Study design and participants.
This cross-sectional study of National Health and Nutrition Examination Survey (NHANES) was to explore the overall picture of nutrition, health, and risk factors among residents in various states across the USA (Centers for Disease Control and Prevention-http://www.cdc.gov/nchs/nhanes.htm). This survey is based on national-scale adopting multistage and clustered sample methods with the figure of being 5000 participants each year. Voluntary participants were asked to finish physical examination and this study was approved by Health Statistics Research Ethics Review Board from national center. All the participants have signed the written consent.
In this present study, data including handgrip strength, depression, sleep disorder, cancer type, baseline characteristics, and other health-related variables were extracted from NHANES ranging from 2011 to 2012 and 2013 to 2014, aggregating for final analysis.
Cancer diagnosis
Cancer patients was confirmed by asking about the question (Have you even been told by the doctor that you had been diagnosed with any type of cancer). We selected the patients whose answer is “YES.”
Handgrip strength
The detail of muscle strength measurement was depicted in NHANES Procedure Manual. Briefly, those participants who are able to finish the test obeyed the standard procedure. Investigators explained the detailed information and asked participants to squeeze the dynamometer as hard as possible for three times, recording the maximum value as the participant’s final handgrip strength. In our present study, the definition of low handgrip strength was <27kg for males and <16kg for females16.
Depressive symptoms
According to the Questionnaire Instruments, Patient Health Questionnaire (PHQ-9) was used to assess the depressive symptoms of participants, consisting of 9-items with each item ranging from 0 to 3 points17. The total points of PHQ-9 ranged from 0 to 27. We categorized participants into depression and non-depression, with the cut-off being 10 points based on the previous study, which was reported good sensitivity (88%) and specificity (88%) for identifying major depression18.
Covariates definition
We extracted demographic characteristics including age, gender, education, race, marital status, and smoking. Of which, the race was defined as Non-Hispanic Black, Non-Hispanic White, and others; marital status was classified as married, widowed or divorced, and others. Regarding of education, four categories including less than 12 grades, high school, some college, and college graduated above were confirmed. Other covariates such as BMI, cancer diagnosis, sleep disorder, history of stroke, and history of Congestive Heart Failure (CHF) were also extracted. In addition, Leisure-time physical activity was assessed by Global Physical Activity Questionnaire19. The detailed calculation method was reported the previous study. In brief, we combined the vigorous and moderate intensities of physical activity each week, and we defined it as inactive and physical activity, respectively, with the cut-off of zero min/week.
Statistical analysis
Continuous variables including age, BMI, Depression score were present as means and SD, and categorical variables including diagnosis, gender, race, education, marital status, sleep disorder, and other health-related variables were displayed as frequency (%). Comparisons for low handgrip strength versus normal handgrip strength and depression versus non-depression were conducted by Student’s-test and chi-squared test or Fisher’s Exact Test and Mann–Whitney tests, for appropriate. In addition, the generalized additive model (GAM) analysis was used to detect whether there is a non-linear relationship between handgrip strength and depression20. Before performing multivariable logistic regression analysis, we used Least Absolute Shrinkage and Selection Operator (LASSO) regression to select variables for the final regression model. The results indicated that these variables (age, race, education, marital status, sleep disorder, history of stroke, polypharmacy, BMI, handgrip strength) were selected. Finally, Multivariable logistic regression analysis was adopted to identify the independent association between handgrip strength and depression after controlling for potential confounding factors including age, race, education, marital status, sleep disorder, history of stroke, polypharmacy, BMI. We also categorized handgrip strength into low handgrip and normal handgrip according to the sarcopenia of the revised European consensus on definition and diagnosis, with the cut-off value being <16kg for females and <27kg for males. The association between low handgrip strength and depression was also detected by multivariable logistic regression analysis with adjustment of the same variables. Unadjusted subgroup analysis between low handgrip strength and depression was performed in terms of different variables( inactive versus active, >=65 versus <65, sleep disorder, congestive heart failure, polypharmacy, history of stroke, marital status, race, and education.) All the statistics were conducted by software packages R and Empowerstats, with the significant P-value being <0.05.