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

Fracture accidents in the elderly reduces their activity of daily life [1] and also increases mortality [2‐4]. Diet is considered an important factor for the maintenance of bone health [5‐7]. Many nutrients, not only calcium [8, 9] and Vitamin D [10], but also phosphorus, vitamin K, strontium and magnesium [11, 12], contribute to bone health Bone is a complex living tissue, however, and a wide spectrum of micronutrients also contribute to its maintenance. Moreover, diets are composed of a wide variety of foods containing complex combinations of nutrients. Therefore, surveys that examine a single nutrient in foods may not adequately account for complicated interactions and cumulative effects on human health.

Tucker et al. [13] and Okubo et al. [14] categorized diets into dietary patterns in order to clarify the relationship between diet and bone mineral density (BMD). Tucker and colleagues reported that a diet with a high fruit and vegetable content appears to have a protective effect on BMD in males while high candy consumption may be associated with low BMD. Okubo et al. demonstrated the possibility that a dietary pattern with high intakes of fish, fruit, and vegetables and a low intake of meats may have a beneficial effect on BMD. Tucker and Okubo's observation, however, was not extended to look for associations between dietary patterns and fractures. There is no report investigating the relation of dietary patterns and fall-related fracture. Furthermore, the population studied by Okubo et al. was made up of pre-menopausal women. Though Tucker et al. studied an elderly population, the dietary habits of people from Western versus Asian countries are entirely different. As is well known, Japanese food is characterized by rice and soy bean products, and contains many types of fish, seafood and vegetables but only small amounts of meat or dairy products [15]. Therefore, in the present study, we examined the relationship between dietary patterns and fractures in elderly Japanese living in a suburb of Sendai, one of the largest cities in Northern Japan.

The present study is a population based prospective study investigating the relationship between dietary patterns and fall-related fractures in elderly Japanese. Three dietary patterns that appeared in our study are similar to the study of Shimazu et al. who studied Japanese middle age to old age (from 40 to 79) [25]. The Vegetable pattern showed a significant trend for the risk of fall-related fracture. The T3 (confirmed group) showed a significant increase in fall-related fracture risk compared to T1 (unconfirmed group) in the Vegetable pattern. In analysis of each food item, vegetables with light green leaves reduced the fall related fracture risk whereas root vegetables and seaweeds increased the risk. Therefore, not all vegetables increases the risk of fall-related fracture, though the Vegetable pattern showed a significant risk increase overall.

In contrast, T2 (moderately confirmed group) in the Meat pattern showed a significant decrease in fall-related fracture risk compared to T1 (unconfirmed group). The trend shown in the meat pattern can be interpreted as T1 group has a tendency to increased risk of fall-related fracture relative to T2 or T3 (see Figure 3).

Our results in dietary pattern analysis appear to contradict previous reports investigating the relationship between dietary patterns and BMD. Tucker et al. [13] reported that a dietary pattern with a high consumption of fruit, vegetables and cereals resulted in greater BMD, while Okubo et al. [14] showed that a Western pattern with a high intake of fat, meat, butter and seasonings was negatively associated with BMD. Single food item analysis in our study also showed that the variety of vegetables reduces the risk of fall related fracture. Only Xu et al. [26] reported that a high intake of meat at a young age reduced the risk of forearm fracture in postmenopausal women. No other researcher has indicated a relationship between the intake of meat and bone health.

Discrepancies between the present and previous studies may be partially explained by differences in population characteristics. All participants in our study were Japanese older than 70 years. The mean meat intake in Japan was only 77.5 g/day in 2002 [27] whereas it reached 242 g/day in the USA in 2000 [28]. Our results should be interpreted as data from a population with low meat consumption. Some nutrients such as proteins and Vitamin Bs contained in meats are known as protective factors for fracture. Figure 2 suggests that the cumulative fall-related fracture incidence in T3 (confirmed) of the Vegetable pattern was higher than T1 or T2. In addition, Figure 3 suggests that the cumulative fracture incidence in T1 (unconfirmed) of the Meat pattern is higher than T2 or T3. In other words, excessive reliance on low meat or high vegetable intake may cause nutritional imbalances and increase the fracture risk in such a population. Nutritional analysis showed an increasing intake of energy, proteins, Vitamin Bs, C, D, K, electrolytes, folic acid and salt from T1 (unconfirmed) to T3 (confirmed) in any dietary pattern. Therefore, all these nutritional factors, not particular ones, may synergistically contribute to the difference of fracture risk among the three dietary patterns.

Dietary patterns should be interpreted in light of regional background. Japanese food culture has been affected by surrounding Asian countries over many years. Interestingly, the greatest naturalist in Chinese history Li Zizhen (1518 to 1593 AD), reported that animal meat such as beef, ram and quail would strengthen bone and muscles in his famous textbook "the General Catalogue of Herbs [29]". Moreover, he stated that light green leaves such as lettuce and cabbage were beneficial for bone health. Our results in a population-based prospective investigation using multivariate analysis may agree more with Li Zizhen than other recent studies.

Our study has several limitations. The number of participants included in the statistical analysis was 877, and the number of fall-related fractures was only 28. Therefore, we were able to adjust few factors in our analysis though many more factors are known to influence the risk of fracture. Also, the limited sample size may affect the statistical detection power. Secondly, though the study design was prospective, dietary data depended on a single cross-sectional investigation in 2002. At that time, all the participants were 70 years old or more, and they were followed for only four years. Therefore, the present study does not reflect long-term dietary habits. Diet at a young age may more strongly influence bone health [26, 30]. However, it is very difficult to avoid biases in longitudinal investigations of lifestyle including diet over decades. The FFQ we used was validated against a 3-day diet record in women, designed to give an accurate description of short-term intake information rather than long-term dietary habits.

Though prior falls were known to lead to increasing risk of subsequent falls [31, 32], we did not record the number of these falls. The Hazard Ratio of fracture, however, showed no difference after adjustment for experience of falls in the previous 6 months. Though our study was conducted on a particular cohort with certain characteristics (people living in a specific region with official health-care insurance), almost all subjects in Japan use the official medical insurance service we consulted. Therefore, our population is representative of the elderly throughout Japan. This is different to other countries i.e. USA. Finally, age at menopause is known to influence BMD [33] but it was not investigated in the present study.

Despite these limitations, the present study suggests that the impact of dietary patterns in the elderly should not be neglected when assessing the risk of fracture. In a population with low meat consumption, such as elderly Japanese, the moderate consumption of meats may reduce the risk of fall-related fracture. In discussions of diet and health, regional dietary habits should be taken into account.

Dietary patterns were related to the risk of fracture in elderly Japanese. The Vegetable pattern increased the risk of fracture. The Meat pattern had a tendency to reduce the risk of fall-related fracture. These results should be interpreted in light of overall low meat consumption in Japan.