Association of dietary variety with body composition and physical function in community-dwelling elderly Japanese

Abstract

Objectives

To examine the associations of dietary variety with body composition and physical function in community-dwelling elderly Japanese

Design

Cross-sectional study.

Setting

Community-based.

Participants

A total of 1184 community-dwelling elderly adults aged 65 and over.

Measurements

Dietary variety was assessed with a food frequency questionnaire (maximum, 10 points) that encompassed the 10 main food components of Japanese meals (meat, fish/shellfish, eggs, milk, soybean products, green/yellow vegetables, potatoes, fruit, seaweed, and fats/oils). Body composition was determined by multifrequency bioelectrical impedance analysis. Physical function was assessed by measuring grip strength and usual walking speed. Multiple linear regression analysis was used to examine the associations of dietary variety with body composition and physical function.

Results

After adjusting for potential confounders, higher dietary variety scores were independently associated with higher lean mass (β (SE): 0.176 (0.049), p<0.001) and appendicular lean mass (β (SE): 0.114 (0.027), p<0.001) but not with body fat mass. Elders with a higher dietary variety score had greater grip strength and faster usual walking speed (β (SE): 0.204 (0.071), p=0.004, and β (SE): 0.008 (0.003), p=0.012, respectively).

Conclusion

Greater dietary variety was significantly associated with greater lean mass and better physical function in Japanese elders. The causal relationship warrants investigation in a prospective study.

Introduction

Sarcopenia is defined as age-related decline in lean mass and physical function (1). Prevention of sarcopenia is a major challenge in aged societies, and effective solutions to prevent or delay sarcopenia onset are urgently needed.

Among the factors that affect age-associated decline in lean mass and physical function, dietary factors are potentially modifiable. Previous studies of the association of diet with age-related changes in body composition and physical function focused on certain nutrients and foods, such as protein, antioxidants, fruits and vegetables, dairy products, and fish (2–7). However, food and nutritional items are consumed in combination, not as isolated substances. Therefore, improving overall diet quality may be a more effective approach for preserving lean mass and physical function in later life.

Recently, studies have investigated the associations of the predefined indexes of overall diet quality as assessed by Mediterranean diet and Healthy Eating Index with functional decline (8–11), and reported the beneficial effects of such indexes of overall diet quality on the age-related functional decline. However, these indexes of overall diet quality have limitations. Because based on adherence to Mediterranean diet pattern or the 2005 Dietary Guideline for Americans (DGA), may be difficult to apply to the entire elderly population. Particularly, Japanese elderly have different dietary characteristics from western populations, another analysis may need to clarify the actual effect of diet on lean mass and physical function.

Dietary variety is an important aspect of diet quality. Previous studies used indices of dietary variety to evaluate overall diet quality (12., 13., 14., 15., 16., 17.) and found that the indices were associated with longer survival and lower risk of cognitive decline in community-dwelling older persons (12, 13, 15–17) and with better nutritional status—as assessed by nutrient intake and biochemical and body composition measures—in elderly nursing home residents (14). However, it is not known whether dietary variety is associated with lean mass and physical function, or with sarcopenia in older persons.

In the present study, we used the dietary variety score (DVS) developed by Kumagai et al. The DVS is a foodbased composite score that is determined by calculating consumption frequencies for 10 food groups, namely, meat, fish/shellfish, eggs, milk, soybean products, green/yellow vegetables, potatoes, fruit, seaweed, and fats/oils, which as a whole constitute a large part of Japanese daily main and side dishes. To reflect differences in frequency consumption patterns and simplify scoring, a score of 1 was given if a food item was consumed every day; otherwise the score was zero. This scale is a simple measure of overall diet quality. High dietary variety as assessed by DVS was found to reduce the risk of higher-level functional decline in Japanese elders (18). DVS has also been used in observational and intervention studies of Japanese older adults (19, 20). Thus, we used DVS to examine associations of dietary variety with lean mass and physical function in Japanese older adults.

Materials and methods

Data were derived from the Hatoyama Cohort Study and Kusatsu Longitudinal Study (21, 22). The study designs and protocols have been reported elsewhere (21, 22). The participants were recruited from an age- and area-stratified random samples of adults aged 65–84 years (in the Hatoyama Cohort Study) and beneficiaries aged 65–74 years in the National Health Insurance system and beneficiaries aged 75 years or older receiving Medical Insurance for the Elderly (in the Kusatsu Longitudinal Study). In the present analysis, we used a combined sample from the 2 studies. Because the 10-item food frequency questionnaire and common covariates were collected in the 2012 survey of the Hatoyama Study and in the 2013 survey of Kusatsu Study, data for both time points were used in the present study. Ultimately, 1184 noninstitutionalized, community-dwelling Japanese adults aged 65–97 years (mean age [SD], 74.0 [5.8]) were enrolled from among participants in a comprehensive health examination in September 2012 at the town of Hatoyama town in Saitama Prefecture (in the Hatoyama Cohort Study; n=576) and participants in a comprehensive health examination in July 2013 at the town of Kusatsu Town in Gunma Prefecture (in the Kusatsu Longitudinal Study; n=608) were. All participants provided written informed consent under conditions approved by the Ethics Committee at Tokyo Metropolitan Institute of Gerontology.

Of the 1184 participants, 204 were excluded because of missing data on covariates (education, living arrangement, smoking habit, drinking habit, exercise habit, chewing ability, medical history, hospitalization in the past year, and body mass index) and 21 were excluded because of a missing DVS. After further excluding participants with incomplete date on body composition (n=41), grip strength (n=41), and gait speed (n=61), the final analytical sample was 972 for body composition, 968 for grip strength, and 960 for gait speed.

Excluded individuals were more likely to be female, to be living alone, and to have chronic diseases (hypertension and heart disease). In addition, they had less soft lean mass and appendicular lean mass and lower grip strength. However, age, BMI, DVS, and usual gait speed did not significantly differ between participants included in the study sample and those who were excluded.

Body composition

Body composition was assessed by bioelectrical impedance analysis, using the InBody 720 (Biospace Inc., Seoul, Korea). The accuracy of this method was previously validated, with dual-energy X-ray absorptiometry as the reference (23, 24). A tetrapolar, eight-point tactile electrode system was used. This device separately measures impedance in the right arm, left arm, trunk, right leg, and left leg at six frequencies (1, 5, 50, 250, 500, and 1000 kHz) for each body segment. The manufacturer's proprietary algorithms were used to estimate whole and regional body composition variables. Lean body mass (LM) refers to bone-free lean mass. Appendicular lean mass (aLM) was calculated as the sum of the lean mass in the arms and legs, with the assumption that all non-fat and nonbone tissue was skeletal muscle.

Physical performance tests

The Asian Working Group for Sarcopenia (AWGS) recommends measuring both muscle strength (grip strength) and physical performance (usual gait speed) as a screening tool for sarcopenia (25). Thus, both these physical performance measures were assessed in this study. In addition, both these measures were reported to be highly predictive of subsequent disability and mortality among a general population of older adults (26–29).

Grip strength was measured using Smedley-type hand dynamometers. Two trials were performed with the dominant hand, and the better of the two results was recorded and analyzed.

Usual gait speed was measured over a straight 11-m walkway marked with tape at 3 and 8 m. The time required to walk 5 m was measured, and gait speed (m/s) was calculated. Usual gait speed was measured once.

Assessment of dietary variety

Dietary variety was assessed by the DVS, which was calculated using the consumption frequencies for 10 food items (meat, fish/shellfish, eggs, milk, soybean products, green/yellow vegetables, potatoes, fruit, seaweed, and fats/ oils) during the week before questionnaire administration. A score of 1 was assigned for items «eaten almost daily», and a score of 0 was assigned for items «not eaten almost daily». Thus, the total score for the DVS ranges from 0 to 10. Higher scores indicate greater dietary variety. During the initial phase of the analysis, we categorized participants into quartiles of DVS. However, because of the distribution of DVS, groups of equal size could not be created. The numbers of participants in the third quartiles of DVS was small; thus, participants from the second and third quartiles of DVS were combined into a category labeled «Medium». DVS was ultimately categorized into 3 groups: Low (DVS = 0–2), Medium (DVS = 3–5), and High (DVS ≥6).

Covariates

The covariates included demographic characteristics (ie, age, sex, study site, and education), living arrangement (single, with spouse only, or other), smoking habit (never, former, or current smoker), drinking habit (never/rarely, sometimes, or every day), exercise habit (defined as engaging in exercise, for 20 minutes or longer, more than three times per week during the previous 6 months), self-perceived chewing ability (can chew anything/almost anything, do not chew much), medical history (hypertension, diabetes mellitus, heart disease, stroke, cancer, chronic obstructive pulmonary disease), hospitalization in the past year, body mass index (computed as weight in kilograms divided by height in meters squared).

Statistical analyses

Participant characteristics, by categories of DVS, were compared using weighted one-way analyses of variance, for continuous variables, and the Mantel–Haenszel chi-square test, for categorical variables. Analysis of covariance was used to calculate adjusted means for body composition and physical performance, by categories of DVS. Testing for linear trends was conducted using linear regression with categories of DVS entered as ordinal variables. The association between body composition and DVS was tested by multiple linear regression analysis. The results of associations are expressed as regression coefficients (β) and standard error (SE). Multiple linear regression analysis was adjusted for age, sex, study site, education, living arrangement, smoking habit, drinking habit, exercise habit, self-perceived chewing ability, medical history, hospitalization in the past year, and body mass index. To determine whether the effect of dietary variety on the two physical performance measures was mediated by lean mass, we performed further analyses with lean mass as a covariate. Selection of these potential confounders was based on previous reports indicating their relationship with both the exposure and outcome of interest. All analyses were performed using IBM SPSS Statistics version 20.

Results

Participant characteristics, by categories of DVS, are shown in Table 1. Those with higher DVS were older, more likely to be female, better educated, were less likely to smoke and drink, more likely to have an exercise habit and less likely to experience hospitalization in past year.

Table 1.

Participant characteristics, by categories of dietary variety score

	Dietary variety score
	Low (0-2 points) n=228	Medium (3-5 points) n=448	High (≥6 points) n=296	P*
Age (y)	73.3 (5.6)	73.4 (5.5)	74.6 (5.5)	0.004
Female (%)	41.2	45.5	58.1	<0.001
Education (y)	10.9 (2.8)	11.8 (3.0)	11.6 (3.0)	0.02
Living alone (%)	20.6	15.0	16.2	0.65
Self-perceived chewing ability (%)
can chew anything/most thing	96.9	97.1	99.0	0.11
do not chew much	3.1	2.9	1.0
Smoking (%)
current	13.2	9.8	6.4	<0.001
former	42.1	37.3	26.0
never	44.7	52.9	67.6
Alcohol (%)
everyday	22.8	27.9	18.6	0.02
sometimes	18.4	15.4	11.1
none/rarely	58.8	56.7	70.3
Exercise habit (%)	46.1	56.3	61.5	0.001
Medical history (%)
Hypertension	49.6	45.3	42.9	0.14
Diabetes	13.2	12.9	12.2	0.73
Cancer	11.8	12.1	8.8	0.23
Stroke	7.9	5.4	6.1	0.44
Cardiovascular disease	15.8	15.2	15.2	0.86
COPD	5.3	4.9	4.4	0.64
Hospitalization (%)	11.4	9.4	6.4	0.04
BMI (kg/m2)	23.5 (3.2)	23.1 (2.9)	23.1 (3.3)	0.18

Values are means (SD) or percentages; *P values are based on weighted one-way analyses of variance, for continuous variables, or the Mantel-Haenszel chi-square test, for categorical variables.

Figure 1 shows the adjusted means for body composition and physical performance by category of DVS. Participants with higher DVS had greater lean mass (low vs high DVS: high, 39.0 kg; low, 37.9 kg; P = 0.003, P for trend = 0.006) and greater appendicular lean mass (low vs high DVS: high, 16.8 kg; low, 16.1 kg; P <0.001, P for trend = 0.001). They also had greater grip strength (low vs high DVS: high, 29.1 kg; low, 27.7 kg; P = 0.008, P for trend = 0.005) and a faster usual gait speed (lowest vs highest quartiles: Q4, 1.34 m/s; Q1, 1.31 m/s; P = 0.245, P for trend = 0.075).

Figure 1.

Adjusted means for body composition and physical performance, by categories of dietary variety score

Adjusted regression coefficients for body composition and physical performance per increase in DVS are shown in Table 2. After adjustment for age, sex, study site, and education, DVS was significantly positively correlated with lean mass (ß=0.155, SE=0.055, P=0.005) and appendicular lean mass (ß=0.102, SE=0.030, P=0.001) but not with body fat mass. Similarly, DVS was significantly positively correlated with both physical performance measures (grip strength: ß=0.235, SE=0.071, P=0.001; usual gait speed: ß=0.010, SE=0.003, P=0.001). These associations remained significant even after full adjustment (lean mass: ß=0.176, SE=0.049, P<0.001; appendicular lean mass: ß=0.114, SE=0.027, P<0.001; grip strength: ß=0.204, SE=0.071, P=0.004; usual gait speed: ß=0.008, SE=0.003, P=0.012). In analyses including adjustment for lean mass, the association between DVS and grip strength was attenuated (ß=0.102, SE=0.065, P=0.12). However, the association between DVS and usual gait speed was independent of lean mass (ß=0.006, SE=0.003, P=0.05).

Table 2.

Adjusted regression coefficients for body composition and physical performance per increase in dietary variety score

	LM (kg)		aLM (kg)		Grip strength (kg)		Usual gait speed (m/s)
	β (SE)	p	β (SE)	p	β (SE)	p	β (SE)	p
Model 1	0.155 (0.055)	0.005	0.102 (0.030)	0.001	0.235 (0.071)	0.001	0.010 (0.003)	0.001
Model 2	0.174 (0.049)	<0.001	0.112 (0.027)	<0.001	0.212 (0.071)	0.003	0.008 (0.003)	0.009
Model 3	0.176 (0.049)	<0.001	0.114 (0.027)	<0.001	0.204 (0.071)	0.004	0.008 (0.003)	0.012

Abbreviationls: LM; Lean body mass, aLM; Appendicular lean mass. Model 1 was adjusted for sex, age, study site, and education. Model 2 was adjusted for variables in Model 1 plus living arrangement, smoking and drinking habit, exercise habit, self-perceived chewing ability, and body mass index. Model 3 was adjusted for variables in Model 2 plus medical history (hypertension, diabetes, cancer, stroke, cardiovascular disease, chronic obstructive pulmonary disease), and hospitalization.

Discussion

In this cross-sectional study of Japanese elders, greater dietary variety was significantly positively associated with lean mass and measures of physical function. To our knowledge, this is the first report to suggest that dietary variety, taken as a whole, helps preserve muscle mass and maintain physical function in older adults.

Our results are in line with previous findings regarding associations between other diet quality scores and functional decline in the elderly. The Healthy Eating Index–2005 is one such diet quality score. It reflects the 2005 Dietary Guidelines for Americans and was associated with better physical performance (walking speed and knee extensor power) (10) and a low prevalence of disability (11). The Mediterranean diet score, which adapts principles of the traditional Mediterranean diet and is one of the most commonly researched indices of diet quality, was associated with slower decline in mobility performance (8) and faster walking speed (9).

Although there are some differences in scoring systems and individual components, all 3 indices consider vegetable, fruits, and legumes. The results of studies using other measures of diet quality support the present findings.

Although the mechanisms linking dietary variety to greater lean mass and better physical performance are unclear, several hypotheses have been advanced. DVS comprises 10 foodbased components: meat, eggs, fish/shellfish, milk, green/yellow vegetables, soybean products, potatoes, fruits, seaweeds, and fats/oils. Variety in food intake may be associated with adequate intake of various nutrients, such as protein and antioxidants, which are associated with age-related loss of lean mass and physical function. Five components—meat, eggs, fish/shellfish, milk, and soybean products—are the main sources of protein. Dietary protein is primarily responsible for muscle protein anabolism in elderly adults (33) and is associated with less loss of lean mass (2) and smaller declines in grip strength and repeated chair stands testing (3). Moreover, fruits and vegetables are important sources of antioxidant vitamins (vitamin C and carotene), which are associated with greater skeletal muscle strength (4, 7). In addition, low plasma carotenoid concentrations were associated with poor physical performance, including gait speed and muscle strength (31, 32). With advancing age, increased oxidative stress may be detrimental for skeletal muscle (34), and antioxidants may have a role in protecting against oxidative stress (35). Therefore, antioxidants in fruits and vegetables may help preserve lean mass and physical function. In the present study, over 80% of participants with higher DVS had reported consuming, almost daily, 3 or more food items from the components of meat, eggs, fish/shellfish, milk, soybean products. Similarly, more than 80% of participants with higher DVS ate fruits and vegetables almost every day. Thus, the combination of these nutrients may partly explain the protective associations of dietary variety with lean mass and physical function.

In this study, the association between dietary variety and grip strength was at least in part mediated by lean mass, but the association between usual gait speed and dietary variety was independent of lean mass. A previous study reported that muscle mass was not correlated with walking speed (36). The factors that influence walking ability can be classified into six main physiological subsystems: the central nervous system, perceptual system, peripheral system, muscles, bone and/or joins, and energy production and/or delivery (37). Thus, a reduction in walking speed may reflect the integrated performance of multiple organ systems, which may explain why the association between usual gait speed and dietary variety was independent of lean mass.

The observed correlations were significant but relatively small. Although this might be interpreted to mean that the results are not clinically relevant, if compounded over a greater length of time, modest differences in intake could result in substantial differences in lean mass and physical function. Lean mass and physical function in older persons are affected by many factors, including medical, physical and socioeconomic factors, as well as dietary factors. Diet can be modified and used as a target for interventions. A recent study of a community-based intervention to improve dietary habits and promote physical activity found that DVS was significantly higher in the intervention group than in the control group (20).

The strengths of the present study include the use of a simple measure of dietary variety and the fact that the effects of dietary variety instead of single nutrients were investigated. Moreover, because the study participants were recruited from urban and rural areas and the sample size was relatively large, the present results may be generalizable to the general population of community-dwelling elderly Japanese.

The present study has several limitations. First, causality cannot be inferred from a cross-sectional study. We thus cannot exclude the possibility that food choices determining dietary variety changed due to lower lean mass and physical function. However, the present results suggest that dietary variety is an important target for interventions against sarcopenia among older adults. Dietary variety should be further investigated for its potential to attenuate age-related declines in lean mass and physical function among older adults. Second, a DVS based on a 10-item food frequency questionnaire may not be fully valid. Because DVS evaluates only the number of different food groups consumed every day, it does not directly provide quantitative estimates of food and nutrient intake. Thus, it is unclear whether the associations of dietary variety with greater lean mass and better physical function are independent of the amount consumed or absolute nutrient composition of the diet. Additional studies that include quantitative evaluation of food or nutrient intakes are therefore needed. Finally, although we collected information on several important covariates and confounders, the possibility of unmeasured or residual confounding cannot be excluded. In particular, we did not examine the hydration status of participants before body composition assessment.

In conclusion, higher dietary variety was associated with greater lean mass and better physical performance. Prospective and/or intervention studies are necessary in order to analyze the causal relation between dietary variety and the risk of sarcopenia.

Conflicts of interest

None of the authors has a conflict of interest.

Contribution of the authors

Study concept and design: Yuri Yokoyama, Mariko Nishi, Hiroshi Murayama, and Shoji Shinkai; Acquisition of data: Yuri Yokoyama, Mariko Nishi, Hiroshi Murayama, Hidenori Amano, Yu Taniguchi, Yu Nofuji, Miki Narita, Eri Matsuo, Satoshi Seino, and Shoji Shinkai; Analysis and interpretation of data: Yuri Yokoyama and Shoji Shinkai; Drafting of the manuscript: Yuri Yokoyama; Critical revision of the manuscript for important intellectual content: Yuri Yokoyama, Mariko Nishi, Hiroshi Murayama, Hidenori Amano, Yu Taniguchi, Yu Nofuji, Miki Narita, Eri Matsuo, Satoshi Seino, Yukari Kawano, and Shoji Shinkai; Statistical analysis: Yuri Yokoyama; Study supervision: Shoji Shinkai.

Ethical Standards

The ethics committee at Tokyo Metropolitan Institute of Gerontology approved this study, which complies with the current laws of Japan.