The study had a cross-sectional study design. All procedures were conducted at a laboratory at our university hospital. Participants underwent the entire process in a day. First, all the participants were provided information about the study and screened for the exclusion criteria. After obtaining informed consent, we assessed their eligibility using an original questionnaire to collect data on participants’ characteristics, Mini-Mental State Examination (MMSE), and plantar sensory function, QFMS, and TGF tests. All the procedures were primarily conducted by an experienced physical therapist with support from a clinical physician.

A total of 54 community-dwelling older adults aged ≥60 years (25 men and 29 women) who responded to advertisements in a local library or hospital were enrolled in this study. All the participants lived independently in the local community and could visit our laboratory at the university hospital independently. Participants with current or previous paralysis, dizziness, body pain, limitations in range of motion and flexibility in body movements including the toe and forefoot, or sensory impairment and those with an MMSE score of ≤21 points were excluded from the study. Details on how the exclusion criteria were identified for participants in this study are reported in our previous study [14]. This study was conducted as a part of our previous study [14] and was approved by the Shiga Medical University Research Ethics Review Committee (reference number R2021–125). All participants provided written informed consent before participation in the study.

Participants who reported falling more than once in the past year were classified into the fall history group, and those who reported no fall history in the past year were classified into the non-fall group.

We examined basic characteristics of participants, such as age, sex, body mass index, DF, history of falls, and cognitive function. DF was identified as the foot used by participants to kick a ball [8, 9]. The participants were interviewed for the presence or absence of falls in the past year [15, 16]. History of falls was determined as “when a body part other than the foot touches the ground unconsciously” based on a previous study [17].

Cognitive dysfunction, which could also pose a risk for falls [3, 4], was evaluated using the MMSE. The MMSE is used internationally to assess the severity and progression of cognitive impairment with a maximum of 30 points for 11 items. The lower the score, the more severe the cognitive impairment [18].

TGF and QFMS were evaluated to assess the lower limb muscle strength of the participants. TGF was measured using a toe muscle strength measuring device (T.K.K3364, Takei Scientific Instruments Corp., Osaka, Japan). To measure TGF, participants sat in a chair barefoot. The participants pulled a bar that was attached to the device using their toes, while their ankle and the sole of their foot were fixed to the device with a belt on the floor (Fig. 1). We paid attention to the participants’ motion to ensure that their ankle and the sole of their foot were attached to the device and to prevent other compensatory movements while measuring the TGF. TGF was calculated by dividing the measured value (kg) by the weight (kg) of the participants and was expressed as a percentage. It was measured twice for the DF and non-dominant foot (non-DF), and the average of the two measurements for each foot was included in the analysis. QFMS was measured using a handheld dynamometer (μTas F-1, Anima Corp., Tokyo, Japan). Participants sat in a chair with their sole lifted off the floor; we attached the device to the distal part of the lower leg and fixed it with a belt and then measured the isometric knee extension muscle strength. To calculate QFMS, the lower leg length (the distance from the knee joint space to the lateral malleolus) of the participants was measured. QFMS was divided by the participant’s body weight (kg) after multiplying the measured value (N) by the lower leg length (m). QFMS was also measured twice each for the DF and non-DF and then averaged.

Evaluation of TGF using a toe muscle strength measuring device. To measure TGF, the participants sat in a chair barefoot. The participants then pulled a bar attached to the device using their toes, while their ankle and the sole of their foot were fixed to the device using a belt on the floor. We paid attention to the participants’ motion to ensure that their ankle and the sole of their foot were attached to the device and to prevent other compensatory movements while measuring the TGF. TGF: toe grip force

The two-point discrimination sense (TPDS) of the sole was measured as a somatosensory function. A digital caliper (AD-5765A-150, A&D Corp., Tokyo, Japan) was used to measure TPDS of both soles of the participants (Fig. 2). To measure TPDS, the participants closed their eyes, their soles were brought into contact with two points of the digital caliper, and the minimum distance between the two points of the calipers at which the participants could identify the two points was measured. The two points—the thenar and heel areas of the bilateral foot soles—were measured, and the average value of the two measurements was used as the adopted value for each sole.

Evaluation of TPDS using a digital caliper. A digital caliper was used to measure TPDS of both soles of the participants’ feet. To measure TPDS, the participants closed their eyes, the soles of their feet were then brought into contact with the two points of the digital caliper, and the minimum distance between the two points of the calipers at which the participants could identify the two points distinctly was measured. TPDS: two-point discrimination sense

To calculate the asymmetry of lower limb function, the asymmetries of TGF, QFMS, and TPDS were calculated using Eq. 1, as reported in a previous study by Carabello et al. [19].

$$ \mathrm=\left(\mathrm\ \mathrm\ \mathrm\ \mathrm-\mathrm\ \mathrm\ \mathrm\ \mathrm\right)/\mathrm\ \mathrm\ \mathrm\ \mathrm\times 100\%. $$

(1)

Fisher’s exact test was used to compare sex, DF, and non-DF between the fall and non-fall groups. An unpaired t-test was used to compare continuous data between the fall and non-fall groups. To identify fall risk factors, we performed logistic regression analysis, with the presence or absence of falls as the dependent variable.

For independent variables, we entered variables with p values < 0.2 after the unpaired t-test and Fisher’s exact test according to the study by Coleman et al. [20]. Age and sex were entered into the logistic regression analysis model as independent variables regardless of the test results. For the logistic regression analysis, the presence or absence of a fall was considered the dependent variable, and age, sex, MMSE, heel TPDS (non-DF), and TGF of the DF and non-DF were considered the independent variables (model 1). Furthermore, to investigate whether lower limb function asymmetry was associated with fall risk, the presence or absence of a fall was considered the dependent variable, and age, sex, MMSE, and the asymmetry of TPDS, QFMS, and TGF were considered independent variables in another logistic regression analysis (model 2).

Statistical analysis was conducted using SPSS Statistics Ver. 27.0. (IBM Corp., Tokyo, Japan). Statistical significance was set as p < 0.05.