survival in older people of different ethnicities
Irene Darmadi-Blackberry MB, PhD1, Mark L Wahlqvist AO, MD2, Antigone Kouris-Blazos PhD2, Bertil Steen MD, PhD3, Widjaja Lukito MD, PhD4, Yoshimitsu Horie PhD5 and Kazuyo Horie BSc6
1Public Health Division, National Ageing Research Institute, Melbourne, Australia
2Asia Pacific Health & Nutrition Centre, Monash Asia Institute, Monash University, Australia
3Department of Geriatric Medicine, Goteborg University, Goteborg, Sweden
4SEAMEO TROPMED, University of Indonesia, Jakarta, Indonesia
5School of Humanities and Social Sciences, Nagoya City University, Nagoya, Japan
6Faculty of Home Economics, Aichi Gakusen, Okazaki, Japan
To identify protective dietary predictors amongst long-lived elderly people (N=785), the “Food Habits in Later Life” (FHILL) study was undertaken among five cohorts in Japan, Sweden, Greece and Australia. Between 1988 and 1991, baseline data on food intakes were collected.
There were 785 participants aged 70 and over that were followed up to seven years. Based on an alternative Cox Proportional Hazard model adjusted to age at enrolment (in 5-year intervals), gender and smoking, the legume food group showed 7-8% reduction in mortality hazard ratio for every 20g increase in daily intake with or without controlling for ethnicity (RR 0.92;
95% CI 0.85-0.99 and RR 0.93; 95% CI 0.87-0.99, respectively). Other food groups were not found to be consistently significant in predicting survival amongst the FHILL cohorts.
Key words: legumes, diet, survival, FHILL, mortality, food intake, elderly, Australia, Greece, Japan, Sweden, fish, fat ratio
Introduction
It is becoming apparent that people of various nations with
different food cultures can have comparable life
expectancy and morbidity rates.1
The ‘Food Habits in
Later Life’ (FHILL)2
is a cross-cultural study conducted
under the auspices of the International Union of Nutritional
Sciences (IUNS) and the World Health Organization
(WHO). The FHILL studies have concentrated on food
intake and food intake patterns as differentiators and
common denominators in health susceptibility and for sur-
vival within and between cultures (people of different
ethnicity living in different localities). It has been
suggested that the dietary patterns in the early 1960s best
characterise what is known today as the "Traditional Greek
Food Pattern or Traditional Medi-terranean Diet".3,4 The
Greek variant of the Traditional Mediterranean Diet has
been documented in the "Seven Countries Study" by Keys et al.
5-7 This diet was characterised by:
1. Plentiful fruits, vegetables, legumes, grains;
2. Olive oil as the principal fat;
3. Lean red meat consumed only a few times per month
or in very small portions;
4. Low to moderate daily consumption of dairy
products;
5. Poultry, fish and eggs consumed a couple of times
per week; and
6. Moderate consumption of wine.
We have previously demonstrated that a plant based diet
(similar to the traditional mediterranean diet) which
includes fish is the most important predictor of survival
amongst FHILL cohorts (in press). This result was
significant after accounting for other non-nutritional
variables such as smoking status, gender, exercise, health
status, social activity, activities of daily living and
memory status in one Cox Proportional Hazard model.
Details of this analysis are discussed elsewhere (in press).
The current article will investigate in greater detail what
aspects of the plant-fish diet contributes to longevity
between and within disparate food cultures. The
objectives of this article include:
• Describing cross-cultural comparisons on average
food consumption amongst FHILL cohorts (grams/
day adjusted to 2500 kcal for men and 2000 kcal for
women)
• Examining individual food groups as predictors of
mortality amongst FHILL cohorts (adjusted and non-
adjusted for ethnic backgrounds)
Correspondence address: Dr. Irene Darmadi-Blackberry,
National Ageing Research Institute, PO Box 31, Parkville 3052.
Tel: 03-83872614; Fax 03-83872153
Email: [email protected]
Accepted 15 January 2004
Methods
Baseline data on food habits, health status and social
variables were collected from five cohorts aged 70 and
over (Japanese in Japan, Swedes in Sweden, Anglo-Celtic
in Australia, Greeks in Australia and Greece). The
validated Food Frequency Questionnaire (FFQ) were used
to collect data on food intakes in all cohorts except
Japanese where the 3-day weighed food record method
was employed. Intakes in gram/week were calculated by
multiplying the serving size by the weekly frequency of
intake. These values were further translated into gram/day
and were adjusted to 2500 kcal (10,460 kJ) for men and
2000 kcal (8,368 kJ) for women for comparison purposes.
Furthermore, these food items were grouped into eight
major food groups based on the Mediterranean diet.8-12
To acknowledge the emerging evidence of beneficial
effects of fish consumption in health and survival,13-15 the
fish and shellfish group was added. Thus the nine major
food groups included vegetables, legumes, fruits and nuts,
milk and dairy products, cereals and potatoes, meat and
meat products, ethanol, monounsaturated to saturated fats
ratio, and fish groups.
All-cause mortality data were obtained from up to
seven years follow-up. Out of the 785 participants in the
baseline, 169 people died during the follow-up period. An
alternative Cox Proportional Hazard model, adjusted to
age at enrolment (in 5-year intervals), gender, and
smoking, was developed to analyse the survival data.
Each Cox model was tested against controlling for
cohorts’ location and ethnicity.
Results
Table 1 describes the average amount of foods consumed
per person daily amongst FHILL cohorts. Most cohorts
consumed vegetables between 283 and 353 g/d. How-
ever, the Swedes in Sweden ate far less vegetables. The
Japanese in Japan, along with both Greek cohorts in
Australia and Greece, consumed more legumes (63-
86g/d) compared to other cohorts. Fruit and nut intakes
were low for Japanese in Japan (140g/d) compared to the
rest of the cohorts (252-330g/d). On the other hand the
intake of cereals (which was mainly rice), as expected,
was high for Japanese in Japan (366g/d). Other cohorts
consumed between 102-280 grams of cereals daily.
Milk and dairy consumption varied between cohorts. It
was noted that Swedes consumed more than 400g/d
whilst Japanese consumed as little as 165g/d. The
variation in meat consumption was enormous. In Japan,
the Japanese ate very little meat (less than 50g/d). In
contrast, the Greeks in Australia ate about 190g of meat
and meat products daily. Little fish and shellfish were
eaten by the Anglo-Celts in Australia (21g/d). However,
the Japanese in Japan consumed five times more than the
Anglo-Celts in Australia (more than 100g daily).
Ethanol consumption appeared to be varied between
cohorts. While most cohorts consumed between 5 and 10
grams of ethanol daily, the Japanese (mainly men) drank
about 15 grams of ethanol per day. The Swedes and
Anglo-Celtic seemed to consume more saturated fats than
the rest of the cohorts. This result was reflected in low
monounsaturated:saturated fat ratio (less than 1). Finally,
there was considerable variation between the cohorts in
energy intakes. The Japanese in Japan reported a very
low energy intake (about 1600 kcal daily). On the other
hand, the Swedes in Sweden consumed around 2500 kcal
per day. Other cohorts reported energy intakes between
2060-2118 kcal daily.
Results from the overall FHILL cohorts suggested that
higher intakes of legumes, fish and shellfish, and olive oil
(reflected in monounsaturated:saturated fat ratio) were
significant predictors of survival in the elderly (Table 2).
There was a reduction in risk of death by up to 8% for
every 20 grams increase in legumes intake. The result
remained significant with or without controlling for ethnic
backgrounds. With every 20 grams increase in fish and
shellfish intake, there was a 6% decrease in the hazard of
death (95% CI 1%-12%), after adjusting for age, gender,
and smoking status only. For every unit increase in the
mono-unsaturated:saturated fat ratio there was a 46%
decrease in the hazard of death (95% CI 9%-68%).
However, this ratio was a significant predictor of
mortality only when ethnic background was included as a
confounding factor in addition to age at enrolment (in 5-
year intervals), gender, and smoking status.
Only for legume intake was the result plausible,
consistent and statistically significant across collective
FHILL cohorts’ data. There is a 7% - 8% reduction in
Table 1. Average amount of foods consumed (grams) per person daily in FHILL cohorts
mortality hazard ratio for every 20g increase in daily
legume intake with adjustment for location/ethnicity (RR
0.92; 95% CI 0.85 – 0.99) and without adjustment for
location/ethnicity (RR 0.93; 95% CI 0.87 – 0.99).
Discussion
Food intake promises to be, not surprisingly, one of the
best measures of nutritional status. Together with body
composition and various performance measures such as
strength and endurance, it represents the inputs, outputs
and the sum total of energy and food component through-
put and status or balance in human biology. Biomarkers
of food intake offer ways in which its validity can be
increased and its perturbations recognised.
There were variations between the five FHILL cohorts
on average daily consumption of foods consumed. Taller
elderly consumed the most calories whilst the shortest
elderly consumed less. These findings about food and
survival apply across cultures with a wide range of energy
intakes (Japanese 1599kcal - Swedes 2501kcal) and wide
range of stature (Japanese 152cm - Swedes 165cm).
Results from the current study were compared with the
results documented by Keys and his colleagues from the
classical international "Seven Countries Study". It was
observed in the current study that the average cereal and
ethanol consumption were lower whilst the average meat
consumption was higher across longevity cultures, in
comparison to the intake from the "Seven Countries
study" in the 1960s..5 Overall, the Japanese in the FHILL
study consumed more vegetables, fruits, meat and dairy
products than the Japanese cohort from the "Seven
Countries Study". On the other hand, they consumed less
legumes, cereals, fish, and alcohol. On average, both the
Greek cohorts in Australia and Greece consumed more
vegetables, legumes, fish, meat and dairy products
compared to two Greek cohorts from the "Seven
Countries Study". However, they consumed less fruits,
cereals, and alcohol.
Only for legumes intake was the result plausible,
consistent and statistically significant from collective
FHILL cohorts data (8% reduction in risk of death for
every 20 grams increase in daily legumes intake).
Legumes have long been associated with longevity food
cultures. For example, the Japanese eat soy, tofu, natto,
miso, the Swedes eat brown beans and peas and the Medi-
terraneans eat lentils, chickpeas and white beans.6,16-18 A
6% decrease in the hazard of death (95% CI 1%-12%) for
every 20g increase in intake of fish and shellfish was
observed when the Cox's model did not include ethnicity/
locality as a confounding factor. It appeared that fish and
shellfish intakes were shown to prolong survival, but it
may be related to certain food cultures that have high
intakes of fish such as the Japanese.19 The monoun-
saturated:saturated fat ratio was associated with a 46%
decrease in the hazard of death (95% CI 9%-68%) for
every unit increase. This ratio was a significant predictor
of mortality only when ethnic background was included
as a confounding factor in the Cox's model. Thus across
longevity cultures in the FHILL study, higher monoun-
saturated: saturated fat ratio (as reflected in intake of olive
oil in the Mediterranean cultures) appeared to be
protective against premature death irrespective of their
ethnic backgrounds. The intake of monounsaturated fat
has been shown to be protective against breast cancer in
Sweden..20
Conclusions
The FHILL longitudinal study shows that a higher legume
intake is the most protective dietary predictor of survival
amongst the elderly, regardless of their ethnicity. The
significance of legumes persisted even after controlling
for age at enrolment (in 5-year intervals), gender, and
smoking. Legumes have been associated with long-lived
food cultures such as the Japanese (soy, tofu, natto, miso),
the Swedes (brown beans, peas), and the Mediterranean
people (lentils, chickpeas, white beans).
Acknowledgements
We would like to acknowledge other Food Habits in Later Life
(FHILL) investigators and collaborators for their contribution to
baseline data collection. We also deeply thank all participants in
the FHILL study for their enormous contribution to the study
