The KNHANES is a cross-sectional, stratified, and multi-stage survey of representative samples of the non-institutionalized Korean population taken since 1998. It comprises a health interview survey, health behavior survey, medical examination, and nutrition survey. For trends in cardiovascular health metrics, we used data from KNHANES III 2005 (n = 34,145), and KNHANES IV 2007 (n = 4,594), 2008 (n = 9,744), and 2009 (n = 10,533).

Figure 1 shows the flow of the number of subjects. The participants of KNHANES were supposed to respond to 4 examinations (health interview survey, health behavioral survey, medical examination, and nutritional survey). Age, sex, current disease, and its medication were included in the health interview survey, and smoking and physical activity were questioned in the health behavioral survey. Blood pressure, body mass index, fasting blood glucose, and total serum cholesterol was measured during a medical examination, and healthy diet under the nutritional survey. Non-pregnant subjects 20 years or older were selected for this study and we count the number of participants responding to each cardiovascular health metric and use this data for analysis of trends. The participants who have all information on the seven health metrics were included for analysis of the sum of the health metrics. In this study, KNHANES I and KNHANES II were excluded because they had quite different questionnaires in the health behavior survey from the recent KNHANES.

The AHA defined ideal, intermediate, and poor cardiovascular health metrics for adults.³⁾ Data from KNHANES III and IV includes the health metrics information, and the data was arranged according to the health metrics defined by AHA with modifications of body mass index, physical activity, and healthy diet.

The subgroup of smoking status is classified as never-smoking group, former-smoking group, and current-smoking group. The smoker was defined as a participant who has smoked over 100 cigarettes. The AHA definition of ideal smoking status includes never-smokers and former smokers who had not smoked for at least 12 months.³⁾ KNHANES III included information from the time of quitting smoking and it was reflected in ideal smoking status. However, the time of quitting smoking is not described in the questionnaire from KNHANES IV. We classified 'never,' 'former,' and 'current' categories as 'ideal,' 'intermediate,' and 'poor' status, respectively in KNHNES IV. Smoking score is classified as either ideal (never and former: quit at least 12 months ago [1 point]), or poor (current smoking and former: quit within 12 months [0 point]) in KNHANES III, and as ideal (never [1 point]) or poor (former or current smoking [0 point]) in KNHANES IV.

Physical activity was converted into numerical values with calculation of metabolic equivalents (METs, min/wk) by utilizing sum of activities per week and its duration time. The physical activity questionnaire of KNHANES consists of duration and the number per week of vigorous activities, moderate activities, and walking. The total METs (min/wk) is the sum of vigorous METs (min/wk) multiplied by 8, moderate METs (min/wk) multiplied by 4, and walking METs (min/wk) multiplied by 3.3.⁹⁾ We defined the subjects as physically active (ideal [high] category) if their vigorous activity on at least 3 days achieved a minimum total physical activity of at least 1,500 METs (min/wk) or 7 or more of any combination of walking and moderate or vigorous activities achieving a minimum total physical activity of at least 3,000 METs (min/wk).⁹⁾ The intermediate category is defined as 3 or more days of vigorous activity of at least 20 minutes per day, 5 or more days of moderate activity and/or walking of at least 30 minutes per day, or 5 or more days of any day combination of walking, moderate or vigorous activities achieving a minimum total physical activity of at least 600 METs (min/wk).⁹⁾ Those individuals who did not meet criteria for intermediate or vigorous category were considered to have a low physical activity level.⁹⁾ For calculation of health metrics, the intermediate physical activity group and low physical activity group were merged with the poor group.

The healthy diet score of AHA is calculated by summing the following items, allotting 1 point each: eating fruits and vegetables(≥4.5 cups/d), fish (≥two 3.5-oz servings/wk), fiber-rich whole grains(≥three 1-oz equivalent servings/d), sodium (<1,500 mg/d), and sugar-sweetened beverages (<36 oz/wk).³⁾ However, it is difficult to estimate the healthy diet score based on the food frequency questionnaire (FFQ) of KNHANES. Nutritional data of KNHANES includes total daily calories and each nutrition fact (i.e., daily protein, daily carbohydrates, daily fat, etc.) individually converted from FFQ. According to the guideline for the DASH (Dietary Approaches to Stop Hypertension) diet,¹⁰⁾ we set up the new healthy diet score modified by the Korean diet guidelines for dyslipidemia:¹¹⁾ total daily fat consumption < 35% of total calories, total daily protein consumption > 15% of total calories, total daily carbohydrate consumption < 55%, total daily sodium consumption < 2,300 mg, and total daily fiber consumption > 20 g. Although AHA defines a healthy diet as meeting the above on 4 out of 5 points, we dichotomized the healthy diet score as fewer than 2 vs. 2 or more. A healthy diet score was defined as 2 points or greater owing to the paucity of the participants with a score of 3 or greater (range, 12.4% to 15.2%). Such a healthy diet score system was applied similarly in a previous study.⁴⁾

The values of fasting plasma glucose and total serum cholesterol were available as part of the medical examination of KNHANES participants. We classified subgroups of fasting plasma glucose level as an ideal group (fasting plasma glucose < 100 mg/dL), intermediate group (100 mg/dL ≤ fasting plasma glucose < 126 mg/dL), and poor group (fasting plasma glucose ≥ 126 mg/dL). Similarly, we classified subgroups of serum total cholesterol level as an ideal group (total serum cholesterol < 200 mg/dL), intermediate group (200 mg/dL ≤ total serum cholesterol < 240 mg/dL), and poor group (total serum cholesterol ≥ 240 mg/dL). The participants who take lipid-lowering medication were excluded among the ideal group.

The participants for KNHANES III, IV's blood pressure was taken and we used the mean blood pressure of the three measurements. We classified subgroups of blood pressure into an ideal group (systolic blood pressure < 120 mm Hg and diastolic blood pressure < 80 mm Hg), intermediate group (120 mm Hg ≤ systolic blood pressure < 140 mm Hg or 80 ≤ diastolic blood pressure < 90 mm Hg), and poor group (systolic blood pressure ≥ 140 mm Hg or diastolic blood pressure ≥ 90 mm Hg). The participants who take antihypertensive medication were excluded among the ideal group.

Although ideal body mass index was defined as < 25 kg/m² by AHA,³⁾ we redefined ideal body mass index as < 23 kg/m² based on the opinion that for many Asian populations, body mass index ≥ 23 kg/m² represents increased cardiovascular risk.¹²⁾ We classified subgroups of fasting plasma glucose level as an ideal group (body mass index < 23 kg/m²), intermediate group (23 kg/m² ≤ body mass index < 25 kg/m²), and poor group (body mass index ≥ 25 kg/m²).

We built up a cardiovascular health metrics score (sum of 7 cardiovascular health metrics) by recoding the 7 metrics as dichotomous variables granting 1 point for the ideal category vs. 0 points for the other categories for physical activity, body mass index, total serum cholesterol, blood pressure, and fasting blood glucose. A healthy diet score is classified as (≥2 components [0 point] vs. <2 [1 point]).

All participants were classified as meeting 0, 1, 2, 3, 4, 5, 6, or 7 cardiovascular health metrics. For this score, we excluded participants missing data on 1 or more of the cardiovascular health metrics.

KNHANES is a dataset organized by using complex sampling design, and we estimated prevalence of each health metric adjusted by weight of the sample in each year in order to minimize selection bias. We calculated prevalence of each metric of KNHANES III (2005), KNHANES IV (2007), KNHANES IV (2008), and KNHANES IV (2009) by Taylor-linearized variance estimation. Predictive probability of each metric was analyzed by logistic regression with adjustment for age, sex, and residential district. And the predictive probability curve was depicted for tangible trends in each metric divided into male and female chronologically with P for trends calculated by logistic regression. Mean of health metric sum for each year was calculated by analysis of covariance, adjusted for age, sex, and residential district, and its P for trend was tested by linear regression. Data were analyzed using Stata SE ver. 12.1 (Stata Co., College Station, TX, USA) and P < 0.05 was considered statistically significant.