Locomotive Syndrome Associated With Daily Moderate to Vigorous Physical Activity in Community‐dwelling Older Adults: A Cross-sectional Analysis

Yoshihiko Ishihara Tokyo Denki University Hayao Ozaki Tokai Gakuen University Takashi Nakagata National Institutes of Biomedical Innovation, Health and Nutrition (NIBIOHN) Toshinori Yoshihara Juntendo University Toshiharu Natsume Tokai University Tomoharu Kitada Seijoh University Masayoshi Ishibashi Juntendo University Pengyu Deng Juntendo University Yasuyuki Yamada Juntendo University Hiroyuki Kobayashi Tsukuba University Hospital Shuichi Machida (  machidas@juntendo.ac.jp ) Juntendo University Hisashi Naito Juntendo University

for successful ageing [13]. Previous sensor-based PA studies have demonstrated that the daily step count and MVPA were associated with musculoskeletal tissue health, calcaneal health, and frailty status [3,5,12,14,15]. Park et al. [14] suggested that older adults aged 65-84 years who walked at least 7,000-8,000 steps/day and/or spent 15-20 min/day performing MVPA using a uniaxial accelerometer were likely to develop muscle mass above the sarcopenia threshold. Imagawa et al. [9] showed that gait ability (10 m gait time and timed-up and go test) and muscle strength (grip and back muscle strength) were lower among older adults with LS compared to those without LS. Thus, the amount and intensity of daily PA (i.e., step counts and MVPA) is likely to be important for the maintenance of ADL, QOL and prevention of LS. Since walking is the main form of spontaneous PA for most older adults, the average daily step count and intensity of daily PA using an accelerometer offers a promising objective indicator of daily PA. To our knowledge, previous studies have used physical functioning tests for the objective assessment of LS; however, speci c, everyday PA characteristics using a three-axis accelerometer have not been quanti ed. Among community-dwelling older adults, the association between the presence of LS and habitual daily PA characteristics is unclear. Wearable sensors have the bene ts of objectivity, portability, and affordability, making these devices useful for LS assessment in the home and community.
Therefore, this study aimed to evaluate the association between LS and daily PA in community-dwelling older adults using a three-axis accelerometer. The primary outcomes were to show the mean value of the daily step counts, and activity time by exercise intensities in LS and non-LS.

Participants
We performed a cross-sectional analysis between May 2015 and November 2016. In total, 143 healthy communitydwelling individuals (ages: 40-85 years) were included in this study. Of the 143 subjects, 63 were excluded because they had a history of regular sports (tennis and cycling), resistance training (body weight, tube and/or machine), participation in exercise intervention experiments, or were under 60 or over 80 years old. A total of 80 healthy, community-dwelling older adults (40 men and 40 women; age: 60-79 years) were included in the present study. The participants had not history of resistance training, cycling, and other sports activity (i.e., tennis, golf, cycling etc.) for at least 1 year prior to the start of the study. We excluded individuals who were unable to follow our instructions and those with chronic orthopedic conditions or any health or medical condition that limited the ability to undertake light-to-moderate walking. In addition, the participants completed a self-report questionnaire regarding medical history and comorbid conditions. All participants were informed of the study methods, procedures, and risks, and each signed an informed consent document before participating in the study. This study was approved by the Ethics Committee of the Juntendo University (Approval Number: 26-75).

LS tests
To screen the presence of LS, the participants were tasked to perform the two-step test and stand-up test as functional tests [8], and were asked to answer the GLFS-25 questionnaire [16]. The participants stood with the toes of both feet behind a starting line. They were instructed to take two long steps (as long as possible) and to align both feet. The length of the two steps from the starting line to the tips of the toes was measured. The two-step test score was calculated using the following formula: length of the two steps (cm) / height (cm). The participants were also instructed to stand using one or both legs from a speci c height. Seats were placed at the following heights: 40, 30, 20, and 10 cm. The participants stood up from each seat (in descending height order) using both legs, and then using each leg separately. A participant passed the test at the speci ed height if they were able to stand up without leaning back to gain momentum and maintain the posture for 3 sec. The GLFS-25 is a self-administered, comprehensive questionnaire consisting of 25 items, that assessed the patient's pain (4), ADL (16), social function (3), and mental health status (2) during the last month. These items are graded using a ve-point scale from 0 (no impairment) to 4 points (severe impairment). The scores are added to produce a total score (minimum = 0, maximum = 100). A higher score is associated with worse locomotive function. Our previous study [17] indicated that the measured variables from the stand-up test, two-step test, and GLFS-25 have enough validity and reliability, with the intra-class correlation coe cients being 0.87, 0.93, and 0.76 and Cronbach's α being 0.93, 0.95, and 0.88, respectively. These tests were used to assess declines in mobility of each participant.
The JOA has proposed clinical decision limits for these tests in the assessment of LS [7]. According to the results of the LS test, the participants were classi ed as having LS when a participant met one or more of the following criteria: (1) two-step test score < 1.3, (2) di culty in standing from a seat at a height of 40 cm using one leg in the stand-up test (either leg), and (3) GLFS score ≥ 7 [7]. All other participants were placed in the non-LS group. In this study, we focused on the stage 1 of JOA de nition of LS considered as LS group (including stage 1 and 2; n=35 and 6), and the independent values were compared between the LS and non-LS groups.

Habitual daily PA
We measured daily PA using a three-axis accelerometer (UW-301, A&D, Toshima, Tokyo). The participants were instructed to wear the accelerometer around the wrist continuously for 9-14 days except during dressing and bathing. Accelerometer readings recorded for a minimum of 7 continuous days, excluding the distribution and collection days, were considered in the study; days wherein the accelerometer was not worn for more than 2 hours METs) was calculated and de ned as RSB.

Anthropometrics and body composition
Anthropometric measurements included height (cm), weight (kg), and body mass index (BMI, weight [kg] / height [m 2 ]). Body composition measurements included body fat, muscle mass, and waist-to-hip ratio; these were estimated by bioelectrical impedance analysis (BIA) using a body composition analyzer (InBody 730, Biospace Co. Ltd, Seoul, Korea). Appendicular skeletal lean mass (ALM) was calculated as the sum of the muscle mass of the arms and legs [19]. We calculated the skeletal muscle index (SMI) as follows: SMI = ALM/height 2 .

Maximal isometric strengths of leg muscle
The maximal voluntary isometric strength of the knee extensors was determined using a dynamometer (Takei, Tokyo, Japan). Each participant was seated on a chair with the hip joint angle at 90° exion (0° = full hip extension). Prior to the test, several warm-up contractions (2-3 submaximal contractions and 1-2 near-maximal contractions) were performed. They were instructed to perform maximum isometric knee extensions two or three times. The best recorded value was used as the representative, and the weight bearing index (knee strength/ body weight; KE-WBI) was calculated. However, ve participants were unable to perform the test due to high blood pressure which was measured before conducting the test. With regard to knee extension strength, the test-retest (inter-session) reliabilities using ICC, SEM, and minimal difference were 0.945, 3.41 kg, and 9.45 kg, respectively.

Statistical analysis
Data are presented as means ± standard deviations (SD). Differences between the non-LS and LS groups were determined using the unpaired Student's t-test. Relationships among the three LS parameters, daily PA, and anthropometric and body composition measurements were examined using Pearson's product-moment correlation analysis. Finally, this study conducted binomial logistic regression analyses (LRA) to examine the statistical relationships between daily PA and category of LS adjusting for sex difference and age. For this analysis, the continuous variables of daily PA were divided into quartile categories (Table 4). All analyses were performed using SPSS software (ver. 24; SPSS Inc., USA). Statistical signi cance was set at p < 0.05.

Results
Overall, the prevalence of LS was 51.3% (41/80; men, 20; women, 21). The characteristics of the non-LS and LS participants are presented in Table 1. Among the participants, the proportions which obtained two-step test scores of < 1.3, di culty in standing from a seat at a height of 40 cm using one leg in the stand-up test, and GLFS-25 scores ≧ 7 were 12.5%, 36.3%, and 21.3%, respectively. The body fat percentage of women was signi cantly higher among non-LS participants compared to LS participants; however, no difference was observed in men. Men non-LS participants were signi cantly taller and heavier compared to LS participants. KE-WBI of women and men was signi cantly higher among non-LS participants than among LS participants (Table 1). No signi cant differences were observed between sexes in the mean step count and time spent performing MVPA (Table 2). For women, the mean step count and time spent performing MVPA were signi cantly higher among non-LS participants than among LS participants ( Table 2). For men, there was no signi cant differences were observed for daily PA (Table 2).

Discussion
The purpose of the present study was to evaluate the association between LS and daily PA in community-dwelling older adults using a three-axis accelerometer. In our study, 41 (men, 20; women, 21) of the 80 participants (51.3%) met the diagnostic criteria for LS. Older women with LS had lower daily PA (step count and time spent performing MVPA) and relative knee extension strength (weight-bearing index of knee extension) than those without. Logistic regression analyses indicated that LS was associated with step count and the time spent performing RSB (< 1.5 METs) and MVPA (≧ 3 METs) using a three-axis accelerometer. We demonstrated that the associations between LS and daily PA were partly dependent on sex differences.
We found that the daily PA of women participants with LS, particularly the step count and time spent performing MVPA, was signi cantly lower than that of women participants without LS. Aoyagi et al. [20] was showed that lower-extremity function such as gait speed and knee extension torque associated more closely with the daily step count and MVPA for older women, which was suggested that indices of lower-extremity function are signi cantly greater in physically more active older individuals. Thus, it is suggested that older women with lower levels of daily PA may be at increased risk for LS due to the weakening and loss of musculoskeletal tissues [7]. In fact, previous LS studies demonstrated that 10-m gait speed and muscle strength of leg were signi cantly lower in the older adults with LS than in those without [21,22]. Moreover, Nishimura et al. [23] suggested that the frequency of PA between the ages of 25 and 50 years was signi cantly related to LS, and LS participants who reported no regular exercise habits between the ages of 25 and 50 years had the highest prevalence of LS. Therefore, past exercise habits and current daily physical activities may be associated with the prevention or delay of LS onset, although it cannot prove a cause-and-effect relationship.
Regular PA can delay age-related deterioration of functional capacity [24], independent of the presence of other health problems [1,25,26]. Thus, low daily PA reduces mechanical stimulus to the musculoskeletal system, increasing the risk of LS [23]. Furthermore, long periods of inactivity may cause fat accumulation, muscle atrophy, and increased central obesity [27]. In the present study, the body fat percentage of older women was higher among non-LS participants than among LS participants, suggesting that a lower daily PA level and higher obesity-related parameters increase the risk of LS. Since older Japanese women traditionally spend long periods performing lowintensity household tasks [28], purposely an increase in the performing of MVPA is necessary.
It is well known that absolute muscle mass [14,19] and absolute muscle strength [29,30] are generally lower in women than in men. Omori et al. [31] reported that the signi cant decline in the quadriceps muscle strength was observed in individuals in their sixties and seventies. Additionally, the level of muscle strength of women in their fties, sixties, and seventies was signi cantly lower than that of men. To support these points, we found that the KE-WBI of older individuals aged 60-79 years was lesser in women than in men. Moreover, the KE-WBI in women was signi cantly correlated with the time spent performing MVPA. Therefore, a decrease in the activity and mobility may be observed earlier in women than in men. Although the present study cannot ascertain the causality due to the cross-sectional data, we found that maintaining PA above the intensity of walking (≥ 3 METS) is essential in daily life. In this regard, we recommend that older women practice an interval or progressive walking program, which has been shown to improve leg muscle strength [32,33]. These research evidences and exercise protocols will be bene cial in increasing intentional MVPA and preventing LS in older women.
The present study conducted an LRA to examine the relationships between daily PA and category of LS adjusting to sex difference and age. We found that the long time spent performing MVPA using a three-axis accelerometer, with over 47.26 min/day, was associated with a lower likelihood of LS compared to performing MVPA under 21.11 min/day. Moreover, a longer period spent on RSB using a three-axis accelerometer with over 1061.76 min/day was associated with the likelihood of developing LS. Previous studies report a positive correlation between MVPA and various components of physical tness and function; in contrast, RSB was negatively associated with the incidence of coronary artery disease, obesity, osteoporosis, and other signi cant causes of morbidity and mortality in the older adults [2,10,12]. It has been suggested that RSB can cause muscle weakness and impaired gait function, resulting in an increased risk of LS and frailty [9]. Although the causal relationship between daily PA and LS risk is unclear, our results suggest an association between daily activity and LS, and the importance of daily PA, including MVPA (≧ 3 METs).

Limitations
The present study had several limitations. The sample size of 80 was small compared to previous studies of LS, sarcopenia, and frailty with sample sizes ≥100. However, given the availability of accelerometer-based measures of physical activity and screening tests, our results provided useful and reliable data. Second, the cross-sectional study design made it di cult to examine a causal relationship between daily PA and LS. Therefore, future studies should utilize a prospective cohort design, such as longitudinal study design, to examine the causal relationship between these factors. Third, our participants were limited to relatively healthy older individuals aged 60-79 years who had no history of resistance training, cycling, and other sports activity (i.e., tennis and cycling etc.); therefore, our ndings may not be generalizable to the entire Japanese population.

Conclusions
In conclusion, our data showed that daily moderate to vigorous physical activity (MVPA; ≧ 3 METs) and sedentary behavior (RSB; < 1.5 METs) using a three-axis accelerometer were associated with LS in community-dwelling older women, but not in men.

Consent for publication: Not Applicable
Availability of data and materials: All data generated or analyzed during this study are included in this article.