Dietary patterns associated with fall-related fracture in elderly Japanese: a population based prospective study

Abstract

Background: Diet is considered an important factor for bone health, but is composed of a wide variety of foods containing complex combinations of nutrients. Therefore we investigated the relationship between dietary patterns and fall-related fractures in the elderly.

Methods: We designed a population-based prospective survey of 1178 elderly people in Japan in 2002. Dietary intake was assessed with a 75-item food frequency questionnaire (FFQ), from which dietary patterns were created by factor analysis from 27 food groups. The frequency of fall-related fracture was investigated based on insurance claim records from 2002 until 2006. The relationship between the incidence of fall-related fracture and modifiable factors, including dietary patterns, were examined. The Cox proportional hazards regression model was used to examine the relationships between dietary patterns and incidence of fall-related fracture with adjustment for age, gender, Body Mass Index (BMI) and energy intake.

Results: Among 877 participants who agreed to a 4 year follow-up, 28 suffered from a fall-related fracture. Three dietary patterns were identified: mainly vegetable, mainly meat and mainly traditional Japanese. The moderately confirmed (see statistical methods) groups with a Meat pattern showed a reduced risk of fall-related fracture (Hazard ratio = 0.36, 95% CI = 0.13 - 0.94) after adjustment for age, gender, BMI and energy intake. The Vegetable pattern showed a significant risk increase (Hazard ratio = 2.67, 95% CI = 1.03 - 6.90) after adjustment for age, gender and BMI. The Traditional Japanese pattern had no relationship to the risk of fall-related fracture.

Conclusions: The results of this study have the potential to reduce fall-related fracture risk in elderly Japanese. The results should be interpreted in light of the overall low meat intake of the Japanese population.

Figure 1

Study protocol. Of these 1178, we excluded 213 subjects who did not agree to the follow up survey, 77 with incomplete dietary data and 11 whose cognitive level was lower than 18 in the Mini Mental State Examination (MMSE) score. Therefore, 877 participants whose medical status, activities of daily living (ADL), and life style including dietary intakes were assessed in July 2002 were followed up until the end of July 2006.

Figure 2

The accumulated rate of fall-related fracture onset in each tertile of the Vegetable pattern. The cumulative fall-related fracture incidence in T3 (confirmed) group of the Vegetable pattern is visibly higher than T1 or T2.

Figure 3

The accumulated rate of fall-related fracture onset in each tertile of the Meat pattern. The cumulative fall-related fracture incidence in T1 (unconfirmed) group of the Meat pattern is higher than T2 or T3.