Original Article

Yonsei Med J 2024 Jun;65(6):363-370

https://doi.org/10.3349/ymj.2023.0455 pISSN: 0513-5796 · eISSN: 1976-2437

The Association between Social Support,

Metabolic Syndrome, and Incidence of
Cardio-Cerebrovascular Diseases in Older Adults:
The ARIRANG Study
Hae-Kweun Nam1, Sei-Jin Chang1,2, Chun-Bae Kim1,3, Kyoung Sook Jeong4, Sung-Kyung Kim4,
Dae Ryong Kang5, Yong Whi Jeong6, Hocheol Lee7, Bo Zhao8, and Sang-Baek Koh1,2
1
Department of Preventive Medicine, Yonsei University Wonju College of Medicine, Wonju;
2
Institute of Occupational and Environmental Medicine, Yonsei University Wonju College of Medicine, Wonju;
3
Hongcheon County Hypertension and Diabetes Registration and Education Center, Hongcheon;
4
Department of Occupational and Environmental Medicine, Yonsei University Wonju College of Medicine, Wonju;
5
Department of Precision Medicine, Yonsei University Wonju College of Medicine, Wonju;
6
Department of Medical Informatics and Biostatistics, Graduate School, Yonsei University, Wonju;
7
Department of Health Administration, College of Software and Digital Healthcare Convergence, Yonsei University, Wonju;
8
Department of Health Administration, Graduate School, Yonsei University, Wonju, Korea.

Purpose: We investigated the association between social support, metabolic syndrome, and incident cardio-cerebrovascular dis-
ease (CCVD) in rural Koreans aged ≥50 years.
Materials and Methods: We conducted a prospective study using the Korean Genome and Epidemiology Study on Atherosclerosis
Risk of Rural Areas in the Korean General Population (KoGES-ARIRANG) dataset. From the baseline of 5169 adults, 1682 partici-
pants were finally included according to the exclusion criteria. For outcomes, myocardial infarction, angina, and stroke were in-
cluded. For independent variables, the social support score and metabolic syndrome were used. Descriptive statistics and multi-
variate logistic regression were performed to investigate the association among the variables. Paired t-test was conducted to
analyze the longitudinal variation of social support scores.
Results: During the 6.37 years of median follow-up, 137 participants developed CCVD. The adjusted odds ratio (aOR) of metabol-
ic syndrome with persistently high social support was 2.175 [95% confidence interval (CI): 1.479–3.119]. The aOR of metabolic
syndrome with persistently low social support was 2.494 (95%CI: 1.141–5.452). The longitudinal variation of the social support
score of persistently high social support group was increased significantly by 4.26±26.32. The score of the persistently low social
support group was decreased by 1.34±16.87 with no statistical significance.
Conclusion: The presence of metabolic syndrome increases the likelihood of developing onset CCVD. Within the metabolic syn-
drome positive group, when social support was persistently low, the cohort developed more cardio-cerebrovascular disease com-
pared to the persistently higher social support group. The social support score of the persistently low social support group could
be improved through proper intervention. To prevent CCVD, metabolic syndrome components and low social support should be
improved in the study participants.

Key Words: ‌Incident cardio-cerebrovascular disease, social support, metabolic syndrome, public health

Received: November 8, 2023 Revised: December 12, 2023 Accepted: December 20, 2023 Published online: April 26, 2024
Co-corresponding authors: Hae-Kweun Nam, MSc, Department of Preventive Medicine, Yonsei University Wonju College of Medicine, 20 Ilsan-ro, Wonju 26426, Korea.
E-mail: namhk@yonsei.ac.kr and
Sang-Baek Koh, MD, PhD, Department of Preventive Medicine, Yonsei University Wonju College of Medicine, 20 Ilsan-ro, Wonju 26426, Korea.
E-mail: kohhj@yonsei.ac.kr
•The authors have no potential conflicts of interest to disclose.
© Copyright: Yonsei University College of Medicine 2024
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (https://creativecommons.org/licenses/by-nc/4.0)
which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

www.eymj.org 363
 Social Support, Metabolic Syndrome, and Cardio-Cerebrovascula Disease

INTRODUCTION MATERIALS AND METHODS

Recently, the World Health Organization (WHO), reported Data and participants
that heart disease and stroke were ranked as the first and sec- We used the Korean Genome and Epidemiology Study on Ath-
ond causes of death worldwide, respectively, in both male erosclerosis Risk of Rural Areas in the Korean General Popula-
and female.1 In the United States, it was predicted that 20% of tion (KoGES-ARIRANG) data. KoGES-ARIRANG is a prospec-
the population will be aged ≥65 years by 2030, and cardiovas- tive cohort project in Wonju and Pyeongchang to observe the
cular disease (CVD) will account for 40% of all deaths.2 A re- prevalence, incidence, and risk factors of chronic diseases. Mini-
cent meta-analysis demonstrated that prevalence stroke in- mal demographic migration in the region made it ideal for a
creased with age.3 long-term observational study.18 The cohort was initiated in
In the United States, socio-structural factors, such as seden- November 2005 with 5169 adult participants aged ≥40 years
tary behavior, longer work, and commuting hours, and a lack (Supplementary Table 1, only online). The baseline was re-
of leisure time resulting in less exercise and unhealthy eating garded as 2005–2011 in the current study. Next, baseline par-
behaviors, including intake of a high-calorie diet, saturated fats, ticipants were gradually excluded according to the study de-
and sugars, caused the development of CVD and metabolic sign and exclusion criteria for further analysis. The exclusion
complications (i.e., metabolic syndrome, diabetes mellitus, and criteria were as follows: 1) age <50 years (n=1520), 2) lost to fol-
hypertension).4,5 In 2021, obesity, hypertension, diabetes, and low-up (n=895); 3) CCVD patients at baseline (n=201); 4) ob-
dyslipidemia were listed as pivotal risk factors of CVD in South served social support score less than two times at baseline
Korea.6 The aforementioned complications are the compo- (n=850), and 5) missing variables (n=21). Finally, 1682 partici-
nents of metabolic syndrome (MetS). A prospective cohort pants were included in the current study, and we conducted a
study conducted in East Asia concluded that MetS was associ- prospective study using the data (Fig. 1). The institutional re-
ated with an increased risk for heart disease and stroke.7 Also, view board of Wonju Christian Hospital approved the study
a prospective cohort study conducted in Scandinavia reported (IRB No. CR105024).
that MetS predicted the incidence of heart disease and stroke.8
The result of the ARIRANG cohort study revealed that the com- Outcome variable
ponents of MetS, especially waist circumference, were the The cohort participants were followed up (median follow-up:
most associated factors of cardiovascular parameters.9 6.37 years), and the diagnosed CCVD were coded using the
Furthermore, the WHO reported that social determinants of 10th revision of International Classification of Diseases (ICD-
health (SDH) account for 30%–55% of health outcomes.10 Ad- 10) code and entered into electronic health records. Myocar-
ditionally, according to the OECD’s Better Life Index, the num- dial infarction was identified as a case in which the ICD-10
ber of social support networks in South Korea was the lowest code was I21–I25 and lab codes of B2640, C3941, C3942, CY277,
among 13 countries. When the data was stratified by age (≥50 CY278, CY279 or procedure codes, such as surgery, percutane-
years), only 60.9% answered to having a social support network, ous coronary intervention, and angiography, were included.
which was 26.2%p lower than the OECD average of 87.1%.11,12 Angina was defined as a case where coronary recanalization
Epidemiological studies have also reported the association be- was performed under the ICD-10 code I20. Stroke was defined
tween social indexes and disease outcomes. The Atherosclerosis as a case where CT or MRI images were performed under the
Risk in Communities Study in the United States reported that ICD-10 code I60–I69 or G45–46.
having a small social network is associated with an increased
risk of incident stroke.13 In Japan, among a cohort of 44152 par- Independent variable
ticipants, those with low social support showed a higher risk of Due to the nature of the current study design, two independent
stroke mortality in male.14 A comprehensive review study also variables were included. First, social support score was collect-
concluded that low social support increased coronary heart ed using the Medical Outcomes Study-Social Support Survey
disease (CHD) events.15 Contrarily, in South Korea, the main (MOS-SSS).19,20 The survey consists of 19 items with a 5-point
concern among the researchers was the state of social support Likert scale. The survey contains four sub-scales: 1) emotional/
status following the occurrence of CHD and stroke.16,17 informational support, 2) tangible support, 3) affectionate sup-
Based on existing research, we hypothesized that cardio- port, and 4) positive social interactions. The MOS-SSS’s Cron-
cerebrovascular disease (CCVD) incidence would be higher in bach a value was 0.962. The measured score was then trans-
participants previously categorized as MetS with persistently formed into a maximum score of 100 based on the equation
low social support compared to participants who are classified provided below:
as MetS-free with persistently high social support. Therefore,
we aimed to investigate the association among social support, MOS-SSS equation20
metabolic syndrome, and incidence of CCVD in rural Koreans (Meansubscalescore–1)
aged ≥50 years. 100×
(5–1=4)

364 https://doi.org/10.3349/ymj.2023.0455
Hae-Kweun Nam, et al.

(2005–2008. 2) Individuals of KoGES-ARIRANG (n=5169)

Excluded
Age <50 years old and CCVD history (n=1520)

(2008. 3–2011. 2) Individuals of KoGES-ARIRANG in older adult (n=3649)

Excluded
CCVD patients at baseline (n=201)

(2008. 3–2011. 2) Subjescts without cadio-cerebrovascular disease (n=3408)

Excluded
- Lost follow-up (n=895)
- Social
‌ support measurement less than 2 times
(n=850)
- Missing varibles (n=21)

Subjects included in this study (n=1682)

Fig. 1. Flowchart of study design. KoGES-ARIRANG, Korean Genome and Epidemiology Study on Atherosclerosis Risk of Rural Areas in the Korean Gen-
eral Population; CCVD, cardio-cerebrovascular disease.

The measured social support score was divided into quartiles, age. As explained in the previous section, we investigated the
and participants who were assigned to the lowest quartile in association among baseline social support level with metabolic
both the first and second wave were considered as “social syndrome and the incidence of new-onset CCVD. The baseline
support persistently low.” Thus, the social support score was di- presence of CCVD was excluded for further analysis. Multivari-
chotomized into “persistently low” and “not persistently low,” ate logistic regression analysis was implemented to examine
namely “higher” social support. the association among the variables. Based on the re-catego-
Second, metabolic syndrome was assessed using the Na- rization of independent variables, participants with high social
tional Cholesterol Education Program Adult Treatment Panel support with no metabolic syndrome were set as a reference
(NCEP-ATP III) criteria. Based on the NCEP-ATP III criteria, group. The results of the multivariate logistic regression were
central obesity was screened using a cutoff of waist circumfer- shown as odds ratios (OR) with 95% confidence intervals (CI).
ence (male: >90 cm; female: >85 cm). High blood pressure was To attenuate potential confounders, sex, age, smoking and
determined if systolic blood pressure (SBP) was ≥130 mm Hg drinking condition, frequent exercise, marital status, education-
or diastolic blood pressure (DBP) was ≥85 mm Hg. Hypergly- al levels, and low density lipoprotein (LDL) cholesterol were ad-
cemia was classified by fasting blood glucose concentration of justed. Paired t-test was used to analyze longitudinal variation
≥100 mg/dL upon 8 hours of fasting. Hypertriglyceridemia was of social support score, which was repeatedly measured in
screened by blood triglycerides ≥150 mg/dL. High density li- identical participants during the baseline and follow-up. Sta-
poprotein (HDL) cholesterolemia was categorized by blood tistical analysis was performed using SPSS version 26.0 (IBM
HDL cholesterol (male: <40 mg/dL; female: <50 mg/dL). Upon Corp., Armonk, NY, USA). P-values<0.05 were considered to
screening of the five metabolic syndrome components, meta- be statistically significant.
bolic syndrome was diagnosed if three out of the five criteria
were satisfied. RESULTS
Finally, social support and metabolic syndrome variables
were re-categorized into four groups: 1) no metabolic syndrome A total of 1682 participants were included in the current study.
with high social support, 2) no metabolic syndrome with per- During the 6.37 years of median follow-up, 137 participants
sistently low social support, 3) presence of metabolic syndrome developed CCVD. Table 1 presents the baseline characteristics
with high social support, and 4) diagnosed metabolic syn- of study participants according to incident CCVD. The propor-
drome with persistently low social support. tion of male participants did not differ between the non-CCVD
and CCVD groups (p=0.553). The CCVD group was older than
Statistical analysis the non-CCVD group (61.31±5.92 years vs. 58.96±5.83 years,
To compare continuous variables, the Student’s t-test was used p<0.001). Marital status (p=0.870) and education level (p=
and the results were presented as mean and standard devia- 0.521) did not affect CCVD occurrence. The proportion of cur-
tion. To compare categorical variables, the χ2 test was per- rent smokers, drinkers, and regular exercisers did not change
formed and the results were shown as frequency and percent- incident CCVD. Body mass index (BMI) was higher in the

https://doi.org/10.3349/ymj.2023.0455 365
 Social Support, Metabolic Syndrome, and Cardio-Cerebrovascula Disease

Table 1. Baseline Characteristics of Study Participants according to syndrome with higher social support group was significantly
Incident Cardio-Cerebrovascular Disease (n=1682) increased compared to the reference group (OR: 2.224, 95%CI:
Cardio-cerebrovascula disease 1.520–3.254). The metabolic syndrome with persistently low
Non-CCVD, CCVD, social support group indicated the highest odds ratio (OR:
p value
n=1545 (91.9) n=137 (8.1) 2.548, 95%CI: 1.194–5.440). The following are the results of the
Sex, male 670 (43.4) 63 (46.0) 0.553 adjusted ORs of CCVD incidence by social support and meta-
Age (yr) 58.96±5.83 61.31±5.92 <0.001 bolic syndrome in the study participants. The CCVD incidence
Marital status 0.870 in the previously no metabolic syndrome with persistently low
Married 1353 (92.0) 118 (8.0) social support group was 1.417-fold higher than the reference
Separated/divorced/widowed 161 (91.0) 16 (9.0) group with no statistical significance (OR: 1.417, 95%CI: 0.617–
Single 4 (100) 0 (0) 3.252). The OR of subjects with metabolic syndrome and higher
Others 27 (90.0) 3 (10.0) social support was 2.136 (OR: 2.136, 95%CI: 1.456–3.134). Study
Education level 0.521 participants with persistently low social support and metabolic
≤Elementary school 862 (91.2) 83 (8.8) syndrome showed the highest OR of 2.445 (OR: 2.445, 95%CI:
Middle school 265 (91.4) 25 (8.6) 1.136–5.264).
High school 279 (93.6) 19 (6.4) Table 3 demonstrates the variation of MOS-SSS and subscale
≥College/University 139 (93.3) 10 (6.7) in groups with persistently higher and low social support. The
Current smoker 246 (15.9) 25 (18.2) 0.441 mean MOS-SSS of the persistently higher social support group
Current drinker 619 (40.1) 58 (42.3) 0.171 of time 1 was 82.13±19.07. When identical participants were
Regular exercise 501 (32.4) 38 (27.7) 0.260 measured in time 2 (follow-up), MOS-SSS was increased by
BMI (kg/m2) 24.59±3.11 25.65±3.27 <0.001 4.26±26.32 points (p<0.001). All subscales were also increased
Waist circumference (cm) 86.48±47.27 87.82±8.57 <0.740 with statistical significance (p<0.001, respectively). The mean
SBP (mm Hg) 132.52±18.16 136.09±21.08 <0.030 MOS-SSS of the persistently low social support group of time
DBP (mm Hg) 82.98±10.99 85.45±12.08 <0.012 1 was 44.70±15.06. The assessed MOS-SSS in follow-up partic-
Fasting plasma glucose (mg/dL) 96.56±18.24 104.76±37.08 <0.001
ipants decreased by 1.34±16.87 with no statistical significance
(p=0.324). Emotional support and tangible support scores were
Triglyceride (mg/dL) 155.25±122.82 167.05±89.27 0.272
decreased, and affectionate support and positive social interac-
HDL cholesterol (mg/dL) 45.82±11.02 43.82±10.33 0.041
tion scores were increased; nevertheless, in the persistently low
LDL cholesterol (mg/dL) 120.47±33.22 122.12±32.61 0.575
social support group, all subscale variations were not statistically
Metabolic syndrome 641 (41.5) 83 (60.6) <0.001
significant (Fig. 2).
BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pres-
sure; HDL, high density lipoprotein; LDL, low density lipoprotein.
The results are presented as mean±SD or frequency (%).
DISCUSSION
CCVD group (25.65±3.27 kg/m2 vs. 24.59±3.11 kg/m2, p<0.001).
SBP was higher in the CCVD group (136.09±21.08 mm Hg vs. The current study found that, in Korean older adults aged ≥50
132.52±18.16 mm Hg, p<0.030). DBP was higher in the CCVD years, the presence of MetS combined with low social support
group (85.45±12.08 mm Hg vs. 82.98±10.99 mm Hg, p<0.012). level was associated with high CCVD incidence. CCVD inci-
Fasting plasma glucose was higher in the CCVD group (104.76± dence was higher in the participant group with persistently low
37.08 mg/dL vs. 96.56±18.24 mg/dL , p<0.001). HDL choles- social support and presence of metabolic syndrome compared
terol was lower in the CCVD group (43.82±10.33 mg/dL vs. to the group with higher social support and no metabolic syn-
45.82±11.02 mg/dL, p=0.041). No difference was found in waist drome. The CCVD incidence of the participants with identified
circumference, triglyceride, and LDL cholesterol between the metabolic syndrome and higher social support was 2.2-fold
non-CCVD and CCVD groups. The prevalence of metabolic higher compared to the reference group. In participants with
syndrome was higher in the CCVD group (60.6% vs. 41.5%, metabolic syndrome, especially, the CCVD incidence was 2.5-
p<0.001). fold when the subject was categorized into persistently low so-
Table 2 provides the crude and adjusted ORs of CCVD inci- cial support compared to the reference group. The analysis of
dence according to social support and metabolic syndrome in longitudinal variation of social support score and its subscale
the study participants. The previously no metabolic syndrome revealed that total and subscale scores were improved signifi-
with a higher social support group was set as the reference cantly in the persistently higher social support group, whereas
group. The CCVD incidence in the previously no metabolic the overall and subscale scores of the persistently low social
syndrome with persistently low social support group was high- support group did not increase.
er compared to the reference group with no statistical signifi- Existing literatures proved the linkage between MetS and
cance (OR: 1.536, 95%CI: 0.672–3.510). The OR of the metabolic CCVD, and the results of the current analysis were consistent

366 https://doi.org/10.3349/ymj.2023.0455
Hae-Kweun Nam, et al.

Table 2. Multivariate Logistic Regression Analysis among Social Support, Metabolic Syndrome, and Incident Cardio-Cerebrovascular Disease with Ad-
justments (n=1682)
Cardio-cerebrovascula disease
Metabolic syndrome
CCVD Non-CCVD
OR 95% CI p value OR 95% CI p value
Model 1 (crude OR)
Social support
Persistently low 2.548 1.194–5.440 <0.001 1.536 0.672–3.510 0.309
Persistently higher 2.224 1.520–3.254 <0.001 Reference
Model 2 (adjusted OR)
Social support
Persistently low 2.445 1.136–5.264 0.022 1.417 0.617–3.252 0.412
Persistently higher 2.136 1.456–3.134 <0.001 Reference
Model 3 (adjusted OR)
Social support
Persistently low 2.494 1.141–5.452 0.022 1.478 0.639–3.416 0.639
Persistently higher 2.175 1.479–3.199 <0.001 Reference
OR, odds ratio; CI, confidence interval; LDL, low density lipoprotein.
Model 2 was adjusted for sex and age; Model 3 was adjusted for sex, age, marital status, education level, drinking and smoking status, exercise status, and LDL
cholesterol.

Table 3. Longitudinal Variation of Social Support Scores in Groups with Persistently Higher and Low Social Support
MOS-SSS and subscales Time 1 Time 2 Mean difference p value
Persistently higher social support (n=1525) 82.13±19.07 86.39±18.55 +4.26±26.32 <0.001
Emotional support 33.82±6.55 35.40±6.43 +1.58±9.01 <0.001
Tangible support 17.48±3.25 17.90±3.17 +0.41±4.20 <0.001
Affectionate support 13.02±2.56 13.54±2.43 +0.51±3.54 <0.001
Positive social interaction 13.03±2.64 13.54±2.43 +0.50±3.58 <0.001
Persistently low social support (n=157) 44.70±15.06 43.36±14.72 -1.34±16.87 0.324
Emotional support 21.87±5.60 21.17±5.28 -0.69±6.29 0.169
Tangible support 12.18±3.63 11.70±3.21 -0.47±4.00 0.137
Affectionate support 8.34±2.63 8.36±2.66 +0.19±3.20 0.940
Positive social interaction 8.20±2.63 8.36±2.66 +0.16±3.00 0.507
MOS-SSS, Medical Outcome Study-Social Support Score.
The results are presented as mean±SD.

SSL-MetS (+)

SSH-MetS (+)

SSL-MetS (-)

SSH-MetS (-)

0 1 2 3 4 5 6
Fig. 2. Adjusted odds ratios of cardio-cerebrovascular disease according to social support and metabolic syndrome.

https://doi.org/10.3349/ymj.2023.0455 367
 Social Support, Metabolic Syndrome, and Cardio-Cerebrovascula Disease

with previous studies. A recent prospective cohort study con- individuals, eventually preventing the occurrence of disease;
ducted in Spain demonstrated that the hazard ratio (HR) of however, the stress-related model is beyond the scope of this
CCVD according to the MetS positive group was significantly research.29,30 The direct effect model of social support may ac-
higher compared to the non-MetS group [HR: 1.32 (95%CI: 1.01– count for the connection to disease outcome.28 The direct ef-
1.74)].21 A prospective cohort study performed in Taiwan showed fect model theorists argue that participating in social networks
that the HR of CCVD in MetS participants was significantly provides meaningful roles that gives self-esteem and purpose
higher compared to non-MetS participants. HR of CHD was 5.5 to life.31 Based on the aforementioned logic, the social control
(95%CI: 2.2–13.7) and stroke was 3.5 (95%CI: 1.9–6.5).7 Anoth- hypothesis was proposed. The social control hypothesis sug-
er prospective cohort study conducted in Iran demonstrated gested that social interaction promotes health by facilitating
that the HR of CVD in MetS participants was significantly high- healthier behaviors, such as smoking and drinking less, exercis-
er compared to non-MetS participants [HR: 1.66 (95%CI: 1.04– ing more, and eating healthy. This social-health behavior route
2.67)].22 Literature documented the association between MetS can provide positive social pressure on group members to im-
and CCVD incidence. When MetS, specifically, increased waist plement healthier habits. For example, in a married couple, a
circumference, fasting plasma glucose, blood pressure, and tri- spouse can advise the husband or wife to eat healthier or to
glycerides, and decreased HDL cholesterol is present, the oc- drink less alcohol.32,33 The result of our analysis is in line with
currence of CCVD was increased. the literature as social support is associated with marital sta-
The relationship between social support and CCVD has been tus (Supplementary Table 2, only online). Participants who
investigated in various studies. A prospective cohort study con- were separated, divorced, or widowed presented persistently
ducted in Australia revealed that in older adults, low social sup- low social support compared to the married group. This result
port resulted in a higher incidence of CVD [HR: 2.05 (95%CI: supports the social control hypothesis that spouses provide
1.31–3.21)] and low social support was associated with a higher positive mutual support to induce healthier behavior, poten-
stroke incidence [HR: 1.76 (95%CI: 1.17–2.65)].23 A cross-sec- tially leading to less occurrence of disease.33
tional study in older Australian adults suggested that social In South Korea, social support should be improved in older
support was inversely associated with CVD risk factors, includ- populations ≥50 years of age. According to the National Assem-
ing BMI and depressive symptoms.24 A longitudinal study in bly of South Korea, the mean retirement age of South Koreans
Swedish male showed that social support and social networks was 49.3 years as of 2022. Moreover, an analysis of OECD data
predicted new CHD events.25 revealed that the social network indicator of older South Kore-
Additionally, the current study performed a longitudinal vari- an adults aged ≥50 years was 60.9%, which is the lowest among
ation analysis of social support scores in the ARIRANG cohort. OECD countries.11,12 Presumably, around this age, social and
The result showed that the mean baseline social support score economic participation decreases, resulting in a lower social
of persistently low social support group was prominently lower network.11 Public health authorities may institutionalize par-
compared to the higher social support group. When persis- ticipation of social activities to improve social support as, at
tently social support was categorized as “higher,” the mean dif- this point, healthcare professionals can only recommend the
ference of the score improved significantly both the overall social involvement of potential patients. The United Kingdom
score and subscale scores. However, if social support was di- launched a new practice of “social prescribing” to improve so-
chotomized into “persistently low,” the overall score did not cial participation to prevent anxiety, loneliness, depression, and
vary, denoting that when social support was persistently low, noncommunicable diseases (NCDs) in the general population.34
there is less of a possibility for natural improvement of social According to the WHO, “social prescribing is a means for health-
support. In older German adults who were aged 75 years or care workers to connect people to a range of non-clinical servic-
older, social support was gradually decreased longitudinally.26 es in the community in order to improve health and wellbeing.”
Similarly, in older French adults (mean age of 72.5 years), so- In the social prescribing scheme, participants are introduced
cial support satisfaction did not vary post-follow-up.27 The lit- to various local social activities, including volunteering, art ac-
erature supports that it is not easy for older adults to improve tivities, group learning, gardening, friend-making, cooking,
social support without proper intervention. In the case of par- eating a healthy diet, and sports.35 As of 2022, social prescrib-
ticipants in our analysis, the persistently low social support ing has been proliferated to 17 countries around the world.36
group could be invited for programs to experience improved There are several strengths of the current study. We incor-
social support. porated social support, which is a widely used index of SDH,
Uchino28 suggest two theoretical models for explaining the in patients with metabolic syndrome and further, progress to
connection between social support and health outcomes. The CCVD. As accentuated by the WHO, it is favorable to include
first model is the stress-buffering model and the second is di- useful social measurements when collecting population health
rect effect model of support. The first model emphasizes the data to prevent increase of NCDs.10 We also analyzed the lon-
role of stress between social support and disease. The main gitudinal variation of social support with its subscales. It is im-
concept is that a positive social support may lessen distress in portant to note that the persistently low social support group

368 https://doi.org/10.3349/ymj.2023.0455
Hae-Kweun Nam, et al.

may not be able to improve social support levels without prop- Chun-Bae Kim https://orcid.org/0000-0002-1979-6833
er intervention. Blood samples from participants who had Kyoung Sook Jeong https://orcid.org/0000-0002-6897-8289
Sung-Kyung Kim https://orcid.org/0000-0002-2742-6410
fasted for 8 hours were collected to prevent interference of di-
Dae Ryong Kang https://orcid.org/0000-0002-8792-9730
etary effects to variables of interest. The current study had few Yong Whi Jeong https://orcid.org/0000-0003-4746-5764
limitations. In terms of social support measurements, there are Hocheol Lee https://orcid.org/0000-0003-1467-8843
a wide variety of scales that are used in different disciplines, Bo Zhao https://orcid.org/0000-0002-6780-1633
countries, and researchers. To obtain a standardized measure- Sang-Baek Koh https://orcid.org/0000-0001-5609-6521
ment of social support in the population, public health authori-
ties may designate a social support scales to fit in the local en- REFERENCES
vironment. Nevertheless, the ARIRANG cohort could use a 1. World Health Organization. Leading causes of death and disabili-
social support scale from a medical outcome study, which is a ty: a visual summary of global and regional trends 2000-2019 [In-
validated and commonly used measurement.20 Since the ARI- ternet] [accessed on 2023 September 13]. Available at: https://
RANG cohort data was collected in rural parts of South Korea, www.who.int/data/stories/leading-causes-of-death-and-disabil-
the result of the study cannot be generalized to the entire popu- ity-2000-2019-a-visual-summary.
2. Heidenreich PA, Trogdon JG, Khavjou OA, Butler J, Dracup K, Eze-
lation. Nevertheless, the results can still represent a rural popu- kowitz MD, et al. Forecasting the future of cardiovascular disease
lation of South Korea. Future studies should include more so- in the United States: a policy statement from the American Heart
cial health measurements (e.g. social capital, loneliness, and Association. Circulation 2011;123:933-44.
social isolation) to obtain more sophisticated results.37-39 3. Rajati F, Rajati M, Rasulehvandi R, Kazeminia M. Prevalence of
In conclusion, onset CCVD increased with the presence of stroke in the elderly: a systematic review and meta-analysis. Inter-
discip Neurosurg 2023;32:101746.
metabolic syndrome. Within the metabolic syndrome group,
4. Fox CS, Coady S, Sorlie PD, Levy D, Meigs JB, D’Agostino RB Sr, et
when social support was persistently low, the cohort developed al. Trends in cardiovascular complications of diabetes. JAMA 2004;
higher CCVD compared to the persistently higher social sup- 292:2495-9.
port group. Social support of the persistently low social support 5. US Preventive Services Task Force; Curry SJ, Krist AH, Owens DK,
group could be improved by applying appropriate intervention, Barry MJ, Caughey AB, Davidson KW, et al. Risk assessment for
cardiovascular disease with nontraditional risk factors: US preven-
such as social prescribing. In order to prevent CCVD, both met-
tive services task force recommendation statement. JAMA 2018;
abolic syndrome components and social support should be im- 320:272-80.
proved. 6. Wang JM, Kim BO, Bae JW, Oh DJ. Implementation of national
health policy for the prevention and control of cardiovascular dis-
ease in South Korea: regional-local cardio-cerebrovascular center
ACKNOWLEDGEMENTS and nationwide registry. Korean Circ J 2021;51:383-98.
7. Chien KL, Hsu HC, Sung FC, Su TC, Chen MF, Lee YT. Metabolic
This study was supported by the Korea Centers for Disease syndrome as a risk factor for coronary heart disease and stroke: an
Control and Prevention (2005-E71013-00, 2006-E71002-00, 11-year prospective cohort in Taiwan community. Atherosclerosis
2007-E71013-00, 2008-E71004-00, 2009-E71006-00, and 2010- 2007;194:214-21.
E71003-00). This research was supported by “Regional Innova- 8. Qiao Q, Laatikainen T, Zethelius B, Stegmayr B, Eliasson M, Jousi-
tion Strategy (RIS)” through the National Research Foundation lahti P, et al. Comparison of definitions of metabolic syndrome in
relation to the risk of developing stroke and coronary heart dis-
of Korea (NRF) funded by the Ministry of Education (MOE) in ease in Finnish and Swedish cohorts. Stroke 2009;40:337-43.
2023 (2022RIS-005). 9. Ahn MS, Kim JY, Youn YJ, Kim SY, Koh SB, Lee K, et al. Cardiovas-
cular parameters correlated with metabolic syndrome in a rural
AUTHOR CONTRIBUTIONS community cohort of Korea: the ARIRANG study. J Korean Med Sci
2010;25:1045-52.
10. World Health Organization. Social determinants of health [Inter-
Conceptualization: Sang-Baek Koh. Data curation: Hae-Kweun Nam.
net] [accessed on 2023 September 13]. Available at: https://www.
Formal analysis: Sang-Baek Koh, Dae Ryong Kang, Yong Whi Jeong,
who.int/health-topics/social-determinants-of-health#tab=tab_3.
and Hae-Kweun Nam. Funding acquisition: Sang-Baek Koh. Investi-
11. Park SN. Aging and economic and social participation of the el-
gation: Sang-Baek Koh. Methodology: Hae-Kweun Nam. Project ad-
derly. KOSTAT: 2019. Internet] [accessed on 2023 July 20]. Available
ministration: Sang-Baek Koh. Resources: Sang-Baek Koh. Software:
at: https://kostat.go.kr/board.es?mid=a90102010100&bid=11918.
Hae-Kweun Nam. Supervision: Sei-Jin Chang, Chun-Bae Kim, Kyoung
12. OECD. Society at a glance 2016: OECD social indicators. Paris:
Sook Jeong, Sung-Kyung Kim, Hocheol Lee, and Bo Zhao. Validation:
OECD Publishing; 2016.
Sang-Baek Koh. Visualization: Hae-Kweun Nam. Writing—original
13. Nagayoshi M, Everson-Rose SA, Iso H, Mosley TH Jr, Rose KM, Lut-
draft: Hae-Kweun Nam. Writing—review & editing: Hae-Kweun Nam.
sey PL. Social network, social support, and risk of incident stroke:
Approval of final manuscript: all authors.
Atherosclerosis Risk in Communities study. Stroke 2014;45:2868-
73.
ORCID iDs 14. Ikeda A, Iso H, Kawachi I, Yamagishi K, Inoue M, Tsugane S;
JPHC Study Group. Social support and stroke and coronary heart
Hae-Kweun Nam https://orcid.org/0000-0002-3353-6990 disease: the JPHC study cohorts II. Stroke 2008;39:768-75.
Sei-Jin Chang https://orcid.org/0000-0001-9347-3592 15. Lett HS, Blumenthal JA, Babyak MA, Strauman TJ, Robins C, Sher-

https://doi.org/10.3349/ymj.2023.0455 369
 Social Support, Metabolic Syndrome, and Cardio-Cerebrovascula Disease

wood A. Social support and coronary heart disease: epidemiologic 1667-76.

evidence and implications for treatment. Psychosom Med 2005; 27. Ferrand C, Martinent G, Bailly N, Giraudeau C, Pothier K, Agli O.
67:869-78. Change of depressive symptoms in a French nine-year longitudinal
16. Kim JM, Kim JW, Kim YJ, Kang HJ, Kim SW, Kim JT, et al. Associa- study of aging: gender differences and relationships between social
tions of systemic inflammation and social support with suicidal support, health status and depressive symptoms. Arch Gerontol
ideation in patients with acute coronary syndrome and stroke. J Geriatr 2020;89:104059.
Affect Disord 2023;323:101-7. 28. Uchino BN. Social support and physical health: understanding
17. Lee Y, Won M. Mediating effects of rehabilitation motivation be- the health consequences of relationships. New Haven (CT): Yale
tween social support and health-related quality of life among pa- University Press; 2004.
tients with stroke. Int J Environ Res Public Health 2022;19:15274. 29. Kang MG, Koh SB, Cha BS, Park JK, Baik SK, Chang SJ. Job stress
18. Jo H, Lee J, Lee S, Lee H, Ahn YS, Koh SB. The longitudinal effect and cardiovascular risk factors in male workers. Prev Med 2005;
of leisure time physical activity on reduced depressive symptoms: 40:583-8.
the ARIRANG study. J Affect Disord 2021;282:1220-5. 30. Chang SJ, Koh SB, Kang MG, Cha BS, Park JK, Hyun SJ, et al. [Epi-
19. Tarlov AR, Ware JE Jr, Greenfield S, Nelson EC, Perrin E, Zubkoff demiology of psychosocial distress in Korean employees]. J Prev
M. The medical outcomes study. An application of methods for Med Public Health 2005;38:25-37. Korean
monitoring the results of medical care. JAMA 1989;262:925-30. 31. Thoits PA. Social support as coping assistance. J Consult Clin Psy-
20. Sherbourne CD, Stewart AL. The MOS social support survey. Soc chol 1986;54:416-23.
Sci Med 1991;32:705-14. 32. Umberson D. Family status and health behaviors: social control as a
21. Guembe MJ, Fernandez-Lazaro CI, Sayon-Orea C, Toledo E, More- dimension of social integration. J Health Soc Behav 1987;28:306-19.
no-Iribas C; RIVANA Study Investigators. Risk for cardiovascular 33. Lewis MA, Rook KS. Social control in personal relationships: im-
disease associated with metabolic syndrome and its components: pact on health behaviors and psychological distress. Health Psy-
a 13-year prospective study in the RIVANA cohort. Cardiovasc Di- chol 1999;18:63-71.
abetol 2020;19:195. 34. National Health Service. Social prescribing [Internet] [accessed
22. Mansourian M, Babahajiani M, Jafari-Koshki T, Roohafza H, Sade- 2023 September 18]. Available at: https://www.england.nhs.uk/
ghi M, Sarrafzadegan N. Metabolic syndrome components and personalisedcare/social-prescribing.
long-term incidence of cardiovascular disease in Eastern Mediter- 35. The King’s Fund. What is social prescribing? [Internet] [accessed
ranean region: a 13-year population-based cohort study. Metab on 2023 September 18]. Available at: https://www.kingsfund.org.
Syndr Relat Disord 2019;17:362-6. uk/publications/social-prescribing?gclid=EAIaIQobChMIhrja98
23. Freak-Poli R, Ryan J, Neumann JT, Tonkin A, Reid CM, Woods RL, uzgQMVWmAPAh3AegiGEAAYAiAAEgLsCPD_BwE.
et al. Social isolation, social support and loneliness as predictors 36. Morse DF, Sandhu S, Mulligan K, Tierney S, Polley M, Chiva Giur-
of cardiovascular disease incidence and mortality. BMC Geriatr ca B, et al. Global developments in social prescribing. BMJ Glob
2021;21:711. Health 2022;7:e008524.
24. Hu J, Fitzgerald SM, Owen AJ, Ryan J, Joyce J, Chowdhury E, et al. 37. Yazawa A, Inoue Y, Yamamoto T, Watanabe C, Tu R, Kawachi I.
Social isolation, social support, loneliness and cardiovascular dis- Can social support buffer the association between loneliness and
ease risk factors: a cross-sectional study among older adults. Int J hypertension? A cross-sectional study in rural China. PLoS One
Geriatr Psychiatry 2021;36:1795-809. 2022;17:e0264086.
25. Orth-Gomér K, Rosengren A, Wilhelmsen L. Lack of social support 38. Nakagomi A, Shiba K, Kondo K, Kawachi I. Can social capital mod-
and incidence of coronary heart disease in middle-aged Swedish erate the impact of widowhood on depressive symptoms? A fixed-
men. Psychosom Med 1993;55:37-43. effects longitudinal analysis. Aging Ment Health 2021;25:1811-20.
26. Hajek A, Brettschneider C, Mallon T, Kaduszkiewicz H, Oey A, 39. Smith RW, Holt-Lunstad J, Kawachi I. Benchmarking social isola-
Wiese B, et al. Social support and health-related quality of life tion, loneliness, and smoking: challenges and opportunities for
among the oldest old—longitudinal evidence from the multicenter public health. Am J Epidemiol 2023;192:1238-42.
prospective AgeCoDe-AgeQualiDe study. Qual Life Res 2022;31:

370 https://doi.org/10.3349/ymj.2023.0455