Participants
The present study used 2017 data from an annual health examination for adult employees conducted with self-report questionnaires in Pingfang District, Haerbin, China. In this study, 2,323 participants provided their written informed consent for the analysis of their data. The protocol for the study was approved by the Human Investigation Ethics Committee of Huaiyin Institute of Technology.
Responses were excluded if the data for grip strength (n = 98), eating behaviors (n = 30), depressive symptoms (n = 156), and PA (n = 3) were missing. The final study population consisted of 2,036 participants (men = 1,191, women = 845).
Eating Behaviors
Eating behaviors were measured using a self-administered questionnaire that included items on breakfast consumption, snacking after dinner, and eating rate. Frequency of breakfast consumption was determined by the question, “In past month, how many times did you eat breakfast per week?” Snacking after dinner was determined by the question, “In past month, how many days did you snack after dinner?” Participants were asked to report the frequency of their breakfast consumption and snacking after dinner by responding with “never,” “once per week,” to “7 times per week.” Regarding breakfast consumption, the participants were categorized as regular consumers (6–7 times per week), occasional consumers (2–5 times per week), and skippers (less than two times per week). Regarding snacking after dinner, participants were categorized as “always” (6–7 times per week), “sometimes” (2–5 times per week), and “seldom” (less than two times per week). Eating rate was examined using the following question: How quickly do you eat? The participants responded by choosing from the following five categories: very slowly, relatively slowly, at a moderate speed, relatively quickly, and very quickly. We combined the very slow and relatively slow categories to form one slow category, and the fast and relatively fast categories to form one fast category.
Muscle Strength
Muscle strength was determined by handgrip strength and measured using a digital grip dynamometer (WCS-100 Qishu Co., Shanghai, China), with participants in the standing position. The dynamometer was adjusted to fit participants’ individual hand sizes. The participants were encouraged to exert maximal grip effort. Grip strength was measured four times: twice for each hand on an alternating basis, with brief intervals between trials. The average value of the highest power recorded for both hands was included in the analysis [20, 21].
Confounding Variables
Blood pressure (BP) was measured twice using an automatic device, which was applied to the upper-left arm (KENTAROU HBP-9021; OMRON Co., Ltd., Japan) subsequent to rest periods of five minutes in the sitting position. The mean of two measurements was used as the BP value. Hypertension was determined based on a systolic BP of ≥ 140 mmHg, diastolic BP of ≥ 90 mmHg, or use of antihypertensive medication [22]. Diabetes was determined based on a fasting blood glucose concentration of ≥ 126 mg/dL or use of antidiabetic medication. Depressive symptoms were assessed using the Chinese version of the Self-Rating Depression Scale (SDS) [23]. An SDS score of ≥ 45 was adopted as the cutoff point indicating relatively mild or severe depressive symptoms [24]. Body mass index (BMI) was calculated as weight/height2 (kg/m2). Educational level was assessed by determining the highest school grade completed, and classified into two categories: < college or ≥ college. Occupation was classified into two categories: desk work or other. Information regarding age, sex, smoking status, drinking status, occupation, and living status was obtained via a questionnaire survey. Levels of daily PA were estimated using the International Physical Activity Questionnaire (IPAQ). Total daily PA (metabolic equivalents [METs] in h/week) was calculated [25]. PA was categorized into high and low categories.
Statistical analysis
Descriptive data are presented as means (with 95% confidence intervals [95%CIs]) and percentages. Grip strength was used as a dependent variable, whereas categories of eating behaviors (all in tertiles) were used as independent variables. Differences between eating behavior categories were examined using analysis of variance for continuous variables and logistic regression analysis for proportional variables. An analysis of covariance was performed to examine the correlation between eating behaviors and grip strength in the crude and adjusted models. Model 1 was adjusted for age, sex, and BMI, whereas Model 2 was adjusted for the items in Model 1, as well as hypertension, diabetes, and depressive symptoms. Model 3 was adjusted for the items in Models 1 and 2, as well as PA, education level, occupation, living status, smoking, and drinking habits. Bonferroni-corrected p values were used to compare differences between groups. All statistical analyses were performed using SPSS Statistics version 24.0 for Windows (SPSS, Inc., Chicago, IL). Means and 95%CIs were reported throughout. All p values for linear trends were calculated using the categories described above; a p value of < 0.05 was considered statistically significant.