Introduction

Age-related loss of muscle strength is a characteristic feature of sarcopenia [1], which is in turn related to falls [2,3], low physical performance [3], prolonged hospital stay [4], and increased healthcare and hospitalization costs [5,6]. Age-related loss of muscle strength starts around the fourth decade of life [7,8], and its early prevention is an important target of public health intervention in the ageing population.

Insufficient dietary protein intake is a cause of muscle weakness in elderly individuals [9]. Protein is a component of the skeletal muscle, and its intake stimulates muscle protein synthesis in humans [10]. Experimental studies reported that the response of muscle protein anabolism to dietary protein is impaired with age and that elderly individuals require more dietary protein to maintain nitrogen balance and to stimulate muscle protein synthesis than the young [10–14]. The protein intake requirement is further increased due to both acute and chronic diseases that are more prevalent among older adults [15–17]. Therefore, health authorities in some countries have adopted higher protein requirement values for elderly individuals in the Dietary Reference Intakes (DRI), including in Japan [18–22].

However, the age by which the protein intake requirement increases to maintain muscle protein remains unclear. Currently, the age threshold by which a higher protein intake is recommended in DRI varies across countries; it starts from age ≥75 years in France [20], from age ≥70 years in Australia, New Zealand, and Japan [19,21]; and from age ≥65 years in Germany, Austria, Switzerland, Denmark, Finland, Iceland, Norway, and Sweden [18,22]. In the present study, we aimed to find the age threshold at which a higher protein intake is associated with higher muscle strength using a population-based panel survey of community-dwelling people aged ≥50 years.

Materials and methods

Results

The mean (± SD) age of the participants in the first-wave survey was 62.4 ± 7.0 years, and 46.6% were men. Among the 9,485 observations, 6,334 (66.8%), 1,703 (18.0%), and 1,448 (15.3%) were obtained from municipalities where the first-wave survey was conducted in 2007, 2009, and 2011, respectively. A total of 2,037 (21.5%) observations were classified as protein intake lower than the recommended limit (13% energy) [19]. The baseline characteristics of the participants among the 10-year age groups are shown in Table 1. The participants in their 70s were less likely to have protein intake lower than the recommended limit (p<0.0001) [19], but they were also more likely to have limitations in IADL and mobility (p<0.0001), reported poorer self-perceived health status (p<0.0001), and had lower grip strength (p<0.0001). Energy and carbohydrate intake for men aged ≥70 years was significantly lower than that for men aged 50–59 years. Total fat intake for women aged ≥70 years was significantly lower than that for women aged 50–59 years.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Baseline characteristics of the study population.

https://doi.org/10.1371/journal.pone.0208169.t001

The marginal effects of protein intake on grip strength at the average of the covariates in two-year age strata are shown in Table 2. After adjustment for all covariates, a higher protein intake associated with higher grip strength was noted for women aged ≥65 years. For men, this association was shown in the older age strata, that is, only those aged ≥75 years. The magnitude of these effects increased as age increased in women. Although positive associations were also shown in men aged 69–74 years and women aged 61–64 years, the associations were not statistically significant.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Adjusted marginal effect for grip strength per unit of energy-adjusted protein intake.

https://doi.org/10.1371/journal.pone.0208169.t002

Discussion

We analyzed the data stratifying the participants into two-year age strata and found that higher protein intake became positively and significantly associated with higher grip strength for men aged ≥75 years and women aged ≥65 years.

Previous experimental studies have shown that the required protein intake for elderly individuals to meet nitrogen balance and to stimulate muscle protein synthesis is higher than that for the young [10–14]. Based on the results of experimental studies, elderly individuals have been encouraged to consume higher protein than the young [11,15,16]. Our results showed the positive effects of protein intake on grip strength only in older participants and supported the findings of the previous experimental studies.

Meanwhile, the results of epidemiological studies examining the association between protein intake and muscle strength are inconsistent. A cohort study conducted in the UK reported a positive association between protein intake and grip strength at baseline and at 1.5 years follow-up [31], and other cohort studies in the US also reported a positive association between protein intake and muscle strength at six years follow-up [32,33]. A cross-sectional study in Italy reported that participants with higher protein intake showed higher grip strength regardless of physical activity level [34]. By contrast, cohort studies conducted in Australia and Italy reported no significant association between protein intake and knee extension strength [35,36], and two cross-sectional studies in the US showed that protein intake had no significant effect on grip strength [37,38]. These inconsistent results among the previous studies may be due to differences in study design, study population, dietary assessment method, or indicator of muscle strength used in the studies. In addition, we speculated that these inconsistencies might be because the participants were not stratified by age.

A few epidemiological studies analyzed data stratifying participants by age. Mulla et al. reported no significant association between protein intake at age 36 and 43 years and grip strength at age 53 years [39]. Thompson et al. reported that protein intake had smaller effect on muscle strength of the leg flexors in middle-aged adults than in older adults [40]. Nevertheless, evidence that indicates the age threshold at which a higher protein intake is required to prevent the age-related loss of muscle strength is scarce. We found that higher protein intake is positively associated with higher grip strength for men aged ≥75 years and women aged ≥65 years. To the best of our knowledge, this study is the first to attempt to find evidence for the age threshold at which higher protein intake is recommended for the prevention of muscle weakness.

Meanwhile, our result showed that dietary protein intake had significant negative association with grip strength for men aged 50–62 years and women aged 50–24 years. Theoretically, it is unlikely that a higher protein intake has a negative effect on the maintenance of muscle strength for the middle-aged and the younger elderly. We speculate that a reverse causation may exist, or those with grip strength weaker for young age would take more protein-rich nutrition. This result implies that dietary intervention alone might be a less effective strategy to prevent age-related loss of muscle strength for middle-aged people. For the younger population, strategies other than dietary interventions to maintain muscle strength might be appropriate, such as exercise.

Another main finding was that the age at which the positive effect of a higher protein intake started to appear was higher in men than in women (≥75 years for men and ≥65 years for women). This result indicated the possibility that women needed a high protein intake from a younger age compared with men. Some previous studies also showed a sex difference in the association between protein intake and muscle strength, which was consistent with our results. The Hertfordshire Cohort Study reported that protein intake was inversely associated to grip strength in women but not in men [41]. Another cohort study showed that lower protein intake was associated with lower baseline grip strength only in women [31]. Since this study is cross-sectional, further studies are needed to clarify from when a higher protein intake is recommended to prevent muscle weakness.

We have two possible explanations for this sex difference. First, the response to ingested protein may be different by sex. An experimental study reported that the response of muscle protein anabolism that was stimulated by protein ingestion was lower in elderly women than in elderly men [42]. This implies that the negative effect of low protein intake is more significant for women than for men. Second, other factors, such as sex hormones, might have a stronger impact on the prevention of muscle weakness in men. Sex hormones have a beneficial effect on maintaining the muscle function [43]. Because testosterone decreases gradually from middle age while estrogen decreases sharply at menopause, age-related decrement in muscle strength might occur later in men than in women [43,44].

Our result showed that protein intake positively associated with grips strength for men aged ≥75 years and for women aged ≥65 years. However, it is difficult to determine whether very high protein intake (such as >30% energy) is recommended for higher muscle strength for the following reasons: First, the tolerable upper intake level of protein is not determined because the available evidence on the tolerable upper intake level of protein is not sufficient. Next, protein intake of many of our study population was less than 20% energy, and we do not have sufficient sample size to examine the effect of higher protein intake (such as >30% energy) on muscle strength in each age stratum. Further studies are needed to examine whether very high protein intake (such as >30% energy) is recommended for higher muscle strength using datasets that include sufficient number of participants with high protein intake.

The strength of this study is the use of a large sample of community-dwelling adults in a panel study that allows data analysis with stratification by 2-year age intervals. However, there are also limitations to our study. First, the cross-sectional study design does not allow assessment of causality because the temporality of the association is uncertain. Although we assessed the dietary habits of the participants for one month before grip strength was measured, the possibility of reverse causality cannot be excluded.

Second, the protein intake obtained from the BDHQ is an estimated protein intake. To minimize the effect of misreporting, we excluded participants whose energy intake was extremely low or high and used energy-adjusted values in the analysis [29]. Also, the validity of the BDHQ for estimating protein intake was limited. The validity of protein intake estimated from dietary questionnaires for Japanese tends to be lower than those for Westerners because of the complexity of the Japanese diet [45]. The median validity for estimated protein intake with 19 dietary questionnaires for Japanese was 0.39 [45]. Although the BDHQ seems to have an acceptable ability to estimate Japanese protein intake, the results should be interpreted with caution.

Third, several variables, such as body height, weight, and limitation of IADL or mobility, were based on self-reports, and thus misclassification may occur. In addition, because grip strength was measured only once, measurement error may occur.

Fourth, the response rate for the first-wave survey varied across cities (45.9%–87.8%) [23–25]. We could not include the city variable as a potential confounder in our analysis because the JSTAR’s public dataset did not include information on the cities where the participants lived. To reduce the effect of the difference in the response rate, we included the survey year variable into the multivariate analysis; however, this difference may bias the results.

Finally, there were some unavailable variables that might affect the association between protein intake and muscle strength. For example, we could not obtain the data on muscle mass, the presence of chronic kidney disease, and physical activity of participants. Although we adjusted for various potential confounding variables, the effects of unobserved confounders may not have been excluded because of these unobserved variables that can influence the maintenance of muscle strength and protein intake. To minimize the effect of unmeasured physical activity, we adjusted the protein intake value for energy intake, which is strongly correlated with physical activity [29]. Nevertheless, this may have been insufficient. Our result showed that the percentage of participants with low protein intake was lower among those in their 70s than those in their 50s. By contrast, older participants showed lower grip strength than younger participants. This implies that multiple factors (including protein intake) are associated with the maintenance of muscle strength. Further studies examining the association between protein intake and muscle strength including physical activity, muscle mass, and chronic kidney disease as covariates are also needed.

In conclusion, protein intake was significantly and positively associated with grip strength for men aged ≥75 years and for women aged ≥65 years, but not in middle-aged and early young-old adults. Further studies to determine the optimal age threshold for a nutritional strategy to prevent age-related muscle weakness are needed.

Acknowledgments

The authors thank the JSTAR team (The Research Institute of Economy, Trade and Industry) for their permission to use the public use dataset. We also thank Satoshi Sasaki and Satomi Kobayashi (The University of Tokyo) for their suggestions in the early version of the manuscript.