Share this article on:

Diet and Overall Survival in a Cohort of Very Elderly People

Fortes, Cristina; Forastiere, Francesco; Farchi, Sara; Rapiti, Elisabetta; Pastori, Gianni; Perucci, Carlo A

Original Articles

We conducted a 5-year cohort study among 162 self-sufficient residents in a public home for the elderly in Rome, Italy, to evaluate the association between the consumption of specific food groups and nutrients and overall 5-year survival. We used a validated, semiquantitative food-frequency questionnaire to assess diet at baseline. Individuals consuming citrus fruit at least twice a week had an adjusted risk of dying that was half that of individuals who consumed citrus fruit less than once a week [relative risk (RR) = 0.52; 95% confidence interval (CI) = 0.28–0.95] (with adjustment for gender, age, education, body mass index, smoking status, cognitive function, and chronic diseases). The adjusted RRs of mortality were 0.38 (95% CI = 0.14–1.01) for consumption of milk and yogurt at least three times a week vs less than once a week; 0.21 (95% CI = 0.08–0.35) for moderate consumption of espresso coffee (1–2 cups weekly) vs less than once a week; and 0.35 (95% CI = 0.17–0.69) for >2 cups a week of espresso coffee vs less than once a week. High levels of intake of ascorbic acid, riboflavin, and linoleic acid were associated with 50–60% decreases in mortality risk. High consumption of meat was associated with a higher risk of mortality (RR = 9.72; 95% CI = 2.68–35.1) among subjects with chronic diseases. Our findings indicate that frequent consumption of citrus fruit, milk, and yogurt; low consumption of meat; and high intake of vitamin C, riboflavin, and linoleic acid are associated with longevity.

From the 1Department of Epidemiology, Lazio Regional Health Authority, and 2National Institute of Nutrition, Rome, Italy.

This study was supported by the Lazio Regional Health Authority.

Submitted June 1, 1999; final version accepted December 1, 1999.

Address correspondence to: Cristina Fortes, Osservatorio Epidemiologico Regionale, Via di Santa Costanza, 53 00198 Rome, Italy.

Aging is associated with a decrease in immune response, a high rate of formation of free radicals, accumulation of free-radical damage, and decreased antioxidant capacity. 1,2 The decreased immune response and the increased lipid peroxidation are, in turn, associated with an increase in the incidence of infections, cardiovascular disease, and cancer. 1,2 In some instances, however, persons 70 years of age and older maintain an immunologic vigor at levels seen among the young. 3 Diet may play a role in maintaining this immunologic vigor and in preventing free-radical damage. Because adults decrease their food intake with age, 4,5 they may also reduce their consumption of antioxidants and nutrients that may be critical to sustaining the immune function. 2 Specifically, deficiencies in vitamin A, zinc, iron, and vitamins B6 and B2 are associated with a depressed immune response, 6 and dietary antioxidants such as vitamins E and C, beta-carotene, and retinol have been reported to inhibit oxidation of low-density lipoprotein, which plays a key role in the development of atherosclerosis and heart disease. 7–9 Furthermore, diets that include consumption of large amounts of fruits and vegetables, which are rich in antioxidants, have been associated with a reduced risk for several types of cancer and cardiovascular disease. 10

Longitudinal studies have indicated that people who follow either the dietary recommendations of the World Health Organization for the prevention of chronic diseases 11–13 or a Mediterranean diet have a lower risk of mortality. 14 To the best of our knowledge, however, no study has specifically evaluated the association between a Mediterranean diet and survival among very old people. The present study was undertaken to examine this association.

Back to Top | Article Outline

Subjects and Methods

Study Population

Between March and May 1993, we contacted 209 individuals 65 years of age and older who were residing in a home for the elderly in Rome, Italy. On the basis of an initial review of clinical records available at the home, interviews with medical staff, and a disability assessment carried out by a senior nurse, we excluded individuals with mental health problems and severe disability. We used the Barthel Index to evaluate disability, with individuals scoring 8 or less out of a possible 20 considered severely disabled. 15 The medical staff completed a checklist for the presence of chronic diseases (diabetes mellitus, cardiovascular disease, stroke, chronic bronchitis, and hepatic and renal disease). Of the initial population, we excluded 8 from the analysis because of mental impairment and 16 because of severe disability; 23 individuals refused to take part in the study. Thus, a total of 162 individuals, mean age 80 years, were included in the analysis. All study participants were basically self-sufficient and mobile at baseline.

A trained person interviewed each participant using a standardized health and nutritional questionnaire that included information on cognitive status, smoking, and eating habits. We used the abbreviated mental test 15 to assess cognitive function, with a score of 7 or less, out of a possible 10, indicating abnormal cognitive function. A single observer took anthropometric measurements (weight, height, and arm span) according to a standardized procedure. 16

Back to Top | Article Outline

Dietary Assessment

We used a semiquantitative weekly food-frequency questionnaire to assess food intake. The 114-item questionnaire was designed and validated by the Italian National Institute of Nutrition. 17 Foods were subdivided into related groups and subgroups based on type of nutrients. On the basis of the reported weekly frequency of consumption, we classified the level of intake of food and food groups as low, medium, or high or as low or high. We categorized daily mean gram intake of micronutrients and percentages of energy intake from macronutrients into quartiles (first quartile = low intake, second or third quartile = medium intake, and fourth quartile = high intake).

Back to Top | Article Outline

Vital Status

We examined files from the Registry Office of the Municipality of Rome to obtain information on vital status. We calculated the length of follow-up for each participant as the number of days from completion of the baseline questionnaire (May 1993) to the date of death or to May 31, 1998, whichever came first.

Back to Top | Article Outline

Statistical Analysis

Overall mortality was the outcome under study. We used the Kaplan-Meier method to calculate crude survival by baseline characteristics and by level of intake of food, food groups, and nutrients. We used the Cox proportional hazards model to investigate the association between mortality and foods, subgroups of foods, and nutrients, with adjustments for age at baseline (in tertiles), gender, number of years of education (<6, 6–8, >8), body mass index (BMI; kg/m 2) (low = ≤21; medium = 21.1–30; high = >30), cigarette smoking (nonsmokers, current smokers, and ex-smokers), cognitive status (normal and abnormal), and presence of chronic disease (diabetes mellitus, cardiovascular disease, stroke, chronic bronchitis, and hepatic and renal disease). We assessed the potential for violation of the proportional hazards assumption by comparing the survival curves for each level of a variable and by fitting models containing an interaction term between the variable of interest and a log-time variable.

Back to Top | Article Outline


Of the 162 individuals included in the analysis, 53 (33%) died during the 5 years of follow-up. Table 1 shows the main characteristics of the study population, 5-year survival, and adjusted relative risks (RRs) of death. The RRs were estimated using a Cox model and have been adjusted for all of the variables presented in the table. Five-year mortality increased with older age, lower educational level, lower BMI, past or current smoking, abnormal cognitive function, and the presence of chronic disease, although in some instances, the effects were not measured with much precision.

Table 1

Table 1

Table 2 shows the 5-year crude survival by level of food intake and the crude RRs and 95% confidence intervals (CIs) estimated with the Cox proportional hazards models. Survival was highest for the following: medium (1–4 times weekly) and high (>4 weekly) intake of pasta, medium (1–3 times weekly) and high (>3 times weekly) intake of milk and yogurt, and medium (1–2 times weekly) and high (>2 times weekly) intake of oranges and tangerines. A protective effect was also found for medium (1–2 times weekly) and high (>2 times weekly) intake of vegetables rich in carotenoids, for the daily consumption of olive oil as a condiment, and for medium (1–2 times weekly) and high (>2 times weekly) intake of espresso coffee. A negative effect on survival was observed for medium (once a week) and high (>1 times weekly) intake of meat.

Table 2

Table 2

The average intake of all nutrients was greater than the levels recommended for persons 65 years of age or over, with the exception of zinc (intake was 34% lower than the recommended nutrient intake) (Table 3). Crude survival was higher among individuals with medium (146–251 mg) and high (>251 mg) intake of vitamin C, high intake (>2.7 mg) of riboflavin, medium (1,063–1,520 gm) and high (>1,520 gm) intake of calcium, medium (7.2–11.3 mg) and high (>11.3 mg) intake of linoleic acid, medium (1.1–1.5 mg) and high (>1.5 mg) intake of linolenic acid, a high percentage (>23%) of intake of vegetable fat, and high caloric intake (>2,900 kcal). Table 4 shows the adjusted RRs of mortality and 95% CIs from multivariate Cox models for the individual components of the diet. Results are presented for food, food groups, and nutrients that appeared in the crude analysis to be somehow associated, either positively or negatively, with mortality (that is, the effect at either bound of the 95% CI was greater than a 20% increase or decrease in risk). After adjusting for potential confounders (gender, age, educational level, BMI, smoking, cognitive function, and chronic disease), the protective effect remained for high intake of milk and yogurt (RR = 0.38; 95% CI = 0.14–1.01), high intake of oranges and tangerines (RR = 0.52; 95% CI = 0.28–0.95), and medium (RR = 0.21; 95% CI = 0.08–0.35) and high (RR = 0.35; 95% CI = 0.17–0.69) intake of espresso coffee. The negative association with mortality also remained for high intake of ascorbic acid (RR = 0.46; 95% CI = 0.19–1.08), riboflavin (RR = 0.38; 95% CI = 0.16–0.90), and linoleic acid (RR = 0.49; 95% CI = 0.24–1.0). By contrast, after adjustment, some of the effect observed for pasta, olive oil, vegetables rich in carotenoids, and high consumption of meat diminished (Table 4). The analysis was also repeated for those subjects without chronic diseases (93 subjects). The negative or positive effect of certain foods, food groups, and nutrients on mortality remained, but with wider CIs because of the small numbers (data not shown). Consumption of meat more than once weekly, however, was associated with a higher risk of mortality (RR = 9.72; 95% CI = 2.68–35.1).

Table 3

Table 3

Table 4

Table 4

Back to Top | Article Outline


Although the dietary intake of this healthy elderly population was on average above the recommended levels for this age group, there was enough variability in nutrient intake to evaluate high and low values and their effect on the long-term outcome. In addition to adjusting for age, gender, and educational level as a marker of socioeconomic status, we also adjusted for other potential confounding factors. We were not, however, able to assess some potential confounding variables, such as the presence of mastication problems that could interfere with eating fruits, or social support from relatives who could help with food shopping or cooking meals.

To assess the relation between diet and survival, we considered foods, food groups, and nutrient composition of the diet for two reasons: (1) epidemiologic analyses based on foods, as opposed to nutrients, are generally most directly related to dietary recommendations, and (2) inference for causality may be strengthened when an association is observed with overall intake of a nutrient and also with more than one food source of the same nutrient. Fat, protein, and carbohydrates were expressed as percentages of energy intake, because energy is derived from these macronutrients; this categorization allowed us to correct for those differences in intake that were due to differences in energy intake among individuals.

In our dataset, fruits rich in vitamin C were correlated with total vitamin C intake (r = 0.60), and both were negatively associated with mortality. These findings are consistent with two studies conducted in elderly populations. 18,19 In a study of 747 elderly persons, Sahyoun et al 18 found that overall mortality was lower among those with higher plasma vitamin C and total intake of vitamin C. Gale et al, 19 in a 20-year follow-up study of a cohort of 730 individuals ages 65 and over, showed that mortality from stroke was highest among those with the lowest vitamin C intake. At least three hypotheses have been proposed to explain the protective effect of vitamin C against overall mortality in the elderly: (1) vitamin C acts as an antioxidant, mediates the oxidation of low-density lipoprotein cholesterol, and protects against cardiovascular disease 20; (2) vitamin C may influence the biological response to infection and hence the risk of respiratory and cardiovascular diseases 21,22; and (3) vitamin C may prevent oxidative damage and mutagenesis of DNA. 23

In our study, milk and yogurt consumption was associated with increased survival; by contrast, Trichopoulou and colleagues 14 found that dairy products were associated with increased mortality. In the latter study, however, dairy products included butter and cheese, which are rich in saturated fat, whereas in the diet of Italian elderly, milk is a very rich source of riboflavin and calcium. 24 We found that riboflavin was also associated with decreased mortality. It has been suggested that high levels of riboflavin protect tissues from oxidative injury and decrease carcinogen DNA binding. 25 Because elderly persons present abnormal values of tissue-bound coenzymes such as erythrocyte glutathione reductase activation coefficient, even with adequate intake of riboflavin, 26,27 they possibly require a greater intake of riboflavin than younger adults, owing to either malabsorption or abnormal metabolism. Regarding the observed protective effect of calcium on mortality, an antihypertensive effect of calcium carbonate was first indicated by an inverse association between stroke mortality and water hardness. 28 Successive epidemiologic studies conducted in different countries have shown a fairly consistent inverse association between blood pressure and calcium intake from foods. 29,30

The role of polyunsaturated linoleic fatty acid on cardiovascular disease and cancer has been a topic of debate in recent years. 31,32 Clinical trials in which diets low in saturated fats were supplemented with polyunsaturated fatty acids (PUFAs) have demonstrated a reduced overall mortality from all causes and reduced cardiovascular mortality, 33 whereas other studies have linked high levels of PUFAs with a predisposition to tumorigenesis. 34 In our study, linoleic fatty acid, which was correlated with PUFAs (r = 0.937) and vegetable fat (r = 0.444), was inversely associated with overall mortality. The main food source of vegetable fat in our population was olive oil, which showed a protective effect in a univariate analysis. The magnitude of this effect decreased after adjustment, however, which may be explained by the way in which olive oil was categorized in our study as “use” and “non-use.” It has been suggested that olive oil has a protective effect against both cancer and cardiovascular disease, 35,36 yet, in addition to the fatty composition of olive oil, this effect may be due to certain natural components, such as polyphenols, as well as to tocopherol and carotenoids with antioxidant action. 36

A meta-analysis has indicated that individuals who drink five cups of coffee per day may have an increased risk of cardiovascular disease, although the evidence remains ambiguous regarding the magnitude of the effect. 37 This increased risk has been attributed to the capacity of coffee to raise the concentration of total low-density-lipoprotein cholesterol. By contrast, high intake of coffee is associated with a lower risk of colorectal cancer. 38 In our study, coffee consumption was associated with a protective effect against overall mortality. Coffee contains antioxidants (for example, polyphenols and isoflavonoids), which may contribute to its protective effect against cancer and overall mortality. 39,40 Moreover, it has been suggested that high levels of serum total cholesterol are not a risk factor of cardiovascular disease in elderly populations and that mortality from cancer and infection is lower among those elderly persons with the highest intake of cholesterol. 41 Another possible explanation for our findings is that some constitutional risk factors or factors related to a disease status led to both the avoidance of coffee and to a higher risk of mortality.

Among subjects without chronic diseases, an increased risk of mortality was found among people who ate meat regularly. These data are in agreement with other studies that have shown a reduction in all-cause, cancer, and cardiovascular mortality among people who do not eat meat. 42–44

Back to Top | Article Outline


We thank Luigia De Gasperis, the director of the elderly home, and her medical and nursing staff for helping us to conduct this study. We thank Margaret Thorogood for her comments on the paper. We also thank Francesca Zampieri for collaborating on data collection. Finally, we thank all of the individuals living in the community for their enthusiastic participation.

Back to Top | Article Outline


1. Nohl H. Involvement of free radicals in ageing: a consequence or cause of senescence. Br Med Bull 1993; 49:653–667.
2. Meydani SN, Blumberg JB. Nutrition and immune function in the elderly. In: Munro HN, Dandford DE, eds. Nutrition, Aging and the Elderly. New York: Plenum Press, 1989; 123–150.
3. Firky ME. Nutrition and the elderly. World Health 1981; 1:13–15.
4. Garry P, Goodwin JS, Hunt WC, Gilbert BA. Nutritional status in a healthy elderly population: vitamin C. Am J Clin Nutr 1982; 36:332–333.
5. Sjogren A, Osterberg T, Steen B. Intake of energy nutrients and food items in a ten-year cohort comparison and in a six-year longitudinal perspective: a population study of 70- and 76-year-old Swedish people. Age Ageing 1994; 23:108–112.
6. Tuormaa TE. A brief review of the immune system and its function in relation to: PVFS, non-antibody mediated allergy, autoimmunity and immune deficiency. Nutr Health 1988; 6:53–62.
7. Witzum JL. The oxidation hypothesis of atherosclerosis. Lancet 1994; 17:793–798.
8. Rice-Evans C, Diplock AT. Current status of antioxidant therapy. Free Radic Biol Med 1993; 15:17–96.
9. Livrea MA, Tesorieri L, Bongiorno A, Pintaudi AM, Ciaccio M, Riccio A. Contribution of vitamin A to the oxidation resistance of human low density lipoproteins. Free Radic Biol Med 1995; 18:401–409.
10. Serdula MK, Byers T, Mokdad AH, Simoes E, Mendlein JM, Coates RJ. The association between fruit and vegetable intake and chronic disease risk factors. Epidemiology 1996; 7:161–165.
11. Farchi G, Fidanza F, Mariotti S, Menotti A. Is diet an independent risk factor for mortality? 20 year mortality in the Italian rural cohorts of the Seven Countries Study. Eur J Clin Nutr 1994; 48:19–29.
12. Huijbregts P, Feskens E, Rasanen L, Fidanza F, Nissinen A, Menotti A, Kromhout D. Dietary pattern and 20 year mortality in elderly men in Finland, Italy, and the Netherlands: longitudinal cohort study. BMJ 1997; 315:13–17.
13. Osler M, Schroll M. Diet and mortality in a cohort of elderly people in a north European community. Int J Epidemiol 1997; 26:551–559.
14. Trichopoulou A, Kouris-Blazos A, Wahlqvist ML, Gnardellis C, Lagiou P, Polychronopoulos E, Vassilakou T, Lipworth L, Trichopoulos D. Diet and overall survival in elderly people. BMJ 1995; 311:1457–1460.
15. The Royal College of Physicians of London and the British Geriatrics Society. Standardized assessment scales for elderly people: a report of joint workshops of the research unit of the Royal College of Physicians and the British Geriatrics Society. London: Royal College of Physicians of London Editors, 1992; 6–15.
16. Lohman T, Roche AF, Martorell R. Anthropometric standardization reference manual. Champaign, IL: Human Kinetics Books, 1988.
17. Freudenheim JL, Krogh V, D’Amicis A, Scaccini C, Sette S, Ferro-Luzzi A, Trevisan M. Food sources of nutrients in the diet of elderly Italians: macronutrients and lipids. Int J Epidemiol 1993; 22:855–868.
18. Sahyoun NR, Jacques PF, Russell RM. Carotenoids, vitamins C and E, and mortality in an elderly population. Am J Epidemiol 1996; 144:501–511.
19. Gale CR, Martyn C, Winter PD, Cooper C. Vitamin C and risk of death from stroke and coronary heart disease in cohort of elderly people. BMJ 1995; 310:1563–1566.
20. Knekt P, Reunanen A, Jarvinen R, Seppanen R, Heliovaara M, Aromaa A. Antioxidant vitamin intake and coronary mortality in a longitudinal population study. Am J Epidemiol 1994; 139:1180–1189.
21. Schartz J, Weiss ST. Relationship between dietary vitamin C intake and pulmonary function in the first national health and nutrition examination survey (NHANES). Am J Clin Nutr 1994; 59:110–114.
22. Khaw KT, Woodhouse P. Interrelation of vitamin C, infection, haemostatic factors, and cardiovascular disease. BMJ 1995; 310:1559–1563.
23. Knekt P, Jarvinen R, Seppanen R, Rissanen A, Aromaa A, Heinonen OP, Albanes D, Heinonen M, Pukkala E, Teppo L. Dietary antioxidants and the risk of lung cancer. Am J Epidemiol 1991; 134:471–479.
24. Krogh V, Freudenhein JL, D’Amicis A, Scaccini C, Sette S, Ferro-Luzzi A, Trevisan M. Food sources of nutrients of elderly Italians. II. Micronutrients. Int J Epidemiol 1993; 22:869–877.
25. Pancrekar J, Krishnaswamy K, Jagadeesan V. Effects of riboflavin deficiency and riboflavin administration on carcinogen-DNA binding. Food Chem Toxicol 1993; 31:745–750.
26. Thurnham DI. The interpretation of biochemical measurements of vitamin status of the elderly. In: Kemm JR, Ancill RJ, eds. Vitamin Deficiency in the Elderly. Oxford: Blackwell Scientific Publications, 1985; 46–67.
27. Madigan M, Tracey F, McNulty H, Eaton-Evans J, Coulter J, McCartney H, Strain JJ. Riboflavin and vitamin B-6 intakes and status and biochemical response to riboflavin supplementation in free-living elderly people. Am J Clin Nutr 1998; 68:389–395.
28. Morris JN, Crawford MD, Heady JA. Hardness of local water supplies and mortality of cardiovascular disease in the county boroughs of England and Wales. Lancet 1961; 1:860–862.
29. García-Palmieri MR, Costas R, Cruz-Vidal M. Milk consumption, calcium intake and decreased hypertension in Puerto Rico. Hypertension 1984; 6:322–328.
30. McCarron DA, Morris CD, Henry HJ, Stanton JL. Blood pressure and nutrient intake in the United States. Science 1984; 224:1392–1398.
31. Ravnskov U. The questionable role of saturated and polyunsaturated fatty acids in cardiovascular disease. J Clin Epidemiol 1998; 6:443–460.
32. Zock P, Katan MB. Linoleic intake and cancer risk: a review and meta-analysis. Am J Clin Nutr 1998; 68:142–153.
33. Oliver MF. It is more important to increase the intake of unsaturated fats than to decrease the intake of saturated fats: evidence from clinical trials relating to ischemic heart disease. Am J Clin Nutr 1997; 66:980S–986S.
34. Cantrill RC, Yung-Sheng H. Fatty acids and cancer. Nutrition 1998; 14:235–236.
35. Lipworth L, Martínez ME, Angell J, Hsieh C-c, Trichopoulos D. Olive oil and human cancer: an assessment of the evidence. Prev Med 1997; 26:181–190.
36. Visioli F, Galli C. The effect of minor constituents of olive oil on cardiovascular disease: new findings. Nutr Rev 1988; 56:142–147.
37. Greenland S. A meta-analysis of coffee, myocardial infarction, and coronary death. Epidemiology 1993; 4:366–374.
38. Centanze S, Boeing H, Leoci C, Guerra V, Misciagna G. Dietary habits and colorectal cancer in a low risk area: results from a population-based case-control study in Southern Italy. Nutr Cancer 1994; 21:233–246.
39. Shi X, Dalal NS, Jain AC. Antioxidant behaviour of caffeine: efficient scavenging of hydroxyl radicals. Food Chem Toxicol 1991; 29:1–6.
40. Mazur WM, Wahala K, Rasku S, Salakla A, Hase T, Adlercreutz H. Lignan and isoflavonoid concentrations in tea and coffee. Br J Nutr 1998; 79:37–45.
41. Weverling-Rijnsburger AWE, Blauw GJ, Lagaay AM, Knook D, Meinders AE. Total cholesterol and risk of mortality in the oldest old. Lancet 1997; 350:1119–1123.
42. Burr ML, Burtland BK. Heart diseases in British vegetarians. Am J Clin Nutr 1988; 48:830–832.
43. Chang-Claude J, Frentzel-Byme R. Dietary and lifestyle determinants of mortality among German vegetarians. Int J Epidemiol 1993; 22:226–228.
44. Thorogood M, Mann J, Appleby P, McPherson K. Risk of death from cancer and ischaemic heart disease in meat and non-meat eaters. BMJ 1994; 308:1667–1671.

fruit consumption; ascorbic acid; riboflavin; aging; mortality; diet; fatty acids

Copyright © 2000 Wolters Kluwer Health, Inc. All rights reserved.