This study utilized two population-based cohorts: the 2021 Korea National Health and Nutrition Examination Survey (KNHANES) and the Korean Genome and Epidemiology Study (KoGES). The KNHANES is a nationwide representative population-based survey. The cohort profile was described in detail in a previous study (20). The 2021 KNHANES dataset included 7,090 participants aged 19 years and older. The KoGES Ansan and Ansung study is a community-based prospective cohort study. The study design and procedures have been described in detail in a previous study (21). The KoGES Ansan and Ansung cohort included 10,030 adults aged between 40 and 69 years. This survey was first conducted in 2001 and 2002, and participants were followed up every 2 years. In the present study, we included participants with up to eight follow-up evaluations conducted between 2017 and 2018.

A flowchart of the study population is provided in Figure 1 (KNHANES and KoGES). Among the 7,090 participants in the 2021 KNHANES dataset, 2,735 participants (1,185 men and 1,550 women) were finally included for analysis after excluding participants who had missing data with which to evaluate MetS (n = 1,197) and to calculate OBS (n = 3,158). Among the 10,030 participants in the KoGES dataset, 5,807 participants (2,921 men and 2,886 women) were finally included for analysis after excluding participants with missing data needed for MetS evaluation (n = 3), those with MetS at baseline (n = 3,197), those who lacked data needed to calculate OBS (n = 457), and those who did not participate in a follow-up after the baseline survey (n = 566).

All participants in KNHANES and KOGES provided informed consent for data collection. The study protocol conformed to the ethical guidelines of the 1964 Declaration of Helsinki and its later amendments. This study was approved by the Institutional Review Board (IRB) of Nowon Eulji Medical Center (IRB number: 2021-09-025).

OBS was calculated from data on six pro-oxidant factors and five antioxidant factors selected based on previous studies (14, 22–24), and data for all 11 factors were available in the both KHNANES and KoGES datasets. Table 1 presents the OBS assignment scheme. Pro-oxidant factors included saturated fatty acid (SFA), total iron intake, smoking status, drinking status, obesity status, and abdominal obesity status (14, 25, 26). Each query was scored 0, 1, or 2, except for abdominal obesity status, which was scored 0 or 1. The scores for SFA and total iron intake were assigned 0–2 points according to sex-specific tertile values (low, 2; intermediate, 1; high, 0) of each variable. The scores for never smokers, former smokers, and current smokers were 2, 1, and 0, respectively. The scores for non-drinkers, mild-to-moderate drinkers (1–29 g/day in men, 1–19 g/day in women), and heavy drinkers (≥30 g/day in men, ≥20 g/day in women) were 2, 1, and 0, respectively. Obesity received 0 points, overweight received 1 point, and normal weight received 2 points. Zero points were given for abdominal obesity and 1 point for no abdominal obesity. Antioxidant factors included physical activity (high intensity, 2; moderate intensity, 1; low intensity, 0) and intake of omega-3 poly-unsaturated fatty acid (PUFA) to omega-6 PUFA ratio, vitamin C, vitamin E, and beta-carotene, which were each assigned 0–2 points according to sex-specific tertile values (low, 0; intermediate, 1; high, 2) for each variable. Using OBS, which had a possible maximum of 21 points, we divided the individuals into tertile groups by sex. Higher OBS indicate higher antioxidant properties. In the KoGES and KNHANES datasets, OBS cut-off points were stratified into three categories: for men, T1 (≦11 for KoGES and ≦10 for KNHANES), T2 (12–13 for KoGES and 11–12 for KNHANES), and T3 (≧14 for KoGES and ≧13 for KNHANES); for women, T1 (≦12), T2 (13–14), and T3 (≧15) in both datasets.

Based on the National Cholesterol Education Program Adult Treatment Panel III definition as amended by the National Heart Lung and Blood Institute and the American Heart Association (27), we defined MetS as the presence of three or more of the following: (1) abdominal obesity as a waist circumference ≥ 90 cm in men and ≥ 85 cm in women based on the 2018 Korean Society for the Study of Obesity guidelines (28); (2) fasting plasma glucose ≥ 5.6 mmol/L or use of oral hypoglycemic medication or insulin; (3) serum triglyceride concentration ≥ 1.7 mmol/L or use of lipid-lowering medication; (4) serum high-density lipoprotein (HDL) cholesterol concentration < 1.0 mmol/L in men or < 1.3 mmol/L in women; and (5) systolic blood pressure (SBP) ≥ 130 mmHg, diastolic blood pressure (DBP) ≥ 85 mmHg, or use of anti-hypertensive medication.

Well-trained medical staff members performed the health examination sand health interviews following the KHANES protocol (20) and KoGES protocol, respectively (29). Body mass index (BMI) was calculated as a person’s weight in kilograms divided by the square of their height in meters. Individuals with a BMI of at least 23 kg/m² but less than 25 kg/m² were considered overweight, and those with a BMI of ≥25 kg/m² were considered obese based on the 2018 Korean Society for the Study of Obesity guidelines (28). Abdominal obesity was defined as a waist circumference ≥ 90 cm in men and ≥ 85 cm in women based on the same guidelines (28). Mean blood pressure (MBP, mmHg) was calculated as DBP + 1/3 × (SBP – DBP) (30). Whole blood white blood count (WBC), plasma glucose, serum insulin, total cholesterol, triglycerides, HDL, cholesterol and C-reactive protein (CRP) were assessed using a Hitachi 700–110 Chemistry Analyzer (Hitachi, Ltd., Tokyo, Japan) after at least 8 h of fasting.

Information on smoking, alcohol consumption, physical activity, education level, and household income were obtained from self-reported questionnaires. A never smoker was defined as someone who had never smoked or had smoked fewer than 100 cigarettes in their lifetime. Former smokers were those who had quit smoking and had smoked more than 100 cigarettes during their lifetime. Current smokers were those who were active smokers at the time of the questionnaire and had smoked more than 100 cigarettes during their lifetime.

We further evaluated the daily alcohol intake (g/day) of each participant. Heavy drinkers were men who drank more than 30 g/day and women who drank more than 20 g/day. Mild-to-moderate drinkers were men who drank less than 30 g/day and women who drank less than 20 g/day. Non-drinkers were those who did not drink alcohol.

Physical activity was assessed using the metabolic equivalent of task–hours per week (MET-hrs/day) using the International Physical Activity Questionnaire (31). Participants were classified into low- (<7.5 METs-hr/day), moderate- (7.5–30 METs-hr/day), or high- (>30 METs-hr/day) intensity physical activity groups. An in-person interview concerning nutrition was conducted in the respondent’s home.

Total energy intake and nutritional status were calculated using a validated 112-item food frequency questionnaire (FFQ) developed for KNHANES (32) and 103-item FFQ developed for KoGES (33). Information for each item was collected based on recalls of the average frequency and amount consumed per serving over the past year. From the FFQ, this study considered the total daily intake values of energy (kcal/day), SFA (g/day), total iron (mg/day), omega-3 PUFA to omega-6 PUFA ratio, vitamin C (mg/day), vitamin E (mg/day), and beta-carotene (μg/day).

Participants were classified into elementary/middle school, high school, and college/university education levels. Monthly household income was divided into three groups: <100 million Korean Won, 100–200 million Korean Won, and > 200 million Korean Won.

We conducted a normality test, and variables with normal distribution are presented as means ± standard deviations. Serum triglyceride, insulin, and CRP with non-normal distribution are presented as medians (25th percentile, 75th percentile). Continuous variables were compared using a one-way analysis of variance or the Kruskal–Wallis test according to the sex-specific OBS tertiles. All statistical analyses were conducted separately for men and women. Categorical variables are described as numbers (%) and were compared using the chi-square test. The dose–response relationship between OBS and the risk of incident MetS was determined using a Cox proportional hazard spline curve. Kaplan–Meier curves with the log-rank test were used to verify the cumulative incidence of MetS according to the sex-specific OBS tertiles. For KNHANES, we calculated the odds ratio (OR) and 95% confidence intervals (CI) for the prevalence of MetS. For KoGES, we completed univariable and multivariable Cox proportional hazard regression analyses to calculate the HRs and 95% CIs for incident MetS of the second tertile (T2) and highest tertile (T3) groups, compared with the referent lowest tertile (T1) group, in a sex-specific manner. In the fully adjusted model, we adjusted for age, education level, monthly household income, total energy intake, MBP, WBC, fasting plasma glucose, and serum total cholesterol levels based on each population characteristics. Confounders were determined as variables influencing exposure and outcome based on a literature review and univariate analysis (34, 35). We calculated variance inflation factor (VIF) values to assess the degree of multicollinearity among the variables, indicating that a VIF value of five or higher represents a high correlation of the variables (36). In this study, the maximum VIF value observed was 2.1, which corresponded to the variable ‘age’ in women (Supplementary Tables 1, 2). All statistical analyses were performed with SAS statistical software (version 9.4; SAS Institute Inc., Cary, NC, USA) and R software (version 4.1.1; R Foundation for Statistical Computing, Vienna, Austria). P-values less than 0.05 were regarded as statistically significant.

Table 2 shows the baseline characteristics of the men according to OBS tertile in the KoGES and KNHANES, respectively. In the KoGES, the T3 group had lower fasting glucose, insulin, total cholesterol, triglyceride, and WBC. T3 men had higher intake of SFA, total iron, omega-3 to omega-6 PUFA ratio, vitamin C, vitamin E, and beta-carotene. Since we excluded participants with existing MetS at the baseline survey of KoGES, the proportion of participants with two MetS components was lowest in the T3 group.

In the KNHANES, the T3 men had older age and lower MBP, insulin, triglyceride, and WBC. The proportion of MetS was significantly higher in T1 men. The T3 men had lower intake of SFA, and higher intake of omega-3 to omega-6 PUFA ratio, vitamin C, vitamin E, and beta-carotene. The T3 men exhibited lower proportions of current smokers, heavy drinkers, obese, and abdominal obesity and greater high intensity physical activity in both the KoGES and KNHANES.

Table 3 shows the baseline characteristics of the women according to OBS tertile in the KoGES and KNHANES, respectively. In the KoGES, the T3 women had lower MBP, insulin, total cholesterol, triglyceride, and WBC. The proportion of participants with two MetS components was lowest in the T3 group. T3 women had higher intake of SFA, total iron, omega-3 to omega-6 PUFA ratio, vitamin C, vitamin E, and beta-carotene.

In the KNHANES, the T3 women had older age and higher HDL-cholesterol, but lower insulin, triglyceride, and WBC. The proportion of MetS was significantly higher in T1 women. The T3 women had lower intake of SFA and higher intake of omega-3 to omega-6 PUFA ratio, vitamin C, vitamin E, and beta-carotene. The T3 women had lower proportions of current smokers, heavy drinkers, obese, and abdominal obesity and greater high intensity physical activity in both the KoGES and KNHANES.

During the mean 13.6-year follow-up period, 1,296 (44.4%) men and 880 (30.5%) women developed new-onset MetS in the KoGES. Baseline characteristics of men and women from KoGES are shown in Supplementary Table 3. Inverse relationships were noted between OBS and the risk of incident MetS in both men (Figure 2A) and women (Figure 2B). Figure 3 presents the cumulative incidence rates of MetS as Kaplan–Meier curves according to sex-specific OBS tertiles in the KoGES. The T3 group showed the lowest cumulative incidence of MetS, followed by the T2 and T1 groups, among both men (Figure 3A) and women (Figure 3B) (both log-rank test p-values <0.001).

Table 4 shows the independent association between OBS tertiles and MetS in the KoGES and the 2021 KNHANES. In the KoGES, the HRs (95% CIs) for new-onset MetS per increment in OBS were 0.90 (0.88–0.92) in men and 0.88 (0.86–0.91) in women. Compared to referent T1, the adjusted HRs (95% CIs) for new-onset MetS were 0.82 (0.72–0.93) in T2 and 0.56 (0.48–0.62) in T3 among men. Compared to referent T1, the adjusted HRs (95% CIs) for new-onset MetS were 0.71 (0.62–0.81) in T2 and 0.63 (0.55–0.73) in T3 among women.

In the 2021 KNHANES, the ORs (95% CIs) for MetS per increment in OBS were 0.89 (0.85–0.93) in men and 0.88 (0.84–0.93) in women. Compared to referent T1, the adjusted ORs (95% CIs) for MetS in T3 were 0.44 (0.29–0.65) in men and 0.34 (0.23–0.50) in women. When we conducted the subgroup analysis in adults aged 40–69 years from KNHANES, the significant association between OBS and incident MetS remained (T3 vs. T1, adjusted ORs and 95% CIs = 0.39 [0.24–0.65] in men and 0.26 [0.16–0.43] in women) (Supplementary Table 4).