Mediterranean diet and age with respect to overall survival in institutionalized, nonsmoking elderly people

¹ From the Departamento de Biología Funcional, Area de Fisiologia, Facultad de Medicina, Universidad de Oviedo, Oviedo, Spain.

² Address reprint requests to C Lasheras, Departamento de Biología Funcional, Area de Fisiologia, Facultad de Medicina, Universidad de Oviedo, Julián Clavería s/n, 33006, Oviedo, Spain. E-mail: lasheras{at}correo.uniovi.es.

ABSTRACT  
Background: In studies from Italy and Greece, a Mediterranean dietary pattern was shown to favorably affect life expectancy in the elderly population. This pattern is thought to reduce the risk of cancer in addition to being cardioprotective.

Objective: The objective of this study was to evaluate the interactive effects of the Mediterranean diet and age with respect to survival after controlling for several other variables that could be considered as confounders: age, sex, body mass index, albumin concentration, physical activity, self-assessment of health, and dieting in response to chronic conditions.

Design: This was a cohort study involving 161 nonsmoking elderly subjects (74 subjects aged <80 y and 87 subjects aged 80 y) living in Spain. The subjects were followed up for 9 y. Diet was assessed with a semiquantitative food-frequency questionnaire.

Results: A diet score based on 8 characteristics of the traditional diet in the Mediterranean region was associated with a significant reduction in overall mortality in elderly subjects aged <80 y but not in subjects aged 80 y. A unit increase in the diet score predicted a 31% reduction in mortality in subjects aged <80 y (95% CI: 7%, 57%).

Conclusion: Efforts to promote adherence to the Mediterranean dietary pattern appear to be worthwhile in persons aged <80 y, in whom the diet predicts survival, but we do not have any available evidence that such a diet benefits subjects aged 80 y.

Key Words: Dietary pattern • life expectancy • elderly • longitudinal study • Mediterranean diet • Spain • nonsmokers

INTRODUCTION  
Over the past 30 y, research in the field of nutrition and causes of chronic disease has led to exciting, significant progress in understanding specific risk factors and chemopreventive agents. The major health problems considered were cardiovascular disease and nutritionally linked cancers.

Although great emphasis has been placed on the different components of the diet, attention recently shifted to the diet as a whole. The Mediterranean dietary pattern is thought to reduce the risk of cancer in addition to being cardioprotective (1–4). Recent prospective studies of elderly persons in other Mediterranean countries related specific diet scores, based on the typical components of the Mediterranean diet, to overall survival (5). These components are a high ratio of monounsaturated to saturated fat; moderate ethanol consumption; high consumption of legumes, cereals (including bread and potatoes), vegetables, and fruit; low consumption of milk and dairy products; and low consumption of meat and meat products.

Data sources for previous studies in Spain (6, 7) suggest that, although changes have occurred in the Spanish diet over the past 10–50 y, the diet still includes the main characteristics of the typical Mediterranean diet. Surveys showing associations between the Mediterranean diet and survival consisted largely of subjects belonging to the "young-old" age group (<80 y) (5, 8, 9) but, as life expectancy increases, it is of particular interest to assess the importance of diet as a risk factor in the "old-old" age group (80 y).

Many factors are known to predict mortality, such as age, sex, smoking status, body mass index (BMI; in kg/m²) (10, 11), albumin concentration (12), history of chronic conditions (13), self-rated health (14), and physical activity (15, 16). However, only age, sex, and smoking status were controlled for in the above-mentioned studies. In addition, the follow-up period was shorter in these studies than in the present study.

To evaluate the interactive effects of the Mediterranean diet and age with respect to survival, a prospective study was performed in elderly Spanish subjects who were followed up for 9 y. The effects of several other variables that could be considered as confounders of the association between this dietary pattern and overall survival were controlled for.

SUBJECTS AND METHODS  
The study was conducted in Asturias, in northern Spain, in a cohort of 161 nonsmoking elderly volunteers (112 women and 49 men) living in 2 institutions supported by State Social Security. The subjects were aged 65–95 y; 74 subjects were aged <80 y [24 (32.4%) of these were men] and 87 subjects were aged 80 y [25 (28.7%) of these were men]. To be included in the study, subjects could not be bedridden or confined to a wheelchair and could not be suffering from any terminal disease. All participants were mentally and physically capable of participating in the initial protocol of the study and gave their consent. The study design was approved by the Committee on Ethical Research of the Oviedo University Hospital.

In a preliminary interview, the subjects were informed of the objective of the study; after they agreed to participate, a personal appointment was made to collect data. Interviews were carried out between March 1989 and July 1989 and study subjects were followed up for mortality from the initial visit until December 1998. The average length of follow-up was slightly more than 9.5 y. To determine the vital status of the cohort, the institutions communicated to us the exact date of death of the 96 subjects who died and confirmed the survival of the remaining 65.

Although the subjects were living in institutions, they could choose their meals daily from several set menus. Dietary intake data were collected by trained dietitians using a semiquantitative food-frequency questionnaire on one occasion. Food items were considered in groups, as recommended by Davidson and Passmore (17). Subjects could indicate their usual patterns of consumption of the different foodstuffs on a daily, weekly, monthly, or yearly basis. Amounts consumed were recorded in household units, by volume, or by measuring with a ruler. When this was not possible, participants were asked whether the portions of food eaten were smaller, equal to, or larger than a previously established standard portion (18). Food items were converted into food quantities (g/d) and were further adjusted to daily intakes of 10460 kJ (2500 kcal) for men and 8368 kJ (2000 kcal) for women. Food intakes were reviewed and analyzed for energy and nutrient contents by using the nutrient database developed in Spain by the Institute of Nutrition and Bromatology (CSIC, Madrid).

A diet score, comparable with the score devised for the Greek study on diet and survival in elderly people (5), was developed. The following 8 dietary components each contributed one point to the total score: high ratio of monounsaturated to saturated fat, moderate ethanol consumption, high consumption of legumes, high consumption of cereals (including bread and potatoes), high consumption of fruit, high consumption of vegetables, low consumption of meat and meat products, and low consumption of milk and dairy products. We used the median values specific to each sex and age group as cutoff points. The possible range of scores was 0–8 points. We hypothesized a priori that the higher the score, the better the beneficial health effects of the diet.

In a general, structured interview, information was collected on potential determinants of diet and health (eg, dieting in response to health problems, such as hyperlipemia, diabetes, gastric pathology, cardiovascular disease, or obesity, as well as physical activity assessed with the question "How many minutes per day do you spend walking?" and self-assessments of health through the question "Would you say your health is good, fair, or poor?").

To calculate BMI, height was registered by using a stadiometer exact to 1 mm (Año-Sayol, Barcelona, Spain). Subjects were barefoot, in an upright position, and with the head positioned in the Frankfort horizontal plane. Weight was measured with a weighing machine with 500-g precision (Seca, Hamburg, Germany). Fasting blood samples were taken and serum albumin was calculated by electrophoresis. Subjects were classified as <80 y or 80 y according to their age at the time of enrollment in the study.

The statistical analysis was performed by using SAS (SAS Institute, Inc, Cary, NC). Analysis was undertaken by modeling the data through Cox's proportional-hazards regression. Eight Cox's models were developed for each age group, which controlled for age at enrollment (in y), sex (0 = male, 1 = female), albumin concentration (per 10 g/L), self-assessment of health (0 = not good, 1 = good), physical activity (in min), BMI (per unit increase), and dieting in response to the above-mentioned health problems (0 = yes, 1 = no); the 8 individual components of the diet score, adjusted for energy, were evaluated alternatively. An additional Cox's model for each age group was developed that controlled for age at enrollment, sex, albumin concentration, dieting in response to chronic conditions, BMI, self-assessment of health, and physical activity and evaluated the total diet score as a predictor of the risk of death. The crude association between diet and survival was assessed by using the Kaplan-Meier method.

RESULTS  
Ninety-six of the study subjects died during the follow-up period; 38 (39.6%) of these subjects were aged <80 y and 58 (60.4%) were aged 80 y. Of the 38 subjects in the younger age group who died, 17 (44.7%) were men with a mean (±SD) age at enrollment of 75.2 ± 4.0 y. Thirty-six subjects in this age group survived; 7 (24.1%) were men with a mean age at enrollment of 74.5 ± 3.4 y. Of the 58 subjects aged 80 y who died, 18 (31%) were men with a mean age at enrollment of 84.9 ± 3.4 y. Of the survivors, 7 (36.8%) were men with a mean age at enrollment of 83.8 ± 2.7 y.

Baseline characteristics for selected variables that have been considered as confounders in elderly people with diet scores 3 and 4 are shown in Table 1. Subjects with higher diet scores did somewhat more physical activity than did those with lower scores and were more likely to state that they were in good health and that they were dieting in response to health problems.

View this table:
TABLE 1.. Baseline characteristics of elderly subjects with diet scores 3 (low score) and 4 (high score)  
Median daily consumption (in g), adjusted to energy intake for each of the 8 components of the diet score used as cutoff points according to sex and age group, is shown in Table 2. Men aged 80 y consumed more fruit and dairy products than did those aged <80 y and women aged 80 y consumed more fruit than did those aged <80 y.

View this table:
TABLE 2.. Daily consumption adjusted to energy intake for 8 components of diet score¹  
Mean (±SD) daily consumption (in g) of the 8 components of the diet score, adjusted to 10460 kJ (2500 kcal) for men and 8368 kJ (2000 kcal) for women, and energy intake according to age group and vital status are shown in Tables 3 and 4 along with the rate ratio estimates and 95% CIs derived from alternative Cox's models with each variable analyzed separately. In subjects aged 80 y, dairy products were associated with an increase in the risk of death for every 20-g increase in daily intake, adjusted to energy intake. The other individual models for this age group and the 8 models for the group aged <80 y showed no significant results for any of the components of the diet score.

View this table:
TABLE 3.. Daily consumption adjusted to energy intake for 8 components of diet score and rate ratio estimates (with 95% CIs) derived from alternative Cox's models with each variable analyzed separately in elderly subjects aged <80 y¹  

View this table:
TABLE 4.. Daily consumption adjusted to energy intake for 8 components of diet score and rate ratio estimates (with 95% CIs) derived from alternative Cox's models with each variable analyzed separately in elderly subjects aged 80 y¹  
The results of the Cox's proportional-hazards models with age, sex, total diet score, albumin concentration, dieting in response to chronic conditions, BMI, self-assessment of health, and physical activity as predictor variables for each age group are shown in Table 5. The composite diet score was significantly associated with a reduced risk of death in subjects aged <80 y (by 31%/unit increase) but not in subjects aged 80 y. In the latter group, there was a 60% lower risk of death for every 10-g/L increase in albumin concentration and subjects who rated their health as good had a 52% lower risk of death than did those who rated their health as poor. Exclusion of death occurring in the first year of follow-up did not alter the regression coefficients related to diet, albumin concentration, and self-perceived health. The Kaplan-Meier survival curves for subjects with diet scores 3 and 4 in each age group are shown in Figures 1 and 2. Subjects aged <80 y with low diet scores (<4) had shorter survival times than did those with higher diet scores (4). No differences in survival according to diet score were found in subjects aged 80 y.

View this table:
TABLE 5.. Rate ratio estimates (with 95% CIs) derived from Cox's proportional-hazards model for diet score as a predictor of survival time¹  

View larger version (8K):
FIGURE 1. . Kaplan-Meier survival curves for subjects aged <80 y with diet scores 3 (low score, n = 40) and 4 (high score, n = 34).

 

View larger version (9K):
FIGURE 2. . Kaplan-Meier survival curves for subjects aged 80 y with diet scores 3 (low score, n = 53) and 4 (high score, n = 32).

 

DISCUSSION  
There were some differences between our results and those obtained with the Greek diet (5) (basically, a lower median intake of legumes, a higher intake of dairy products, and a lower ratio of monounsaturated to saturated fat) in the elderly subjects in our study. However, even when there are several variants of the Mediterranean diet, data indicate some similarities in the food patterns of Mediterranean countries: a greater intake of olive oil, cereals, legumes, fruit, and vegetables and relatively less meat and milk consumption than in northern Europe (5, 7, 19–21).

Asturias is a region of Spain, a Mediterranean country, although it does not actually border the Mediterranean Sea. The region is 1 of the 5 Spanish regions participating in the European Prospective Investigation into Cancer and Nutrition. Recent cross-sectional analysis from that project showed that the consumption of fruit and vegetables in the Asturian population aged 29–69 y (as in the other Spanish regions participating in the investigation) was remarkably higher than that in most European countries and the United States (22). Also, only fruit had a consistent trend of increased intake with age, which is contrary to dietary patterns in other countries. Nevertheless, in previous studies (23, 24) we found that the consumption of fruit and vitamin C in the elderly population living in the Asturian region was lower than that in other regions of Spain and other Mediterranean countries. Regarding fat consumption and main sources of specific fatty acids, data from the European Prospective Investigation into Cancer and Nutrition showed that dietary patterns of northern and southern Spain were very similar and were characteristic of the Mediterranean diet (25).

This study, which was conducted in a nonsmoking elderly population, showed that the Mediterranean diet score remained a predictor of mortality after age, sex, BMI, albumin concentration, physical activity, self-assessment of health, and dieting in response to chronic conditions were controlled for. Our results also provided evidence that the ability of this Mediterranean dietary pattern to predict mortality is not consistent across age groups. When evaluating the dietary pattern as a predictor of mortality in subjects aged <80 y, we found results similar to those described previously for Greek and Danish populations (5, 8). That is, a substantial and significant effect of the overall score and a weak and generally nonsignificant association of the individual components of the diet score with survival were noted. However, these results are not applicable to subjects aged 80 y because diet score did not predict mortality in this group, which suggests a more complex relation between diet and survival.

Other studies of the predictors of cardiovascular disease, morbidity, and mortality also showed changes in the strength of an association between a specific risk factor and mortality with age, the most frequent observation being a decline in risk factor–mortality associations with increasing age (11, 26, 27). There are several possible explanations to consider when interpreting this pattern, the first of which is the "survivor explanation": a selective earlier mortality of persons most susceptible to a particular risk factor and relative immunity of survivors (28), a higher prevalence of illness at baseline in subjects aged 80 y, different earlier life experiences altering the later effect of various risk factors, and increased influence of competing causes of death in a population with such high mortality. Furthermore, it is of course also possible that the effect of the diet on mortality changes with age.

Another explanation we postulated was a recent modification of dietary habits as a consequence of diverse health problems that lead to better adherence to dietary recommendations. For this reason we controlled for subjects following such a diet. Results indicated that the different relation between diet and survival between age groups was not due to greater consumption of vegetables and fruit by subjects with health problems and consequently with an increased risk of mortality as a result of these problems.

The only individual components of the Mediterranean diet for which we found a positive association between intake and mortality was dairy products consumed at an age of 80 y. The same result was seen in the Greek study (5). A high intake of dairy products implies a higher consumption of saturated fatty acids and total lipids, which can be related to this significant increase in overall mortality (29). Further studies are required to explore the role of the different fatty acids in survival at 80 y of age. Also, in subjects aged 80 y, we found that poor self-assessed health and lower albumin concentrations were strong, significant predictors of mortality. One study showed an increasing risk of mortality in subjects who assessed their health as poor (14). In another study, low albumin concentrations were found to be a predictor of short-term mortality in institutionalized elderly people (12).

The present study is the only one of its type with data on several variables that could be used to eliminate suspected confounding of the associations between the Mediterranean dietary pattern and survival; it is also the first to adduce evidence for interactive effects of diet and age with respect to survival.

In summary, our findings provide evidence that a Mediterranean diet is beneficial to health, being associated with survival in the young-old age group (<80 y). Efforts to promote adherence to this dietary pattern appear to be worthwhile in elderly persons aged <80 y. Nevertheless, we do not have any available evidence about the benefits that can be expected from initiating changes in diet during later years of life. Perhaps people who have survived beyond 80 y must be allowed to eat their favorite foods and attempts should not be made to modify their dietary habits, particularly if this dietary modification affects the quality of their last years of life.

REFERENCES  

1. Mancini M, Parfitt VJ, Rubba P. Antioxidants in the Mediterranean diet. Can J Cardiol 1995;11(suppl):105G–9G.
2. Tavani A, La Vecchia C. Fruit and vegetable consumption and cancer risk in a Mediterranean population. Am J Clin Nutr 1995; 61(suppl):1374S–7S.
3. Lupo A. Nutrition in general practice in Italy. Am J Clin Nutr 1997;65(suppl):1963S–6S.
4. de Lorgeril M. Mediterranean diet in the prevention of coronary heart disease. Nutrition 1998;14:55–7.
5. Trichopoulou A, Kouris Blazos A, Wahlqvist ML, et al. Diet and overall survival in elderly people. BMJ 1995;311:1457–60.
6. Rodríguez Artalejo F, Graciani MA, Banegas JR, Martín Moreno JM, Sabatér J, Rey Calero J. Food and nutrient consumption in Spain in 1940–1988 (and II). Comparative study of the main sources of information on food consumption. Med Clin (Barc) 1996; 107:446–52.
7. Arija V, Salas Salvadó J, Fernandez Ballart J, Cucó G, Martí-Henneberg C. Consumption, dietary habits and nutritional status of the Reus (IX) population. Evolution of food consumption, energy and nutrient intake and relationship with the socioeconomic and cultural level, 1983–1993. Med Clin (Barc) 1996;106:174–9.
8. Osler M, Schroll M. Diet and mortality in a cohort of elderly people in a north European community. Int J Epidemiol 1997;26:155–9.
9. de Groot LC, van Staveren WA, Burema J. Survival beyond age 70 in relation to diet. Nutr Rev 1996;54:211–2.
10. Woo J, Ho SC, Yuen YK, Yu LM, Lau J. Cardiovascular risk factors and 18-month mortality and morbidity in an elderly Chinese population aged 70 years and over. Gerontology 1998;44:51–5.
11. Davis MA, Neuhaus JM, Moritz DJ, Lein D, Barclay JD, Murphy SP. Health behaviors and survival among middle-aged and older men and women in the NHANES I Epidemiologic Follow-up Study. Prev Med 1994;23:369–76.
12. Sahyoun NR, Jacques PF, Dallal G, Russell RM. Use of albumin as a predictor of mortality in community dwelling and institutionalized elderly populations. J Clin Epidemiol 1996;49:981–8.
13. Takezaki T, Ohashi Y, Ohta T, Tajima K, Tominaga S. Survival factors in healthy aging among rural Japanese residents of Aichi. Nippon Koshu Eisei Zasshi 1996;43:901–8.
14. Idler EL, Angel RJ. Self-rated health and mortality in the NHANES-I Epidemiologic Follow-up Study. Am J Public Health 1990; 80:446–52.
15. Lindsted KD, Tonstad S, Kuzma JW. Self-report of physical activity and patterns of mortality in Seventh-Day Adventist men. J Clin Epidemiol 1991;44:355–64.
16. Rakowski W, Mor V. The association of physical activity with mortality among older adults in the Longitudinal Study of Aging (1984–1988). J Gerontol 1992;47:M122–9.
17. Davidson S, Passmore R. Human nutrition and dietetics. Edinburgh, United Kingdom: Churchill Livingstone, 1979.
18. Ministerio de Sanidad y Consumo. Guia nutricional para centros de ancianos, hogares y residencias. (Nutritional guide for old peoples' home.) Madrid: Secretaría General Técnica, 1985 (in Spanish).
19. Helsing E. Traditional diets and disease patterns of the Mediterranean, circa 1960. Am J Clin Nutr 1995;61(suppl):1329S–37S.
20. Trichopoulou A, Lagiou P. Healthy traditional Mediterranean diet: an expression of culture, history, and lifestyle. Nutr Rev 1997; 55:383–9.
21. Willett WC, Sacks F, Trichopoulou A, et al. Mediterranean diet pyramid: a cultural model for healthy eating. Am J Clin Nutr 1995; 61(suppl):1402S–6S.
22. EPIC Group of Spain: dietary intake of vegetables and fruits among adults in five regions of Spain. Eur J Clin Nutr 1999;53:174–80.
23. Lasheras C, Gonzalez C, García A, Patterson AM, Fernández S. Dietary intake and biochemical indicators of nutritional status in an elderly institutionalized and noninstitutionalized population. Nutr Res 1999;19:1299–312.
24. Lasheras C, Gonzalez C, Patterson AM, Fernández S. Food habits and anthropometric measures in a group of independent and institutionalized elderly people from Northern Spain. J Nutr Sci Vitaminol (Tokyo) 1998;44:757–68.
25. EPIC Group of Spain. Patrones de consumo y principales fuentes de ingestión de lípidos y ácidos grasos en la cohorte española del Estudio Prospectivo Europeo Dieta y Cáncer (EPIC). [Patterns and sources of fat and specific fatty acid intake in the participants of the European Prospective Investigation into Cancer and Nutrition (EPIC) in Spain.] Med Clin (Barc) 1999;112:125–32 (in Spanish).
26. Taylor JO, Cornoni Huntley J, Curb JD, et al. Blood pressure and mortality risk in the elderly. Am J Epidemiol 1991;134:489–501.
27. Langer RD, Ganiats TG, Barrett Connor E. Paradoxical survival of elderly men with high blood pressure. BMJ 1989;298:1356–7.
28. Kaplan GA, Haan MN, Cohen RD. Risk factors and the study of prevention in the elderly: methodological issues. In: Wallace R, Woolson R, eds. The epidemiologic study of the elderly. New York: Oxford University Press, 1992:20–36.
29. Frisoni GB, Franzoni S, Rozzini R, Ferruci L, Boffelli S, Trabucchi M. Food intake and mortality in the frail elderly. J Gerontol A Biol Sci Med Sci 1995;50:M203–10.