Introduction

The concept of “locomotive syndrome (LS)” was proposed by the Japanese Orthopaedic Association in 2007 to address Japan’s super-aging population1,2,3,4,5. This concept is halfway between musculoskeletal disease and disability and is defined as a state in which locomotor function in daily activities such as walking, climbing stairs, and rising from a chair is impaired due to musculoskeletal disease1,2,3,4,5. To quantify LS, the Japanese Orthopaedic Association developed LS risk tests consisting of one self-report questionnaire called “the 25-question Geriatric Locomotive Function Scale (GLFS-25)”5,6 and two physical examinations called “the two-step test”5,7 and “the stand-up test.”5,8 Many potential patients with LS have osteoporosis and/or osteoarthritis, along with other musculoskeletal conditions in healthy populations who do not visit hospitals; thus, early detection and prevention of LS through health checkups and other measures are urgently needed9,10.

Walking speed, also a “functional vital sign,” is a valid and reliable measurement method in many populations11. Subjective walking speeds are suitable for evaluating physical functional status, such as decline in skeletal muscle function, and overall health status, such as hypertension, type 2 diabetes, dyslipidemia, cardiovascular disease, cognitive decline, and mortality12,13,14,15,16,17. However, measuring objective walking speed requires sufficient space, manpower, and time; and therefore not be feasible when many people are examined in a limited settings such as routine health checkups18. Considering this evidence and test characteristics, health checkups in Japan generally include a self-report questionnaire regarding subjective walking speed “Is your walking speed faster than that of people of the same age and sex?”19. This question is routinely used in health checkups in Japan because it enables rapid and standardized assessment of functional mobility without the need for additional equipment, space, or examiner training, making it suitable for population-based screening.

Objective usual walking speed has significant associations with all the LS risk tests20,21,22,23,24,25,26,27. However, objective gait speed assessment is not routinely feasible in the clinical health checkup settings. Subjective walking speed has been shown to correlate strongly with objectively measured usual walking speed over a 3-m distance, supporting its validity as a proxy measure of gait and physical performance28. Despite this evidence, previous studies examining the relationship between walking speed and LS have primarily relied on objective gait assessments. To date, no large-scale studies have comprehensively investigated the relationship between subjective walking speed—commonly assessed in routine health checkups—and the LS risk tests. Consequently, the clinical utility of subjective walking speed as a screening indicator for LS in health checkup populations remains unclear. Therefore, we investigated the relationship between LS and subjective walking speed in adults who underwent health checkups.

Methods

Study design

This cross-sectional study assessed adults who underwent medical checkups at Omiya City Clinic, Saitama, Japan, in 2023. The Omiya City Clinic Health Checkup Database (OCC-HCD), a large-scale database launched on April 1 st, 2016, systematically collected data from all participants undergoing health checkups. This database includes demographic information, medical and family history, medication use, lifestyle questionnaire data, blood test results, and imaging studies. In Japan, all full-time employees must undergo annual health checkups, and employers must encourage their employees to participate in these checkups. Additionally, older individuals who have retired from full-time employment continue to receive health checkups individually as part of their personal health management. OCC-HCD has been used in various research studies29.

Of the 40,800 adults who underwent the health checkup, those who did not conduct LS risk tests (the GLFS-25, two-step test, and/or stand-up test) (n = 5,105) and those lacking assessable data (n = 760) were excluded. Accordingly, 34,935 adults were considered eligible for inclusion in this study. The study was approved by the institutional review board of Saitama Medical University (Registration number: 957).

Subjective walking speed

Subjective walking speed was investigated using the question “Is your walking speed faster than the speed of those of your age and sex?” in a health examination questionnaire. Those who answered “yes” were classified as having “fast walking speed” and those who answered “no” were classified as having “slow walking speed”19.

LS assessment

The GLFS-25 includes 25 items graded on a 5-point scale (0–4 points; possible scores range 0–100)5,6. The domains covered by this scale include body pain (items 1–4), movement-related difficulty (items 5–7), usual care (items 8–11 and 14), social activities (items 12, 13, and 15–23), and cognition (items 24 and 25)6. Total scores of 0–6, 7–15, 16–23, and 24–100 reflect non-LS, LS-1, LS-2, and LS-3, respectively5,6. The two-step test measures the maximum length of the two-step stride of participants5,7. Individuals carry out this test from the standing posture without losing balance. The two-step test score is a normalized value calculated by dividing the maximum two-step stride length (cm) by the individual’s height (cm)5,7. Two-step test scores of ≥ 1.3, < 1.3, < 1.1, and < 0.9 points are equivalent to non-LS, LS-1, LS-2, and LS-3, respectively5,7. The stand-up test evaluates lower limb strength according to stand—in a single-leg or double-leg stance—from four different heights (10, 20, 30, and 40 cm)5,8. The test is scored as 0–8, and the scores are defined as follows: 0 (unable to stand); 1–4 (able to stand—using both legs—from 40, 30, 20, and 10 cm, respectively); and 5–8 (able to stand—using one leg—from 40, 30, 20, and 10 cm, respectively)5,8. Stand-up test scores of 5–8, 3–4, 2, and 0–1 points are equivalent to non-LS, LS-1, LS-2, and LS-3, respectively5,8. In the LS total assessment, the risk level was deemed LS-1, LS-2, or LS-3 if the adults met even one of the relevant criteria5.

Covariates

We investigated covariates that have been reported in previous studies examining factors associated with LS30,31,32,33,34,35,36,37,38: sex32,33,34,35,36,37,38, age (< 65 or ≥ 65 years)30,31,32,33,34,35,36,37,38, body mass index (< 18.5, 18.5–24.9, or ≥ 25.0 kg/m2)32,33,34,35,36,37,38, smoking status (current smoker)30,31,33,35, alcohol consumption (everyday, sometimes, or no/rare)33, exercise habits31,33,35,36,37, and comorbidities (cerebrovascular, cardiovascular, and renal diseases)31,33.

Body mass index was calculated from measured height and weight using the following formula: weight (kg) divided by height squared (m2). The other covariates were investigated using self-reported questionnaires. Current smoking status was defined based on the response to the question, “Do you currently smoke cigarettes habitually?” Participants who answered “yes” and met both of the following criteria were classified as current smokers: (1) having smoked during the past month, and (2) having smoked for six months or more in their lifetime or having smoked a total of 100 or more cigarettes. Alcohol consumption was assessed using the question, “How often do you drink alcohol (sake, shochu, beer, Western liquor, etc.)?” Participants who answered “every day” were classified as everyday drinkers; those who answered “5–6 days per week,” “3–4 days per week,” “1–2 days per week,” “1–3 days per month,” or “less than 1 day per month” were classified as sometimes drinkers; and those who answered “I have quit” or “I do not drink (cannot drink)” were classified as no/rare drinkers. Exercise habits were defined based on the response to the question, “Do you engage in exercise that causes light sweating for at least 30 minutes per session, at least two days per week, for at least one year?” Participants who answered “yes” were classified as having exercise habits. Participants who answered “yes” to the question, “Have you ever been told by a doctor that you have had a stroke (cerebral hemorrhage, cerebral infarction, etc.) or received treatment for it?” were classified as having cerebrovascular disease. Participants who answered “yes” to the question, “Have you ever been told by a doctor that you have heart disease (angina pectoris, myocardial infarction, etc.) or have you ever received treatment for it?” were classified as having cardiovascular disease. Participants who answered “yes” to the question, “Has a doctor told you that you have chronic kidney disease or renal failure, or are you receiving treatment (such as dialysis)?” were classified as having renal disease.

Statistical analyses

Variables are expressed as means and standard deviations for quantitative data and percentages for qualitative data. Student’s t-test was used to compare quantitative data, and Fisher’s exact test was used to compare qualitative data between participants with fast and slow walking speeds. We conducted logistic regression analysis to estimate the odds ratios (ORs) with a 95% confidence interval (CI) to investigate the relationship between LS risk tests and subjective slow walking speed. Categorization of LS (LS-1 or more [LS-1, LS-2, or LS-3] vs. non-LS; LS-2 or more [LS-2 or LS-3] vs. LS-1 or less [non-LS or LS-1]) was used as a dependent variable, and subjective walking speed was used as an independent variable. The regression models were constructed in multiple steps. First, the univariate models were created. Second, we created multivariate models adjusted for sex, age, and body mass index. Third, we created multivariable models adjusted for factors with statistical significance (p < 0.050) or marginal significance (0.050 ≤ p < 0.100) in comparison analyses, in addition to sex, age, and body mass index. Fourth, we created multivariable models adjusted for all the factors investigated in this study. These factors accounted for potential confounding, recognizing that apparent associations may exist even when no significant differences exist in comparison analyses30,31,32,33,34,35,36,37,38. As walking speed decreases and LS severity progresses with age32,39, subgroup analyses by age groups (< 65 years and ≥ 65 years6 were performed. The threshold for significance was set at a p-value < 0.050. All analyses were performed using JMP® pro 17 (SAS Institute; Cary, North Carolina, USA).

Results

Demographic data

The demographic data for the study participants are shown in Table 1. This study included 34,935 participants (males, n = 20,994 [60.0%]; females, n = 13,991 [40.0%]; mean age, 50.1 ± 9.7 years [range, 19–90 years]). The mean GLFS-25 total score was 2.7 ± 4.2 points, and the mean two-step test score was 1.4 ± 0.1. Based on the GLFS-25, 31,141 (89.1%), 3,135 (9.0%), 454 (1.3%), and 205 (0.6%) participants were categorized as non-LS, LS-1, LS-2, and LS-3, respectively. Based on the two-step test, 34,108 (97.6%), 761 (2.2%), 59 (0.2%), and 7 (0%) participants were categorized as non-LS, LS-1, LS-2, and LS-3, respectively. Based on the stand-up test, 32,243 (92.3%), 2,604 (7.5%), 84 (0.2%), and 4 (0%) participants were categorized as non-LS, LS-1, LS-2, and LS-3, respectively. Based on the total assessment, 28,970 (82.9%), 5,194 (14.9%), 557 (1.6%), and 214 (0.6%) participants were categorized as non-LS, LS-1, LS-2, and LS-3, respectively. These participants were classified as fast walking speed groups (n = 18,222 [52.2%]) and slow walking speed groups (n = 16,713 [47.8%]).

Table 1 Characteristics of the study participants (n = 34,935).
Full size table

Association between LS and subjective slow walking

In all the LS risk tests, the percentage of subjective slow walking speed increased significantly as the LS severity progressed (p < 0.001; Fig. 1).

Fig. 1
Fig. 1
Full size image

Percentage of participants with subjective slow walking speed according to the LS severity categorized by the total assessment, GLFS-25, two-step test, and stand-up test. LS, locomotive syndrome; GLFS-25, the 25-question geriatric locomotive function scal.

Table 1 presents the results of the comparison analyses between participants with fast and slow subjective walking speeds. Compared with the fast walking speed group, participants with slow walking speed were more likely to be female, younger, and have a higher body mass index. They were also less likely to report alcohol consumption and regular exercise habits, and showed marginally higher prevalences of cerebrovascular and cardiovascular diseases.

Accordingly, four regression models were developed: (1) unadjusted; (2) adjusted for sex, age, and body mass index; (3) adjusted for sex, age, body mass index, alcohol consumption, exercise habits, cerebrovascular disease, and cardiovascular disease; and (4) adjusted for sex, age, body mass index, smoking status, alcohol consumption, exercise habits, cerebrovascular disease, cardiovascular disease, and renal disease. In all the regression models, the categorization of LS was associated with walking speed (Table 2). In model 4, after adjusting for sex, age, body mass index, smoking status, alcohol consumption, exercise habits, cerebrovascular disease, cardiovascular disease, and renal disease, subjective slow walking was associated with LS categorized by all LS risk tests and total assessment. The OR of total assessments in LS-2 or more for subjective slow walking speed (OR = 2.49; 95% CI 2.12–2.91) were greater than that of total assessments in LS-1 or more (OR = 1.51; 95% CI 1.43–1.60). Similarly, the ORs of all LS risk tests in LS-2 or more for subjective slow walking speed were greater than those of all LS risk tests in LS-1 or more. Furthermore, age (< 65 and ≥ 65 years) did not modify the associations between LS and walking speed, except for the associations between LS-2 or more categorized by the stand-up test and the walking speed (Table 3).

Table 2 Association between categorization of LS and subjective slow walking in the overall participants (n = 34,935).
Full size table
Table 3 Association between categorization of LS and subjective slow walking in participants aged < 65 years (n = 32,252) and ≥ 65 years (n = 2,683).
Full size table

Discussion

This cross-sectional study involved 34,935 adults to investigate the relationship between LS and subjective walking speed. Our study findings revealed a relationship between subjective walking speed, all LS risk tests, and total assessment. Our results unveiled that the ORs of all LS risk tests and total assessment in LS-2 or more for subjective slow walking speed were greater than those of all LS risk tests and total assessment in LS-1 or more. Finally, our findings showed that age did not modify the association between LS and subjective slow walking speed.

Similar to our findings, a cross-sectional preliminary study of 172 participants that collected health checkup data reported significant relationships between LS risk tests (the GLFS-25, two-step test, and stand-up test) and subjective walking speed40. To the best of our knowledge, no large-scale studies investigated the relationship between LS and subjective walking speed. Subjective usual walking speed was reported to be strongly associated with objective usual walking speed among 730 males and 999 females aged 61–73 years28.

Previous studies reported a significant relationship between LS categorized by the GLFS-25 and objective usual walking speed20,21,22,23,24. In the ROAD study20, LS-1 or more and LS-2 or more categorized by the GLFS-25 were reported to be associated with objective usual walking speed < 0.8 m/s (OR, 2.65 [95%CI, 1.82–3.86] and 3.49 [95%CI, 2.15–5.65], respectively). The Nagahama study21 revealed that the more severe the LS categorized by the GLFS-25, the slower the objective usual walking speed; the mean objective usual walking speed values of non-LS, LS-1, LS-2, and LS-3 were 1.33 m/s, 1.26 m/s, 1.24 m/s, and 1.13 m/s, respectively. In the Miyagawa study22, LS-2 or more categorized by the GLFS-25 showed slower objective usual walking speed (mean, 0.78 vs. 1.04 m/s). Nakamura et al.23. assessed 126 healthy females and reported that those with LS-2 or more had a slower usual 6-m walk time (mean, 5.18 vs. 4.45 s) than those with LS-1 or less. Saito et al.24 assessed 125 local residents and investigated the relationship between LS categorized by the GLFS-25 and spatiotemporal gait parameters and kinematics during the 10-m walk test at usual speed using wearable gait sensors. Compared with non-LS, LS-2 or more had shorter objective usual walking speed (mean, 1.1 vs. 1.3 m/s), slower cadence (mean, 111.4 vs. 120.6 steps/min), shorter step length (mean, 0.52 vs. 0.78 m), smaller hip extension (mean, 4.2° vs. 9.5°), smaller hip flexion (mean, 28.5° vs. 34.2°), smaller hip abduction (mean, 5.9° vs. 7.9°), and smaller knee flexion (mean, 50.6° vs. 65.2°).

According to previous studies20,25, LS categorized by the two-step test was significantly associated with objective usual walking speed. In the ROAD study20, LS-1 or more and LS-2 or more categorized by the two-step test were reported to be associated with objective usual walking speed < 0.8 m/s (OR, 4.24 [95%CI, 2.18–8.22] and 4.19 [95%CI, 2.75–6.39], respectively). Sato et al.25 assessed 36 community dwellers and assessed the relationship between LS categorized by the two-step test and gait kinematics at the usual speed using eight cameras and four force plates. Compared with non-LS, LS-1 or more had shorter objective usual walking speed (mean, 1.23 vs. 1.40 m/s), slower cadence (mean, 124.99 vs. 126.03 steps/min), shorter step length (mean, 0.58 vs. 0.66 m), smaller muscle strength during hip extension (mean, 0.64 vs. 1.05 Nm/kg), hip flexion (mean, 1.04 vs. 1.68 Nm/kg), hip abduction (mean, 0.84 vs. 1.00 Nm/kg), and smaller knee extension (mean, 1.59 vs. 2.25 Nm/kg).

Two cross-sectional studies20,26 observed a significant association between LS categorized by the stand-up test and objective usual walking speed. In the ROAD study20, LS-1 or more and LS-2 or more categorized by the stand-up test were reported to be associated with objective usual walking speed < 0.8 m/s (OR, 2.01 [95%CI, 1.35–3.16] and 3.40 [95%CI, 1.99–5.82], respectively). Nishimura et al.26. revealed that older community-dwelling females with LS-2 or more categorized by the stand-up test tended to have shorter objective usual walking speed (mean, 1.49 m/s vs. 1.54 m/s) than those with LS-1 or less.

LS categorized by total assessment was reported to have a significant association with objective usual and maximum walking speed20,27. In the ROAD study20, LS-1 or more and LS-2 or more categorized by the total assessment were reported to be associated with objective usual walking speed < 0.8 m/s (OR, 32.21 [95%CI, 9.64–107.7] and 61.93 [95%CI, 24.92–153.87], respectively).

Subjective slow walking speed was associated with LS-1 or more and LS-2 or more categorized by the GLFS-25 regardless of age group. The GLFS-25 widely evaluates subjective health in motor functions, including body pain, movement-related difficulty, usual care, social activities, and cognition6. This test significantly correlates with walking speed and the other physical assessments, such as the two-step test, stand-up test, grip strength, one-leg standing time, timed up-and-go, and five times sit-to-stand, 4-m walking test31.

This study reported that subjective slow walking speed was associated with LS-1 or more and LS-2 or more, which were categorized by the two-step test regardless of the age groups. The two-step test represents actual transverse movement ability (walking ability), which correlated well with maximum walking speed7. Furthermore, the degree of change in the two-step test score due to strengthening the hip flexor muscles and the degree of change in usual walking speed are reported to be correlated25. This test widely evaluates the function of bilateral lower extremities, such as lower limb strength, standing balance, flexibility, and walking ability41.

This study revealed that subjective slow walking speed was associated with LS-1 or more, and it was categorized by the stand-up test regardless of the age group. The stand-up test represents actual vertical movement ability (standing ability), evaluating the knee extension strength of the quadriceps femoris muscle42, range of motion and flexibility at the hip and knee joint43, and the balance function44, which are important factors for walking. In contrast, the lack of association between subjective slow walking speed and LS-2 or more categorized by stand-up test in participants aged ≥ 65 years could be partially explained by lumbar spinal canal stenosis, whose prevalence in those aged ≥ 65 years is 9.4–19.4%45. Lumbar spinal canal stenosis is related to slow walking speed46, but not to the stand-up test42. Although leg pain or low back pain may decrease the ability to stand up, the muscle strength of the quadriceps femoris is not directly affected by stenosis at the level of L4/5, which is the most frequently responsible level in lumbar spinal stenosis43.

Therefore, adults with subjective slow walking speed are recommended to take LS risk tests irrespective of their age to accurately identify and assess the decline in mobility. To our knowledge, this is the first study to investigate the relationship between LS and subjective walking speed, and our study findings may contribute to the early detection and prevention of LS.

Clinical implications

This study provides important clinical implications. Subjective walking speed—widely used in routine Japanese health checkups—can serve as a simple, low-cost, and time-efficient initial screening tool to identify adults aged 19–90 years at risk of LS. A stepwise screening approach—referring individuals with subjective slow walking speed for LS risk tests—may improve the efficiency of routine screening programs, facilitate earlier intervention, and contribute to the prevention of future disability and the extension of healthy life expectancy.

Strengths and limitations

This study has two main strengths. First, the large sample size facilitated good statistical power to precisely estimate the association between LS and subjective walking speed. Second, there were very few missing values; the response rate for the GLFS-25 was 87.5%, and all of the participants completed the two-step and stand-up tests. This resulted in fewer participants being excluded and was less likely to result in selection bias.

However, this study has certain limitations. First, to categorize LS, only stand-up tests from 30 cm on both legs, 20 cm on both legs, and 40 cm on one leg were investigated. Thus, the stand-up test value was not available. Second, there may be unmeasured confounding factors such as education level47. Third, compared with the results of a nationwide study of LS48, this study population was relatively young, and many of the participants had non-LS (Table 1; Supplementary Table 1). Further investigations are warranted considering these limitations to generalize our findings.

In conclusion, subjective walking speed was related to all LS risk tests and total assessment. Adults with a subjective slow walking speed are more likely to have severe LS. Age did not modify the associations between LS and subjective slow walking speed. Therefore, adults aged 19–90 years with subjective slow walking speed are recommended to take LS risk tests to accurately identify and assess the decline in mobility.