We analyzed the participants of the second year (2008) of KNHANES IV. KNHANES was initiated in 1998 and was designed to assess the health and nutritional status of adults and children over 1 year of age [10]. Specifically, KNHANES targeted the noninstitutionalized civilian Korean population. The KNHANES IV was a cross-sectional and nationally-representative survey conducted by the Division of Chronic Disease Surveillance, Korea Centers for Disease Control and Prevention from 2007 to 2009. The data collected consisted of a health interview survey, a health examination survey, and a nutrition survey. A detailed description of the sampling method used was published in the 2007 KNHANES [11]. In this survey, 12,528 individuals in 4,600 households were sampled, with 9,308 of them participating in the health interviews and health examination surveys, and 8,641 participating in the nutrition surveys. Among those who participated in the survey between February 2008 and December 2008, serum 25(OH) D levels were obtained in 3,047 males and 3,878 females aged 10 years and older from all 16 administrative districts of South Korea. The present study was restricted to participants over 50 years of age who completed the health examination survey and had their serum 25(OH) vitamin D levels measured (n=2,710). This study analyzed publicly available data sets and was therefore exempted from institutional review board approval.

In the KNHANES, the patients' body heights, body weights, and waist circumferences (WCs) were measured by a specially trained examiner. We calculated BMI by dividing weight in kilograms by height in square meters (kg/m²). BMI was used to categorize the participants as normal weight (BMI <23.0 kg/m²), overweight (BMI ≥23 and <25 kg/m²), or obese (BMI ≥25 kg/m²). We also categorized patients as abdominally obese if their WC was over 90 cm (in women, the cutoff was 85 cm) [12]. We categorized patients as normal if their WC was under 90 cm for men and under 85 cm for women.

Blood samples were obtained after a minimum fasting time of 8 hours. Blood samples were collected from each participant during the survey and were processed, immediately refrigerated, and transported in cold storage to the Central Testing Institute in Seoul, Korea. Serum 25(OH) D concentration was measured by RIA (DiaSorin Inc., Stillwater, MN, USA) using a γ-counter (1470 Wizard, PerkinElmer, Turku, Finland). The interassay coefficients of variation were 11.7%, 10.5%, 8.6%, and 12.5% at 8.6, 22.7, 33.0, and 49.0 ng/mL, respectively. The serum levels of total cholesterol (TC), triglycerides, high density lipoprotein cholesterol (HDL-C), aspartate aminotransferase, and alanine aminotransferase (ALT) were measured enzymatically using a Hitachi Automatic Analyzer 7600 (Hitachi, Tokyo, Japan). Insulin concentration was measured by immunoradiometry (INS-IRMA, Bio-Source, Nivelles, Belgium). The homeostasis model assessment estimate of insulin resistance (HOMA-IR) was calculated using the following formula: HOMA-IR=fasting plasma glucose (mmol/L)×fasting insulin (mIU/L)/22.5 [13].

The measurement of lifestyle-related and metabolic parameters was conducted via interviews by trained interviewers. Patients were categorized by smoking status as current smokers, past smokers, or nonsmokers. We also categorized participants based on their alcohol consumption as drinkers or nondrinkers based on questions about their drinking behavior. We used the International Diabetes Federation (IDF) criteria for metabolic syndrome [14]. The classification of metabolic syndrome included the presence of central obesity (>90 cm in men, and >85 cm in women, based on the recommendation of the Korean Society for the Study of Obesity [12]), in addition to any two of the following four risk factors: hypertension ≥130/≥85 mm Hg or treatment of previously diagnosed hypertension, triglyceride levels ≥150 mg/dL (1.7 mmol/L), low HDL-C: in men <40 mg/dL (1.03 mmol/L) and in women <50 mg/dL (1.29 mmol/L), and fasting plasma glucose ≥100 mg/dL (5.6 mmol/L), a history of diabetes mellitus, or currently taking antidiabetic medications.

Statistical analyses were performed using SPSS version 15.0 (SPSS Inc., Chicago, IL, USA). The data are presented as means±SD unless stated otherwise. Categorical variables are reported as frequencies and proportions. An independent two-sample t test was used to compare two independent groups. The Pearson's chi-squared test was employed to analyze categorical data, as appropriate. The distributions of the continuous variables were examined for skewness and kurtosis, and the logarithm-transformed values of the nonnormally distributed variables were used for comparison between the two groups. We performed multiple regression analysis to investigate the association between serum 25(OH) vitamin D and several variables (BMI, WC, fasting insulin, and HOMA-IR) after adjusting for age, sex, education, physical activity, smoking, alcohol consumption, and history of cardiometabolic disorders. Analyses were also conducted after the subjects were stratified by obesity to identify the association between serum 25(OH) vitamin D and insulin resistance when considering obesity status. All P values were two-tailed and results were considered statistically significant if the P values were <0.05.

The cross-sectional analyses included data on 2,710 participants (1,119 men and 1,591 women), aged ≥50 years. The mean age of the study participants was 63.8±9.2. Table 1 shows the characteristics of the study participants according to their obesity status (nonobese, BMI <25 kg/m²; obese, BMI ≥25 kg/m²). In comparison with the nonobese group, the obese group had a higher systolic blood pressure, diastolic blood pressure, TC, triglyceride, and ALT levels. The fasting glucose, insulin, and HOMA-IR of the obese group were all significantly higher than the nonobese group. Metabolic syndrome, as defined by IDF criteria, was more frequently observed in the obese group (59.5% vs. 11.2%, P<0.001). The mean serum 25(OH) Vitamin D level in the obese group was 51.3±18.6 nmol/L, significantly lower than that of the nonobese group (53.3±20.5 nmol/L, P=0.009).

After adjusting for age and sex, BMI was found to be inversely associated with 25(OH) vitamin D levels (P=0.002) (model 1 in Table 2). However, WC was not significantly associated with 25(OH) vitamin D levels (P=0.123). After adjusting for education, physical activity, smoking, alcohol consumption, and history of cardiometabolic disorders, the associations between 25(OH) vitamin D level and BMI or WC did not change (P=0.008 and P=0.433, respectively) (model 2 in Table 2).

In the stratified analyses, we adjusted the regression coefficients of 25(OH) vitamin D with insulin resistance according to BMI (BMI <23 kg/m² [n=1,043], BMI ≥23 and <25 kg/m² [n=700], and BMI ≥25 kg/m² [n=700]). After adjusting for age and sex, serum 25(OH) vitamin D was found to be inversely associated with fasting insulin and HOMA-IR in all of the participants regardless of obesity status. Serum 25(OH) vitamin D was found to be statistically associated with fasting insulin and HOMA-IR in all participants except for the fasting insulin in participants with BMIs <23 kg/m² (P=0.065) (model 1 in Table 3). After adjusting for education, physical activity, smoking, alcohol consumption, and history of cardiometabolic disorders, serum 25(OH) vitamin D was found to be negatively associated with fasting insulin and HOMA-IR in participants with BMIs ≥25 kg/m² (P=0.003 for both insulin and HOMR-IR) but not in those with BMI <23 kg/m² (P=0.158 and P=0.069 for insulin and HOMA-IR, respectively) (model in Table 3).

We also evaluated the regression analysis between serum 25(OH) vitamin D and insulin resistance, as stratified by abdominal obesity. The participant's abdominal obesity status was categorized as normal (n=1,163) or obese (n=1.047) (WC ≥90 cm for men, ≥85 cm for women). After adjusting for age and sex, 25(OH) vitamin D was statistically associated with fasting insulin and HOMA-IR in both the normal and abdominally obese group (P<0.001 for both fasting insulin and HOMA-IR in the normal group, P=0.002 and P=0.001 for fasting insulin and HOMA-IR in the abdominally obese group) (model in Table 4). After adjusting for education, physical activity, smoking, alcohol consumption, and history of cardiometabolic disorders, the correlations did not change in both groups (P=0.001 and P<0.001 for fasting insulin and HOMA-IR in the normal group, P=0.001 and P=0.009 for fasting insulin and HOMA-IR in the abdominally obese group) (model in Table 4).