Demographic data on sex, age, education, marriage, cohabitation during the past year, occupation, household income, and subjective economic status are collected. The medical history covers the participant’s history of diseases, family history, medication, and female reproductive factors. The presence of a diagnosis by a physician and the age at the first diagnosis of each of the following diseases are investigated: stroke, transient ischemic attacks, myocardial infarction, angina pectoris, heart failure, chronic kidney disease, hypertension, dyslipidemia, diabetes, thyroid disease, fatty liver, chronic hepatitis, liver cirrhosis, asthma or chronic obstructive pulmonary disease, osteoporosis, arthritis, autoimmune disease, and malignant tumors. If a participant reports a history of stroke, myocardial infarction, heart failure, or active cancer, he/she is excluded from the study. Participants are also required to report whether a family member (father, mother, and siblings) has been diagnosed with or died from specific diseases (myocardial infarction, hypertension, stroke, or diabetes), and to provide their family member’s age at the first diagnosis of the disease or at death. Female participants are required to provide additional information on reproductive health, including age at menarche, menstruation during the past 3-month, menopause, age at menopause, number of pregnancies (total, childbirth, miscarriage, and stillbirth), the presence of gestational diabetes or hypertension, oral contraceptive use, and female hormone use.

Information on cigarette smoking and alcohol drinking are collected using standardized questions that gather information about lifetime smoking, current smoking, age at smoking initiation, the amount and duration of smoking, lifetime drinking, current drinking, and frequency of consumption and average amount consumed of each type of alcohol. Dietary exposures are evaluated using a semi-quantitative food frequency questionnaire (FFQ) that was developed and validated for the KNHANES [9]. The FFQ asks for participants’ average consumption frequency and amount of 112 food items during the preceding year [10]. The FFQ has an acceptable level of validity and reliability for assessing usual dietary intake [11]. Physical activity is evaluated using a Korean version of the International Physical Activity Questionnaire-Short Form [12], which asks for the frequency of each activity (walking, moderate-intensity activity, and vigorous activity) and the duration thereof during the past 7-day. The sleep questionnaire covers quantity (bedtime, rise time, sleep duration) and quality (difficulty in falling asleep, and in sleeping again after waking) of sleep and a screening for obstructive sleep apnea known as the Berlin questionnaire [13]. All the health behaviors listed above are obtained based on participants’ subjective reports through a face-to-face interview. Additionally, physical activity and sleep habits are measured using a wrist-worn accelerometer (GENEActive, Activinsight, Kimbolton, Cambridge, UK) for a subsample of the participants at the YUCM clinic. The device applied in this study is a 3-axial seismic acceleration sensor housed in a small (43×40×13 mm), lightweight (16 g), and waterproof case, and makes objective measurements of participants’ free-living physical activities [14]. To assess participants’ usual data on sleep and physical activity, they are required to wear the device for seven consecutive days, and also to record a log of details about how they wore the device, such as the time of detachment/reattachment and the reason for putting it on or taking it off.

Depressive symptoms and psychosocial responses to stress are assessed in a self-administered manner, not by a face-to-face-interview, and all answers are reviewed by research personnel to check for any instances of misreading, miswriting, or missing answers. Depressive symptoms during the past 2-week are assessed using the Korean version [15] of the Beck Depression Inventory-II [16], which contains 21 questions that reflect the physical and psychological aspects of depression. We also collect participants’ general psychosocial response to stress during the past 6-month, using the Korean version of the life experiences survey questionnaire [17]. The original questionnaire lists 57 possible life events, such as marriage, divorce, death of family members or friends, lifestyle changes, and educational and occupational successes and failures [18]. Of the 57 total events, 10 items that are applicable to students have been omitted in this study, and participants are additionally asked to report any important event not included in the list of 47. Participants are asked to report the presence of stressful life events and the positive or negative effects of each life event. Cognitive function is assessed only for participants aged 50 years or older using the Korean version of the Mini-Mental State Estimation for dementia screening [19]. This questionnaire includes psychometric properties such as orientation to time, place, and person; verbal memory; concentration and calculation; and the functions of language, praxis, and visuospatial construction.

We collect information about social network properties and social support via face-to-face interviews. For the measurement of social network properties, we use questionnaires that were developed for the US. General Social Survey [20] and National Social Life, Health, and Aging Project [21], and translated into the Korean language for the Korean Social Life, Health, and Aging Project [22]. Collecting a complete graph of the social network for all people in a specific community can generate all the social relations between all the people in the community. This type of social network is a complete network or a global network. In contrast, collecting network data from a sample of the population produces a more limited number of networks, in which only the respondent and his or her own social network members are listed. This type of social network is an ego-centric network that consists of the respondent (ego) and his or her social network members (alters). The CMERC cohort measures the ego-centric social networks of the participants. Participants are asked to list their spouse, acquaintances (up to 5) with whom they have discussed important matters, and another individual not in the list of up to 5 members provided earlier but still close and important to the participant. The following question is used to identify acquaintances: “From time to time, most people discuss things that are important to them with others. For example, these may include good or bad things that happen to you, problems you are having, or important concerns you may have. Looking back over the last 12 months, who are the people with whom you most often discussed things that were important to you?” They are also asked to provide information about their relationship with each network member, each network member’s demographic characteristics, such as sex, age, cohabitation status, and educational status, how often they talk to (or meet) each network member, and the degree to which they feel emotional closeness to each network member. In addition, they are asked for information about social interactions among the members of their social network.

Analyzing the collected information, we can measure ego-centric network properties, consisting of the participant (ego) and his or her social network member (alters) including network size, network density, content, composition, emotional closeness of alters, volume of contact with alters, bridging potential, and many others. More details regarding the data collection procedures and measurement of social network properties were presented in a previous study [22].

Height is measured to the nearest 0.1 cm using stadiometers: a DS-102 (Jenix, Seoul, Korea) and a BSM-330 (Inbody, Seoul, Korea) in the YUCM clinic, and a BSM-330 (Inbody) in the AUSM clinic. Body weight is measured to the nearest 0.1 kg on a digital scale: a DB-150 (CAS, Seongnam, Korea) and a BSM-330 (Inbody) in the YUCM clinic, and a BSM-330 (Inbody) in the AUSM clinic. Although the stadiometers and weight scale are not identical in the two research clinics, the measurement is made according to a standardized protocol. In addition, a zero-point adjustment is conducted in the same manner at least once a week using a standard ruler (170 cm) and weights (20, 40, and 60 kg). Body circumferences are measured to the nearest 0.1 cm using a plastic measuring tape (SECA 201, SECA, Hamburg, Germany), while maintaining the level of the measuring tape. Waist circumference is measured at the midpoint between the lower point of the rib cage and the upper point of the iliac crest during exhalation. Hip circumference is measured at the widest region of the hip. Mid-thigh circumference is measured at the midpoint between the inguinal crease of the right leg and the proximal border of the patella in a standing position with feet 20 cm apart and the right foot in front of the left. Mid-arm circumference is measured at the midpoint of the right arm between the acromion and the olecranon.

Single arm BP is measured in the right arm in both clinics. Participants take at least a 5-minute rest in the seated position and they are asked to remain still and relaxed during all measurements. A cuff tailored to their individual mid-arm circumference is used, and trained research personnel conduct BP measurements using an automated oscillometric device (HEM-7080, Omron Health, Matsusaka, Japan) in accordance with a standardized protocol. Systolic and diastolic blood pressures are measured three times at 2-minute intervals in the right arm. Measurements are made on the left arm only if right arm measurements are impossible. Additionally, at the YUCM clinic, dual-arm BP measurements are made three times at 60-second intervals, using a device with two cuffs developed for simultaneous dual-arm measurements (WatchBP Office, Microlife, Widnau, Switzerland) [23]. At the YUCM clinic, radial arterial pulse waves are measured and the augmentation index is calculated using a non-invasive device (HEM-9000 AI, Omron Health).

Carotid artery ultrasonography is performed at both research clinics, and all operators have been trained according to a predefined study protocol. However, the ultrasonography machines are not identical at the YUCM clinic (Accuvix XG, Samsung Medison, Seoul, Korea) and the AUSM clinic (Logiq S8 ECG module, GE Healthcare, Chalfont St. Giles, UK). While being measured, participants are in the supine position with their neck extended and their head turned 30° toward the opposite direction of the measurement site. Carotid intima-media thickness (CIMT) is scanned at the 1-cm segment of the common carotid arteries proximal to the bulb region, at the time of the R-wave on the electrocardiogram, and is calculated using the dedicated software installed in the equipment in each clinic. The mean and maximum CIMT values are measured on both the right and left sides. A carotid artery plaque is considered present if the thickness is greater by 50% or more than the surrounding intima-media thickness, or if its size is 1.5 mm or more. The presence and thickness of a carotid plaque is observed by an operator and assessed in both the right and left common, external, and internal carotid arteries, as well as at the bulb area. More details and the inter-rater reliability of the CIMT measurements have been reported separately [24].

Resting electrocardiography is conducted in the supine position using a standard 12-lead electrocardiogram. The electrocardiography machines used in this study are made by the same company, but the model numbers are different in the YUCM clinic (MAC-5500, GE Healthcare, Chicago, IL, USA) and the AUSM clinic (MAC-3500, GE Healthcare).

Participants’ bone mineral density is measured by quantitative computerized tomography (QCT) at the YUCM clinic, and by whole body dual-energy X-ray absorptiometry (DXA) at the AUSM clinic. At the YUCM clinic, the bone mineral density of the lumbar spine (T12-L2) and hip is measured using a Somatom Definition AS+ 128 channel CT (Siemens Healthcare, Forchheim, Germany), a Somatom sensation 64 channel CT (Siemens Healthcare), or a GE Lightspeed VCT apparatus (General Electric Medical System, Milwaukee, WI, USA). Scanning is done at 120 kV and 150 mAs and images are reconstructed with a 3-mm slice thickness. All scanned data are analyzed using the QCT PRO software (Mindways Software, Austin, TX, USA) and the CTXA Hip Exam Analysis protocol (Mindways Software). The AUSM clinic measures bone mineral density using whole-body DXA scans (Lunar iDXA, GE Healthcare, Madison, WI, USA), and the acquired data are analyzed with the enCORE software (GE Healthcare). A trained operator conducts all examination procedures according to the standardized protocol of each clinic.

Bioimpedance analysis is used to assess participant’s body composition parameters, such as body fat and muscle content. During the test, participants hold onto hand electrodes and stand barefoot on the foot electrodes, and spread out their arms and legs naturally.

Different models from the same company were used for measuring bioimpedance: the InBody 370 (Biospace Co., Seoul, Korea) at the YUCM clinic and the Inbody 720 (Biospace Co.) at the AUSM clinic. Both machines estimate information about body circumference (chest, waist, hip right arm, left arm, right thigh, and left thigh) and the body composition of each location (right arm, left arm, trunk, right leg, and left leg), as well as estimating total body composition (body fat, fat-free mass, and total body water). These properties are measured across the same wide range of frequencies (5, 50, and 250 kHz). In addition to the bioimpedance analysis, QCT or DXA is also used to measure body composition. The YUCM clinic assesses abdominal adipose tissue distribution, mid-thigh muscle, and liver adiposity using QCT. Images scanned at the lumbar spine, the L1 level, and the L4 level, are used to estimate the amount of visceral and subcutaneous fat and to confirm the presence of fatty liver, respectively. Voxels are separated into fat and muscle tissue by Hounsfield units (HU): −190 to −30 HU for fat and 0 to 100 HU for muscle. The AUSM clinic estimates total mass, tissue mass, lean mass, and fat mass using the DXA android region of interest [25].

Overnight-fasting blood samples and casual urine samples are obtained in the morning, and bioassays are performed at a single laboratory (Seoul Clinical Laboratories R&D Center, Seoul, Korea). Concentrations of blood glucose, blood urea nitrogen, creatinine, uric acid, total protein, and albumin are evaluated using a colorimetric method with an analyzer (ADIVA 1800 AutoAnalyzer, Siemens Medical Sol., Deerfield, IL, USA). Insulin levels are determined using a radioimmunoassay with a SR-300 apparatus (Stratec, Birkenfeld, Rhineland-Palatinate, Gemany). Lipid profiles, including total cholesterol, triglycerides, high-density lipoprotein cholesterol and low-density lipoprotein cholesterol levels, and liver function indices, including aspartate aminotransferase, alanine aminotransferase, and γ-glutamyl transferase, are analyzed enzymatically with an ADIVA 1800 AutoAnalyzer (Siemens Medical Sol.). The blood concentration of 25-hydroxy vitamin D is assessed by a chemiluminescence immunoassay using a Liason apparatus (Diasorin, Darmstadt, Hesse, Germany). High-sensitivity C-reactive protein is evaluated using the turbidimetry method with an ADIVA 1800 AutoAnalyzer (Siemens Medical Sol.). Hemoglobin A1c is determined using high-performance liquid chromatography with a Variant II Turbo (Bio-Rad, Hercules, CA, USA). Complete blood cell counts are determined using flow cytometry with an ADIVA 2120i (Siemens, Medical Sol.). Routine urinalysis is done using a urine reagent strip test with a Uriscan Pro II apparatus (YD Diagnostics, Yonginsi, Korea). The AUSM clinic additionally performs an oral glucose tolerance test, with measurements of blood glucose and insulin levels 0, 30, 60, and 120 minutes after a 75-g glucose load. The YUCM clinic additionally measures inflammatory and hemostatic markers, including prothrombin time, activated partial thromboplastin time, factor VII, factor VIII, fibrinogen, fibrin D-dimer and high sensitive troponin T in a subset of enrolled participants.

We measure cellular immunity and immunosenescence biomarkers, such as CD4, CD8, CD28, and IL-6. These immunological characteristics and clinical features are analyzed using fluorescence-activated cell sorting analysis and enzyme-linked immunosorbent assays [26] for subsamples of the participants enrolled at each clinic. Chronic inflammation has been established to be responsible for the development of CVD [27], meaning that it may be a therapeutic target.