OBJECTIVE

A narrow range of food choices may lead to dietary inadequacies, a particular concern in elderly people. We hypothesized that consumption of a more diverse diet would predict better nutritional status in frail elderly persons.

SUBJECTS

Subjects included 98 frail nursing home residents (36 men, 62 women), mean age 87.1+/-5.5 (72 to 98) years.

METHODS

3-day dietary variety scores ranging from 23 to 48 and fruit and vegetable variety scores ranging from 5 to 20 were calculated from weighed 3-day food records as the number of different food or fruit and vegetable choices consumed. A higher score indicates a more varied diet. Nutritional status was assessed by weight, height, body mass index (BMI), skinfold thickness, circumference measures, calculated mean arm muscle area, total body water, computerized tomography of the thigh, and total body potassium, as well as nutritional analysis, biochemical measures, and subject medical history.

STATISTICAL METHODS

Univariate regression analyses were performed to investigate the relationship between clinical and nutrition variables. Multiple linear regressions were used to develop models relating dietary variety scores to possible etiologic factors as well as indicators of nutritional status. Models were controlled for age, BMI, and energy intake when appropriate.

RESULTS

Mean dietary variety score was 35.2+/-4.5, and mean fruit and vegetable variety score was 11.3+/-3.0. Higher dietary variety score was associated with higher energy intake (beta = 20.5, P < .001) and both high dietary variety score and fruit and vegetable variety score were positively associated with intake for many nutrients (P < or = .05). High dietary variety score was related to high fruit and vegetable variety score and total intake of fruits and vegetables. In men, higher dietary variety score and fruit and vegetable variety score were associated with higher high-density lipoprotein (beta = 1.02), lower very-low-density lipoprotein (beta = -3.58) and triglycerol (beta = -3.51), and higher blood folate (beta = 4.72) concentrations in women (P < or = .05). In women, high dietary variety score was associated with higher BMI (beta = 0.34, P < .001) and higher total body potassium (beta = 1.30, P = .02); high fruit and vegetable variety score was associated with higher BMI (beta = 0.41), mid-arm circumference (beta = 0.34), and mid-arm muscle area (beta = 2.94) (P < or = .03). Dietary variety score was higher (mean 37.6+/-5.38 vs 34.6+/-4.14) in those who received assistance with feeding (beta = 2 .67, P = .01). History of cancer (beta = -2.04) and gastrointenstinal cancer (beta = -3.54) were associated with low dietary variety score (P < or = .05).

CONCLUSIONS

The results of this study suggest that a highly varied diet in elderly nursing home residents is associated with better nutritional status as assessed by nutrient intake, biochemical measures, and body composition measures. Dietary variety score is a straightforward tool for screening and identifying people at nutritional risk, as well as a mechanism for monitoring response to nutritional, medical, and environmental interventions. Preventive measures to improve dietary variety, as measured by the dietary variety score, should be evaluated and introduced before nutrition and health complications arise.