Study design and population
This cross-sectional study is based on secondary data from the 2011–12 and 2013–14 NHANES. NHANES, conducted by the National Center for Health Statistics, interviewed approximately 5000 individuals of all ages in their homes. The details of the design, target population, and content of the public use data files are described elsewhere . Participants were selected using a stratified, multistage probability design. Hispanics, non-Hispanic Blacks, non-Hispanic/non-Black Asians, low-income non-Hispanic Whites, and persons 80 years or older were oversampled to provide stable estimates of these groups.
The present study sample consisted of 1247 NHANES 2011–12 and 2013–14 participants aged 50 years and older who had serum Zn level measurements and self-reported physical functioning information. We excluded participants with missing data (n = 111) on covariates included in the multivariable model. The final sample comprised 1136 participants. This study was approved by the Institutional Review Board.
Basic physical functioning difficulty
In this study, basic physical functioning status is defined based on the four questions under the module “Physical Functioning” that are available in both NHANES 2011–12 and 2013–14 dataset in the elements of the Katz index of independence of basic ADL : “Without using any special equipment, how much difficulty do you have walking from one room to another on the same level?”, “Without using any special equipment, how much difficulty do you have getting in or out of bed?”, “Without using any special equipment, how much difficulty do you have eating, such as holding a fork, cutting food or drinking from a glass?”, “Without using any special equipment, how much difficulty do you have dressing yourself, including tying shoes, working zippers, and doing buttons?” The response options to the question range from 1 to 5 to reflect no difficulty, some difficulty, much difficulty, unable to do, or do not do this activity, respectively. We coded the response as 0 if the participant reported performing an activity with no difficulty and responses were coded as 1 if the participant reported some difficulty, much difficulty, or “unable to do”. Following existing literature, participants who reported “do not do this activity” were excluded from the analysis . The distribution of self-reported difficulty in four basic physical functioning was shown in supplementary Additional file 1: Table 1A and 1B. Because of the problems that were encountered with small sample sizes in some cells, an issue of convergence, and unstable odds ratio estimates, our study adopted the coding for the basic physical functioning status as having no difficulty vs. having difficulty (0 vs. 1). While there are well-known limitations of self-reported measures of physical functioning, they are reliable , and both men and women generally report their disability accurately [29, 30].
Serum Zn concentrations
Serum Zn levels were measured by inductively coupled plasma dynamic reaction cell mass spectrometry (ICP-DRC-MS) method as described by NHANES 2011–14 protocols . Participants were categorized into three groups by serum Zn tertile: low (range: 0.409–0.748 μg/mL, n = 404), middle (range: 0.749–0.870 μg/mL, n = 380), and high (range: 0.872–1.698 μg/mL, n = 352) group. Serum Zn levels in SI unit (μmol/L) can be converted by multiplying the factor 15.3. Serum albumin and blood hemoglobin concentrations which are significantly associated with serum Zn levels  were obtained from laboratory files of NHANES 2011–14 data .
Age, sex, race/ethnicity, and health insurance information was assessed using the questionnaire. Body mass index (BMI) calculated as weight in kilograms divided by height in meters squared was categorized into four groups: obese (BMI ≥ 30); overweight (BMI 25.0–29.9); normal (BMI 18.5–24.9); and underweight (BMI < 18.5). Race/ethnicity was based on self-reported information and was categorized into three groups: non-Hispanic white, non-Hispanic black, and others. Education was categorized into three groups: less than high school, high school, and more than high school.
Self-reported physical activity was categorized into two groups, non-sedentary vs. sedentary, based on the ‘yes’ vs. ‘no’ response to the question “Do you do any moderate-intensity sports, fitness, or recreational activities that cause a small increase in breathing or heart rate such as brisk walking, bicycling, swimming, or golf for at least 10 minutes continuously?” Participants were categorized as drinkers or non-drinkers based on self-reported alcohol use, defined as having at least one drink per week: “In the past 12 months, on those days that you drank alcoholic beverages, on average, how many drinks did you have?” Smoking status was categorized as never, former, and current. Current smokers were those who responded yes to the question “Do you now smoke cigarettes?” Former smoker and never smokers were categorized based on the response to questions “Have you smoked at least 100 cigarettes in your entire life?” and “Do you now smoke cigarettes?” Diagnosis of diabetes mellitus was based on self-reported response to the question “Have you ever been told by a doctor or health professional that you have diabetes or sugar diabetes?”
Descriptive statistics were performed to describe and summarize the data. Using Rao-Scott chi-square and independent sample t-tests, bivariate analyses were performed to compare groups. We then used a multivariable logistic regression model to calculate the adjusted odds ratio (aOR) and 95% confidence interval (95% CI) for basic physical functioning difficulty by serum Zn level tertile, using the lowest tertile as the reference group. Pairwise interaction terms between serum Zn levels and covariates were included in the multivariable regression to examine subgroup differences in the association between serum Zn status and basic physical functioning difficulty. The covariates that were included in this study for the multivariable model regression were age, sex, BMI, race, education, status of health insurance, smoking, alcohol use, physical activity, diabetes, serum albumin, and blood hemoglobin levels. Covariates were retained in a multivariable model regardless of the results of univariable analyses. Potential multicollinearity was assessed using a variance inflation factor. Regression analysis was not prone to multicollinearity. Statistical significance was set at p ≤ 0.05. SAS® version 9.4 software (SAS Institute, Inc., Cary, NC) was used to perform all statistical analyses. All analyses incorporated sample weights that account for the unequal probabilities of selection, oversampling, nonresponse, and complex survey design elements of NHANES.