroup analysis
of the effects of the DASH diet and three dietary sodium levels on blood pressure:
results of the DASH-Sodium Trial. Am. J. Cardiol. 94: 222–227.
WILLCOX et al. 455
58. MASORO, E.J. 1998. Hormesis and the antiaging action of dietary restriction. Exp.
Gerontol. 33: 61–66.
59. TAKATA, H., M. SUZUKI,T.ISHII,et al. 1987. Influence of major histocompat-
ibility complex region genes on human longevity among Okinawan-Japanese
centenarians and nonagenarians. Lancet 2: 824–826.
60. AKISAKA, M., M. SUZUKI &H.INOKO. 1997. Molecular genetic studies on DNA
polymorphism of the HLA class II genes associated with human longevity. Tissue
Antigens 50: 489–493.
61. GOTO, A., S. YASUMURA,Y.NISHISE,et al. 2003. Association of health behavior
and social role with total mortality among Japanese elders in Okinawa, Japan.
Aging Clin. Exp. Res. 15: 443–450.
62. WILLCOX, B.J., D.C. WILLCOX &M.SUZUKI. 2006. Exceptional human longevity.
In Aging and Age-Related Diseases. The Basics. M. Karasek, Ed.: 459–509.
Nova Science Publishers. New York, NY.
63. MASORO, E.J. 1990. Assessment of nutritional components in prolongation of life
and health by diet. Proc. Soc. Exp. Biol. Med. 193: 31–34.
64. SHO, H. 2001. History and characteristics of Okinawan longevity food. Asia Pacif ic
J. Clin. Nutr. 10: 159–164.
65. OLSHANSKY, S.J., D.J. PASSARO, R.C. HERSHOW,et al. 2005. A potential decline in
life expectancy in the United States in the 21st century. N. Engl. J. Med. 352:
66. SHRAUWEN, P., K. WALDER &E.RAVUSSIN. 1999. Human uncoupling proteins and
obesity. Obes. Res. 7: 97–105.
67. LEE, C.K., R.G. KLOPP,R.WEINDRUCH,et al. 1999. Gene expression profile of
aging and its retardation by caloric restriction. Science 285: 1390–1393.
68. BLUHER, M., B.B. KAHN & C.R. KAHN. 2003. Extended longevity in mice lacking
the insulin receptor in adipose tissue. Science 299: 572–574.
69. SPEAKMAN, J.R., D.A. TALBOT,C.SELMAN,et al. 2004. Uncoupled and surviving:
individual mice with high metabolism have greater mitochondrial uncoupling
and live longer. Aging Cell 3: 87–95.
70. HOLICK et al. 1996. Vitamin D and bone health. J. Nutr. 126: 1159S–1164S.