Subjects and measurements
We performed a cross-sectional analysis between May 2015 and November 2016. In total, 223 untrained community-dwelling middle-aged and elderly individuals (104 men and 119 women; ages: 40–85 years; mean height: 159.3 ± 8.2 cm; mean weight: 58.4 ± 9.5 kg) were included in this study. The participants had not performed any resistance training 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 about the contents of the study, and they provided a signed informed consent. This study was approved by the Ethics Committee of the Juntendo University (Approval Number: 27–10).
Body weight, skeletal muscle mass, fat mass, and percentage of body fat (% Fat) were measured via bioelectrical impedance analysis using a body composition analyzer (InBody730; Biospace). Venous blood samples of about 13 mL were obtained between 10:00 am and 5:00 pm after at least a 3-h fast, and the following parameters were analyzed: white blood cell (WBC) count, hematocrit (Hct) value, and levels of hemoglobin (Hb), total protein (TP), triglyceride (TG), low density lipoprotein (LDL)-cholesterol (C), high density lipoprotein (HDL)-C, HDL/LDL-C ratio, glycated hemoglobin (HbA1c) (National Glycohemoglobin Standardization Program [NGSP]), aspartate amino transferase (AST/GOT), alanine amino transferase (ALT/GPT), alkaline phosphatase (ALP), leucine aminopeptidase (LAP), γ-glutamyl transpeptidase (γ-GTP), growth hormone (GH), insulin-like growth factor-1 (IGF-1), testosterone, cortisol, dehydroepiandrosterone-sulfate (DHEA-S), and albumin. Of 223 participants, 101 (41 men, 60 women) were analyzed for albumin levels because we adopted it as an indicator after October 2015. All assays were performed by SRL Inc. (Tokyo, Japan).
After blood collection, all participants were measured for LS risk, walking speed, and maximal isometric muscle strength. LS risk tests (the stand-up test, two-step test, and 25-question GLFS) were performed as described previously [
14]. Briefly, in the stand-up test, the participants were asked to stand from a sitting position on two legs or one leg, and from four different seat heights (40, 30, 20, and 10 cm). Participants were instructed to stand up without leaning back to gain momentum and to maintain the standing posture for at least 3 s. A score from “0” to “8” was allocated based on the difficulty, as described by Ogata et al. [
5]. For the two-step test, participants stood with the toes of both feet behind the starting line and performed two maximal stride lengths, one with each foot, and the distance from the starting line was measured. The score was calculated as the length of the two steps (cm)/height (cm) [
5]. The 25-question GLFS is a self-reported questionnaire, which is a comprehensive measure consisting of 25 items, including pain, activities of daily living, and mental health during the last month [
3]. These 25 items were graded on a five-point scale, from 0 (no impairment) to 4 (severe impairment) points, with the scores summed to provide the total GLFS score; a higher GLFS score was associated with a higher risk of developing LS. Our previous study indicated that the measured variables from the stand-up test, two-step test, and GLFS have enough validity and reliability, with the intra-class correlation coefficients being 0.87, 0.93, and 0.76 and Cronbach’s α being 0.93, 0.95, and 0.88, respectively [
14].
Walking speed was evaluated by timing each subject as they walked across a 10-m corridor on a hard-surfaced floor. They were asked to walk down the corridor as fast as possible without running. The maximum isometric strength of the knee extension was also measured (Takei, Tokyo, Japan) as described previously [
14]. Each subject was seated on a chair with the hip joint angle at 90° flexion (0° = full hip extension), and they were instructed to perform maximum isometric knee extensions two or three times. The best recorded value was used as the representative one, and the weight bearing index (knee strength/body weight) was calculated.
Statistical analysis
The data are presented as the mean ± standard deviation (SD). To compare the variables between the LS and non-LS groups, we used Student’s unpaired
t-test. For variables that did not show normal distribution (for HbA1c levels and DHEA-S levels in males), the Mann-Whitney U test was used. The medians [interquartile ranges (IQRs)] are also presented for variables that did not show a normal distribution (i.e., for HbA1c and DHEA-S levels in males). The Benjamini-Hochberg procedure was performed to control the false discovery rate at 0.05 (used 14 tests for blood characteristics and biochemical examinations and 6 tests for hormone levels in males and females, respectively) [
16]. To determine whether blood parameters were associated with LS risk, a binary logistic regression analysis (adjusted for age) was performed for HbA1c, DHEA-S levels in males, and albumin. We used the odds ratio (OR) and 95% confidence intervals (OR
95%CI) to estimate the relative risk. A receiver operating characteristics (ROC) curve was used to select an appropriate cutoff and to determine the area under curve (AUC) and maximum sensitivity and specificity. Correlations between HbA1c and albumin levels and characteristics of participants [% Fat, body mass index (BMI), skeletal muscle mass, 10-m walking speed, and the weight bearing indices of knee extension] were analyzed using simple linear regression and Pearson’s correlation analysis. All of the statistical analyses were performed using the EZR package (Saitama Medical Center, Jichi Medical University) [
17], which is a modified version of the R programming environment (The R Foundation for Statistical Computing, Vienna, Austria). The statistical significance level was set at
p < 0.05.