Study design
The NHANES updated at regular biennial intervals by the National Center for Health Statistics (NCHS) and the Centers for Disease Control and Prevention (CDC), representing the national population of American of all ages(11). A complex and multistage probability sampling strategy was used to optimize the reliability and precision among representative samples(12). Family interviews, physical exams, laboratory tests, and questionnaires were collected. A total of 3 cycles were used into our study, covering the years 2007 to 2012 (2007-2008, 2009-2010, 2011-2012). There were 17713 individuals aged from 20 years to older. After excluding individuals with missing baseline daily energy intake data (n=3,684), we further excluded those with cancer (n=1388), pregnant (n=110) and daily intaking less than recommended calories (male: <800 kcal/day or >4200 kcal/day, female: <500 kcal/day or >3500 kcal/day) (n=105). Additionally, those individuals without available pulmonary function (FVC and FEV1, n=1585) and missing data on potential covariates (n=1914) were also not enrolled. Finally, a total of 8926 participants with complete data were enrolled in our study. The screening process is shown in Figure 1. The data and detailed information can be traced: https://www.cdc.gov/nchs/nhanes/index.htm.
Pulmonary function assessment
All operation adhered to the guideline of the American Thoracic Society (ATS). At the beginning standardized patient (SP) took a deep breath (at least 6s), with the chart driver started scrolling, SP can take the tube at their mouth and blow out as quickly as possible. Following the prompting from the spirometry system, the SP repeated the tests, resulting in acceptable/repeatable results. FVC and FEV1 were obtained by the above methods and applied in this study(13).
Composite dietary antioxidant index assessment
Highly trained professionals collected the information of dietary and their components on two days through a 24h dietary recall interview. In present investigation, all participants had two 24h dietary recall record, the first recall was carried out by face-to-face interviews, and the second proceed 3-10 days later over the telephone. By reviewing dietary data and total nutrient intake files, the average daily intake for 2 days was calculated to calculate daily total energy and nutrients intakes from foods and beverages. The NHANES provided the intake of specific antioxidants, including Zinc, Se, carotenoid, vitamins A, C, and E. These antioxidants were recommended by a previous study for calculating the CDAI scores (10). The calculation formula of CDAI can be referred in a previously published study. Firstly, the standardization of individual components was achieved by subtracting the mean specific to sex and dividing by the standard deviation specific to sex. Secondly, we summed into the standardized intake of individual antioxidant to obtain the CDAI(10).
Potential covariates
The factors of reference for potential confounding variables associated with CDAI and pulmonary function were common clinical confounders and previous studies. Demographic variables include age, gender/sex, ethnicity/race, education level, and family income to poverty ratio. Divide the weight in kilograms by the height in meters squared to determine the body mass index (BMI). BMI was classified as dichotomous variable: normal (<25 Kg/m2), overweight/obese (≥25 Kg/m2)(14). Participants with an average blood pressure ≥140 mmHg systolic and/or ≥90 mmHg diastolic, or told of hypertension by physicians, or taking antihypertensive medication can be regarded as having hypertension(15). Hyperlipidemia can be defined as participants who had one of triglyceride ≥150mg, total cholesterol ≥200mg/dl, low-density lipoprotein ≥130mg/dl, high-density lipoprotein <40mg/dl, or taking lipid-lowering medications(16). The diagnosis of type 2 diabetes was self-reported diagnosed of diabetes, serum biochemical indicators reached the standards, or using of diabetes medications or insulin(17,18). Those individuals with pre-diabetes, referring to abnormal fasting glucose or impaired glucose tolerance, have not yet reached the diagnosis standard of diabetes. The Healthy eating index-2015 (HEI) contains common and additional dietary components, the total score is 100, and individuals who have higher scores mean consuming a healthier diet(19). According to the participant’s smoking status, it’s defined as never (smoking less than 100 cigarettes in life), former (smoking more than 100 cigarettes in life and smoking not at all now), and now (smoking more than 100 cigarettes in life and smoking some days or every day). Alcohol use can be divided into never (had <12 drinks in life), current mild alcohol use (female and male, had ≤1 or ≤2 drinks a day, respectively), current moderate alcohol use (female and male, had ≥2 but <3 drinks a day or ≥3 but <4 drinks a day, respectively, or binge drinking ≥2 and <5 days per month ), and current heavy alcohol use (female and male, had ≥3 or ≥4 drinks a day, respectively, or binge drinking ≥5 days per month ), former alcohol use (had ≥12 drinks in 1 year and did not drink last year, or did not drink last year but drank ≥12 drinks in a lifetime). Cardiovascular disease (CVD) was acquired from the Medical Conditions questionnaire survey, individuals who told to the physician that they had coronary heart disease, congestive heart failure, heart attack, stroke, or angina can be deemed as having CVD(20). Respiratory disease (RD) diagnosed as fulfilling one of the following criteria: 1) physician ever told participants have asthma, chronic obstructive pulmonary disease (COPD), or both; 2) use drugs (selective phosphodiesterase-4 inhibitors, mast cell stabilizers, leukotriene modifiers, and inhaled corticosteroids); 3) FEV1/FVC <0.7 (Post-Bronchodilator).
Statistical analysis
The data employed in our final analysis was weighted complying with the guideline of the stipulated analytical. The CDAI was distributed into quartiles, from Q1 (≤ -2.50), Q2 (-2.50, -0.58), Q3 (-0.58, 1.90), to Q4 (>1.90) respectively. Categorical variables were presented as the frequency with proportion, and continuous variables were exhibited by mean ± standard error (SE). Chi-square test (categorical variables) and one-way Analysis of Variance (ANOVA) (continuous variables) were carried out to calculate differences among baseline characteristics of the individuals. To account for the influence of the potential confounders, unadjusted and adjusted weighted linear regression models were performed to investigate the correlation of CDAI (continuous, quartiles) and pulmonary function (FVC, FEV1). In addition, a trend test was also carried out by CDAI quartiles to confirm the association consistency. Furthermore, Restricted Cubic Spline (RCS) was conducted to verify the linearity and the dose-response relationship between CDAI and outcome variables. We also performed subgroups analysis to test the stability in various populations and carried out interaction tests to detect whether other factors jointly affect lung function with CDAI. Software R (4.4.1) were carried out to conduct all statistical analyses. Two-tailed P<0.05 was considered statistically significant.