Simple regression analysis revealed that FM values of the legs and total body as well as the waist circumference were significantly correlated with scores of each of the three tests, whereas LSTM values for legs and total body were not. In addition, CSAs of the three muscle groups were also not significantly correlated with scores of the three performance tests except for the relationship between QF CSA and Two-step length. These results are consistent with previous findings of cross-sectional studies [12, 15, 17, 21] and suggest that the magnitude of FM is more influential than LSTM with regard to lower extremity performance in older women. However, results of the multiple regression analysis (Table 3) deny the aforementioned consideration based on the absolute values of the measured parameters. Notably, for TUG time and Two-step length, QF CSA/BM2/3 is more influential than CSA/BM2/3 values of the other thigh muscles and the body composition variables of the total body and three body segments.
In terms of CSA/BM2/3, only QF was significantly correlated with scores of the three tests. This may be attributable to the fact that the performance tests adopted here are body mass-based movements in which practitioner body mass acts as a load on lower limb muscles. [35] conducted a cross-sectional study involving men and women aged 19 to 90 years and found that electromyogram activities of the quadriceps femoris during body mass-based squat movements increased nonlinearly with decreasing knee extension torque relative to body mass. A certain degree of knee extension torque relative to body mass is required to successfully perform activities of daily living [35–38]. Taking the above into account together with what is known about the relationship between QF size and knee extensor torque [11], it is assumed that QF CSA/BM2/3 corresponds to knee extensor torque relative to body mass, which may explain the significant correlation with scores of the three performance tests.
For 5-m walking time, leg LSTM/BM was selected as a significant contributor; simple correlation analysis also found QF CSA/BM2/3 to be significantly correlated, possibly because the 5-m walking test was performed under conditions of the participant’s preferred walking speed. Knee extensor strength has been shown to be associated with either preferred [39, 40] or maximal [41] walking speeds. However, [42], who examined how leg strength was associated with preferred and maximal walking performance within the same individual, reported that knee extensor torque was significantly correlated with maximal gait speed, but not with preferred walking speed. In their study, leg muscle mass estimated using a bioelectrical impedance method was found to be significantly correlated to stride length for both preferred and maximal walking speeds. In addition, [43] observed that in older adults, stride length during 6-min walking at a preferred pace decreased markedly as body fat percentage increased. Combined with the findings obtained here, it is said that LSTM relative to body mass for legs (rather than the specific muscle group such as QF) would more readily influence gait performance at a habitual or usual pace.
For Two-step length, waist circumference was selected as a negative factor, and QF CSA/BM2/3 as a positive factor. The Two-step test was developed to assess walking ability, including muscle strength, balance, and flexibility of the lower limb muscles [28]. Among these factors relating to two-step length, balance has been shown to be associated with the magnitude of obesity [34, 44–46]. [34] observed that static postural instability, assessed by the path length of the center of gravity, was positively associated with abdominal visceral fat area and negatively associated with QF CSA-corrected body mass. In addition, [44] reported that a high BMI demands more displacements to maintain postural balance. Meanwhile, reducing weight is known to increase postural stability in obese men [46]. The present study found that Two-step length was negatively correlated with body mass, BMI, waist circumference, FM, and FM/BM. This supports the hypothesis that the magnitude of obesity may negatively affect the Two-step length by reducing postural balance during a given task. Notably, only waist circumference was identified by multiple regression analysis as a negative contributor, suggesting that the magnitude of fat mass per se would not negatively influence Two-step length.
From a biomechanical point of view, an increased waist circumference might negatively influence Two-step length. It is known that postural stability in obese women is strongly dependent on fat mass distribution within the body [47]. [47] reported that women with abdominal obesity, i.e., the android type, in which fat localizes mostly in the upper part of the body, especially the abdomen and chest, show less stability in standing posture than women with gynoid fat distribution, for whom the fat localizes mainly in the thighs and buttocks. Human locomotion is affected by human body structure, of which two-thirds of the total body mass is in the upper body (head-arm-trunk) segment [48]. Theoretically, a greater waist circumference should become a source yielding a large amount of potential energy when in an upright position. Based on this idea, a higher waist circumference would negatively influence Two-step length by increasing practitioner efforts to control upper body movement while conducting the test.
For the TUG score, multiple regression analysis selected age as a negative factor and QF CSA/BM2/3 as a positive factor. Age as well as gender contribute directly to explain variation in TUG performance [49]. It is well documented that the TUG score increases with age [8, 49–52]. The present study found that Two-step length was also negatively correlated with age. However, it remains unclear why, of the three performance scores obtained here, only the TUG score had age as a significant factor explaining its variation. One possibility is the complicated structure of the TUG test, which might explain why age negatively influenced the performance of this test. Namely, the TUG test requires practitioners to move from sitting to standing, walk 3 m, turn, and walk back to the chair and sit down again as quickly as they safely can while being timed [9]. This relatively complex structure of the TUG test may be more difficult for older individuals relative to tests consisting of a single action; accordingly, age might have been selected as a negative contributor to the TUG score.
The present study has some limitations. First, participants were between the ages of 60 to 77 years, and all were functionally independent in daily life. All participants successfully completed the three performance tests. Accordingly, it is possible that our results may have been influenced by some selection bias and generalizing our results to other populations of older adults may be problematic. Second, the present study did not quantify the size of individual muscles located in the trunk. One systematic review demonstrated the importance of trunk muscle strength for balance, functional performance, and fall prevention in older adults [53]. In addition, muscle cross-sectional ([54] and muscle thickness values ([27] of the rectus abdominis are related to functional ability in older adults. Trunk exercise training improves the size of trunk muscles, trunk muscle strength and physical function [53, 55]. Thus, we cannot rule out the possibility that, if the size of individual trunk muscles is added to that of the thigh muscles in multiple regression analyses, the results pertaining to factors contributing to each of the performance scores might have differed from those obtained here. Further research, including that to determine the size of individual trunk muscles as well as thigh muscles is needed to clarify this issue.