Skip to main content
NCBI home page
European Journal of Ageing logoLink to European Journal of Ageing
. 2020 Feb 20;17(4):531–546. doi: 10.1007/s10433-020-00561-y

Positive and negative aspects of social relations and low-grade inflammation in Copenhagen Aging and Midlife Biobank

Charlotte Juul Nilsson 1,4,✉, Signe Nørgaard 1, Else Foverskov 1,4, Helle Bruunsgaard 2, Per Kragh Andersen 3, Rikke Lund 1,4
PMCID: PMC7752933  PMID: 33381004

Abstract

The association between social relations and health outcomes is well described, but pathways are relatively poorly understood. Inflammation has been suggested as a potential physiological pathway, linking social relations to adverse health outcomes. However, previous studies have shown ambiguous results and have for the vast majority been based on studies small in sample size. The aim of the present study is to examine the association between comprehensive measures of structural and positive as well as negative functional aspects of social relations, across four relational domains—partner/spouse, children, other family and friends, and the level of systemic low-grade inflammation in a large population-based middle-aged cohort and to examine variation by gender and socioeconomic position in these associations. The study comprised of 5576 participants in the Copenhagen Aging and Midlife Biobank. The inflammatory biomarkers collected in late midlife included C-reactive protein, Interleukin-6, and TNF-alpha. Multiple linear regression models were implemented to explore associations between social relations and inflammatory measures controlling for gender, age, socioeconomic position, marital status, early major lifeevents and morbidity. Results show weak and ambiguous associations in all analyses. There were no strong indications of interaction with socioeconomic position. Concluding cautiously, men appear to be more vulnerable toward living alone and low contact frequency with family compared to women as regards high level of low-grade inflammation. In conclusion, this large-scale population-based study among middle-aged men and women showed no association between social relations and low-grade inflammation.

Electronic supplementary material

The online version of this article (10.1007/s10433-020-00561-y) contains supplementary material, which is available to authorized users.

Keywords: Low-grade inflammation, Social relations, Social strain, Gender, Socioeconomic position

Introduction

The link between social relations and health is well established (Holt-Lunstad 2018; Holt-Lunstad et al. 2010; Uchino et al. 2018; Yang et al. 2016), and one of the suggested mechanisms behind this link is chronic low-grade inflammation (Kiecolt-Glaser et al. 2010; Uchino et al. 2018; Yang et al. 2016), which is involved in several chronic public health diseases such as diabetes, cardiovascular disease, arthritis, dementia, as well as aging (Cevenini et al. 2013; Eisenberger et al. 2017; Ershler and Keller 2000; Lara et al. 2015; Wang et al. 2013; Yang et al. 2016). Inflammatory processes are partly driven by the nervous system, body fat, life style, and comorbidity in addition to the immune system, thus allowing environmental features and perceptions of for example inadequate social interaction and support to impact inflammatory activity by a state of chronic stress (Eisenberger et al. 2017).

A recent review and meta-analysis on social integration, social support and inflammatory cytokines present weak significant associations, indicating that higher levels of social integration, perceived and received social support are associated with lower levels of inflammatory cytokines. Ambiguity in findings is also highlighted, and more studies on the association between social relations and inflammatory markers have been called for to strengthen the evidence (Uchino et al. 2018). Structural aspects of social relations are often defined by their quantitative characteristics such as contact frequency, social integration and network diversity, whereas functional aspects of social relations include qualitative characteristics such as access to emotional or instrumental support, but also negative aspects such as excessive demands, frequent conflicts or serious worries (Due et al. 1999). The association between negative aspects of social relations and health as well as potential pathways has received relatively less attention within health research (Chiang et al. 2012; Donoho et al. 2013; Ekmann et al. 2012; Lund et al. 2010, 2014a, b; Rook 2015; Walen and Lachman 2000; Wilson et al. 2015).

Few studies within the field of social relations and low-grade inflammation (Bajaj et al. 2016; Nakata et al. 2014; Yang et al. 2014) have looked into the association with different types of relations such as spouse, partner, family, friends and colleagues. Moreover, there is still a lack of studies assessing variation by other social environmental factors such as gender and socioeconomic position (SEP) (Elliot et al. 2018; Uchino et al. 2016; Yang et al. 2014). The study of variations in associations by sociodemographic factors requires large-scale studies in order to perform stratified analyses or test interactions, but large-scale population studies in this field are scarce (Uchino et al. 2018).

The aim of the present study was to examine the association between structural and positive as well as negative functional aspects of social relations, across four relational domains—partner/spouse, children, other family and friends, and the level of low-grade inflammation in a large population-based middle-aged cohort. A second aim was to examine variation by gender and SEP in these associations.

Methods

The present study was based on data from the Copenhagen Aging and Midlife Biobank (CAMB) study (Lund et al. 2015). CAMB includes participants from three earlier established cohort studies: The Metropolit 1953 Danish Male Birth Cohort (MP), the Copenhagen Perinatal Cohort (CPC), and the Danish Longitudinal Study on Work, Unemployment and Health (DALWUH). In total, 17,937 cohort members aged 49–61 living in the Eastern parts of Denmark were invited to participate in CAMB (7750 from MP, 5282 from CPC, and 4906 from DALWUH). In total, 5576 (31%) participated in all parts of the study, physical and cognitive testing, clinical examination, blood samples, and extensive questionnaire data. For this study, participants diagnosed with immune deficiency, HIV, Hepatitis C/B or inflammation in heart valves were excluded (n = 64). The statistical analyses were performed on subsamples with complete information on included variables, and for each inflammatory marker separately. The samples for analyses of the association between each aspect of social relations and each of three markers of low-grade inflammation varied between n = 4401–5169. The smaller sample sizes were primarily due to respondents not having a partner or children. The Ethical Committee of the Capital Region of Denmark approved the study (No: H-A-2008-126), and written, informed consent was obtained from all participants. CAMB is also registered at the Danish Data Protection Agency as a combined database (No: 2008-41-2938).

Measures of social relations

Social relations were assessed by the validated Copenhagen Social Relations Questionnaire (CSRQ), showing satisfactory content and face validity as well as reliability (Lund et al. 2013). Questions on functional aspects of social relations included perceived emotional and instrumental support, perceived demands, worries and conflicts with partner, children, other family and friends, respectively. Cohabitation status, face-to-face as well as non-face-to-face contact frequency with other family and friends were included as markers of structural aspects of social relations.

Perceived emotional and instrumental support were measured by two questions: ‘Can you talk with any of the following people, if you need support?’ and ‘Would any of the following people help you with daily practical matters, if necessary?’ each including one item for each of the following social domains: partner, children (your own or partners), other family, and friends. The response key was always, often, sometimes, seldom/never, have none. The perception of worries, demands and conflicts were measured by three questions: ‘In your everyday life, do you experience that any of the following people demand too much of you?’, ‘In your everyday life, do you experience that any of the following people seriously worry you?’ and ‘In your everyday life, do you experience conflicts with any of the following people?’, each with one item for each of the following social domains: partner, children (your own or partners), other family, and friends. The response key was: always, often, sometimes, seldom/never, have none. For the positive aspects of functional social relations, the response categories seldom and never were collapsed, and for the negative aspects always and often were collapsed, due to a low number of responses in these categories.

Contact frequency with other family and friends was assessed by the questions ‘How often are you together with any of the following people, who you do not live with?’ and ‘How often do you have contact with any of the following people without seeing them (e.g., by telephone, letters, e-mail, sms)?’ with the response key several days a week, about once a week, one to three times a month, less often than once a month/never.

Systemic low-grade inflammation

The included inflammatory markers were Interleukin 6 (IL-6), Tumor Necrosis Factor–Alpha (TNF-α) and high-sensitivity C-reactive protein (hsCRP). Non-fasting blood samples were collected and initially fractionated in plasma, serum and DNA. IL-6 and TNF-α were analyzed in EDTA plasma by electro-chemiluminescence multiplex system on a Sector 2400 Imager. The lower limit of detection (LOD) for IL-6 was 0.21 pg/ml and interassay coefficient of variation (CV) was 11–21%. LOD was 0.28 pg/ml (and interassay CV 9–13%) for TNF-α. There were 0.4% values below the LOD for IL-6, and these values were substituted using simple imputation. HsCRP was analyzed from plasma samples by Roche/Hitachi MODULAR P, with a measuring range 0.1–20 mg/L (0.95–190 mmol/L). Analyses were conducted by the Clinic Biochemical Laboratory and Centre of Inflammation and Metabolism, Rigshospitalet, Copenhagen. The inflammatory markers were analyzed individually as continuous variables.

Covariates

The variables identified as potential confounders, based on previous literature, were gender, age, SEP, marital status, early major lifeevents (eMLE) and morbidity. BMI, acute inflammatory events and medicine intake were included in the analytical model, as they are highly associated with inflammatory levels.

SEP was categorized according to the Danish Occupational Social Class measure which is based on the assessments of the occupational skills and competencies necessary for the job as well as the power and control associated with the position (Christensen et al. 2014). Social class I constitute the highest occupational levels, and V is unskilled manual workers. VI represents people on transfer income, including sickness benefits and disability pension. Social class VII and VIII were not used in this study. Social class was organized in four categories: social class I and II, social class III, social class IV and V, and social class VI. To assess eMLE participants were asked to specify if they had experienced the death or serious illness of a parent, had a conflict-ridden childhood, had experienced financial difficulties during childhood or had been placed in foster-care.

Participants were interviewed at the enrollment in the clinical examination part of CAMB in order to obtain information on acute inflammatory events including recent illness, surgery, bone-fracture or a dentist visit within the past three weeks. Information on chronic diseases was collected as self-reported data and as part of the enrollment interview. Participants were asked whether they have or have had any of a list of 21 individual diseases. A comorbidity index was constructed, encompassing those disease categories of immediate relevance to low-grade inflammation: (1) hypertension, (2) cardiovascular disease including angina, myocardial infarction and stroke, (3) pulmonary disease including asthma, chronic obstructive pulmonary disease, emphysema and chronic bronchitis, (4) allergy, (5) diabetes, (6) inflammatory diseases including rheumatoid arthritis, osteoarthritis, other connective tissue disorders and inflammatory bowel disease, (7) cancer including leukemia, (8) depression, anxiety, and (9) other psychiatric disorders, ranging 0–9. Information on use of medication was obtained at the enrollment interview and an index of pharmaceuticals known to influence inflammatory levels was constructed. The index included NSAID, statins, steroids and immunosuppressive agents including cytostatic and anti-cytokine therapy, ranging 0–4.

Table 1 shows the characteristics of the study sample. The majority of the participants were men (68.7%) due to the MP being part of CAMB. Mean age was 54.41 years. A quarter of the participants were manual workers (24.9%), another quarter was from social class III (non-manual occupations medium level), and 42% of participants were in social Class I & II (non-manual occupations high level), leaving 9.5% on transfer income. More than 50% experienced none of the measured eMLE, whereas 36% experienced 1–2 events. Fifteen percent were living alone. More than 50% of the population was overweight or obese. Thirty-two percent experienced a recent episode of acute systemic event/disease related to inflammation, 37.1% reported having no and 54% reported having 1–2 chronic diseases. The majority reported no medication intake (80.3%). Mean of TNF-α was 5.80 (standard deviation (SD) 12.54) pg/ml, IL-6 3.55 (SD 14.72) pg/ml and hsCRP 2.30 (SD 4.20) mg/L. Standard deviations of TNF-α and IL-6 were relatively high displaying the skewed distribution before log-transformation. Significantly, more men than women had high social class, were overweight and scored high on the index of pharmaceuticals, whereas more women than men had experienced eMLE (data not shown). No gender differences were found regarding the other included variables.

Table 1.

Descriptive statistics stratified by gender

All (n = 5169) Women (n = 1617) Men (n = 3552)
Mean (SD)
Age 54.41 (3.89) 52.46 (4.34) 55.30 (3.32)
Inflammatory markers
TNF-α (pg/ml) 5.80 (12.54) 5.54 (15.01) 5.93 (11.24)
IL6 (pg/ml) 3.55 (14.72) 3.41 (17.27) 3.62 (13.41)
hsCRP (mg/L) (n = 5154) 2.30 (4.20) 2.24 (4.21) 2.33 (4.19)
Frequency (%)
Socioeconomic position
Social Class I + II 2172 (42.0) 620 (38.3) 1552 (43.7)
Social Class III 1223 (23.7) 396 (24.5) 827 (23.3)
Social Class IV + V 1285 (24.9) 454 (28.1) 831 (23.4)
Social Class VI 489 (9.5) 147 (9.1) 342 (9.6)
Early major life events
0 2782 (53.8) 784 (48.5) 1998 (56.3)
1 1152 (22.3) 384 (23.8) 768 (21.6)
2 701 (13.6) 259 (16.0) 442 (12.4)
3 343 (6.6) 110 (6.8) 233 (6.6)
4 130 (2.5) 51 (3.2) 79 (2.2)
5–6 61 (1.2) 29 (1.8) 32 (0.9)
Cohabitation status
Not living alone 4395 (85.0) 1354 (83.7) 3041 (85.6)
Living alone 774 (15.0) 263 (16.3) 511 (14.4)
BMI
Underweight (< 18, 5) 54 (1.0) 39 (2.4) 15 (0.4)
Normal weight (18, 5–25) 2183 (42.2) 858 (53.1) 1325 (37.3)
Overweight (25–30) 2178 (42.1) 504 (31.2) 1674 (47.1)
Obese (> 30) 754 (14.6) 216 (13.4) 538 (15.2)
Acute systemic event < 3 weeks
Yes 1651 (31.9) 503 (31.1) 1148 (32.3)
No 3518 (68.1) 1114 (68.9) 2404 (67.7)
Morbidity index (0–7)a
0 1919 (37.1) 611 (37.8) 1308 (36.8)
1 1864 (36.1) 586 (36.2) 1278 (36.0)
2 923 (17.9) 286 (17.7) 637 (17.9)
3 312 (6.0) 89 (5.5) 223 (6.3)
4 111(2.2) 34 (2.1) 77 (2.2)
5-7 40 (0.8) 11 (0.7) 29 (0.8)
Use of pharmaceutical index (0–3)b
0 4152 (80.3) 1314 (81.3) 2838 (79.9)
1 951 (18.4) 294 (18.2) 657 (18.5)
2–3 66 (1.3) 9 (0.6) 57 (1.6)
Open in a new tab

TNF-α: Tumor Necrosis Factor-α. IL-6: Interleukin-6. hsCRP: high-sensitivity C-reactive Protein

aMorbidity index includes hypertension, cardiovascular disease, pulmonary disease, allergy, diabetes, inflammatory diseases, cancer, and psychiatric morbidity

bPharmaceutical index includes NSAID, Statins, Steroids, and Immunosuppressive medication

Statistics

Multiple linear regression models with the inflammatory markers as dependent variables were performed. Analyses were performed for each item of social relations separately. To approach normality, the inflammatory markers were log-transformed. Regression coefficients are judged statistically significant if within the 95% confidence intervals after back-transformation, meaning that an estimated coefficient of 1.05 is interpreted as a 5% higher level of the inflammatory marker compared to a reference level. The first model included only the respective measures of social relations. In the second model, we adjusted for gender, age, SEP, cohabitation status, eMLE, BMI, acute systemic inflammatory event, comorbidity and pharmaceutical intake. Only the estimates from the adjusted models are presented. To assess any differences in the associations between social relations and the inflammatory markers by gender or SEP, in additional models interaction terms were added. Overall significance of the association with inflammatory markers across each measure of social relations was tested by the F-test. Reference categories were the category hypothesized to have the lowest level of inflammatory markers, i.e., least adverse score on the relevant measure of social relations, see Tables 2, 3 and 4. Analyses were performed in SAS, version 9.4, using PROC GLM.

Table 2.

Multiple linear regression analysis of positive functional aspects of social relations (emotional support and instrumental support) and TNF-alpha, IL-6 and hsCRP. Statistically significant estimates are presented in bold

TNF-alpha IL-6 hsCRP
Coefficient 95% CI P values F-test Coefficient 95% CI P values F-test Coefficient 95% CI P values F-test
n = 4436 n = 4436 n = 4421
Emotional support from partner
Always Ref. Ref. 0.40 Ref. 0.76
Often 0.99 0.96 1.03 0.26 1.03 0.97 1.10 0.96 0.89 1.05
Sometimes 1.06 1.00 1.13 1.06 0.96 1.18 1.01 0.88 1.16
Seldom/Never 1.03 0.93 1.14 1.10 0.93 1.31 0.94 0.75 1.17
Interaction with gender 0.96 0.86 0.48
Interaction with social class 0.17 0.15 0.53
n = 4697 n = 4697 n = 4680
Emotional support from children
Always Ref. Ref. Ref.
Often 1.03 0.99 1.06 0.49 1.02 0.97 1.08 0.23 0.99 0.92 1.06 0.58
Sometimes 1.01 0.97 1.05 1.01 0.95 1.08 0.93 0.86 1.01
Seldom/Never 1.01 0.96 1.05 1.00 0.93 1.08 0.94 0.85 1.03
Interaction with gender 0.42 0.71 0.60
Interaction with social class 0.98 0.42 0.96
n = 5115 n = 5115 n = 5097
Emotional support from other family
Always Ref. Ref. Ref.
Often 1.01 0.98 1.05 0.90 1.04 0.98 1.10 0.44 1.06 0.99 1.14 0.03
Sometimes 1.01 0.98 1.04 1.01 0.96 1.07 1.03 0.96 1.11
Seldom/Never 1.00 0.97 1.04 1.04 0.98 1.10 0.94 0.87 1.02
Interaction with gender 0.46 0.76 0.25
Interaction with social class 0.31 0.96 0.02
n = 5133 n = 5133 n = 5115
Emotional support from friends
Always Ref. Ref. Ref.
Often 0.99 0.96 1.02 0.62 1.01 0.96 1.06 0.11 1.05 0.98 1.12 0.50
Sometimes 1.02 0.98 1.05 1.04 0.98 1.10 1.03 0.96 1.11
Seldom/Never 1.01 0.96 1.05 1.09 1.01 1.17 0.99 0.90 1.09
Interaction with gender 0.37 0.29 0.54
Interaction with social class 0.96 0.95 0.66
n = 4432 n = 4432 n = 4417
Instrumental support from partner
Always Ref. Ref. Ref.
Often 1.00 0.96 1.03 0.92 1.00 0.94 1.06 0.43 0.92 0.85 0.99 0.12
Sometimes 1.01 0.95 1.07 0.99 0.90 1.08 0.96 0.85 1.08
Seldom/Never 1.03 0.94 1.12 1.13 0.97 1.30 1.00 0.83 1.21
Interaction with gender 0.11 0.82 0.92
Interaction with social class 0.03 0.01 0.01
n = 4695 n = 4695 n = 4678
Instrumental support from children
Always Ref. Ref. Ref.
Often 1.00 0.97 1.03 0.96 1.01 0.96 1.07 0.23 0.96 0.90 1.03 0.58
Sometimes 1.00 0.97 1.04 1.06 1.00 1.13 0.96 0.89 1.04
Seldom/Never 0.99 0.94 1.04 1.03 0.95 1.12 0.95 0.86 1.06
Interaction with gender 0.86 0.32 0.60
Interaction with social class 0.38 0.14 0.96
n = 5108 n = 5108 n = 5090
Instrumental support from other family
Always Ref. Ref. Ref.
Often 1.01 0.97 1.04 0.57 1.00 0.94 1.06 0.75 0.97 0.89 1.05 0.16
Sometimes 1.02 0.98 1.06 1.02 0.96 1.09 0.98 0.91 1.06
Seldom/Never 1.00 0.96 1.03 0.99 0.93 1.06 0.92 0.85 1.00
Interaction with gender 0.64 0.46 0.59
Interaction with social class 0.60 0.64 0.98
n = 5131 n = 5131 n = 5113
Instrumental support from friends
Always Ref. Ref. Ref.
Often 0.99 0.95 1.02 0.49 0.96 0.90 1.02 0.57 0.97 0.90 1.05 0.32
Sometimes 1.01 0.97 1.04 0.98 0.92 1.04 0.94 0.87 1.02
Seldom/Never 0.98 0.94 1.02 0.98 0.91 1.05 0.93 0.85 1.01
Interaction with gender 0.86 0.84 0.44
Interaction with social class 0.13 0.42 0.57
Open in a new tab

Estimates are from models adjusted for age, gender, social class, cohabitation status, major life events, morbidity, medication, acute inflammatory event, body mass index

Table 3.

Multiple linear regression analysis of negative functional aspects of social relations (perceived demands, perceived worries and perceived conflicts) and TNF-alpha, IL-6 and hsCRP. Statistically significant estimates are presented in bold

TNF-alpha IL-6 hsCRP
Coefficient 95% CI P values F-test Coefficient 95% CI P values F test Coefficient 95% CI P values F test
n = 4423 n = 4423 n = 4408
Perceived demands from partner
Always/often 0.99 0.93 1.04 0.73 0.98 0.89 1.07 0.91 1.02 0.90 1.15 0.62
Sometimes 1.01 0.98 1.05 0.99 0.93 1.05 1.03 0.95 1.11
Seldom 1.01 0.98 1.04 0.98 0.93 1.04 1.05 0.98 1.13
Never Ref. Ref. Ref.
Interaction with gender 0.46 0.16 0.76
Interaction with social class 0.86 0.09 0.05
n = 4688 n = 4688 n = 4671
Perceived demands from children
Always/often 1.01 0.95 1.06 0.17 0.97 0.89 1.07 0.30 1.00 0.88 1.12 0.75
Sometimes 0.97 0.93 1.06 0.94 0.89 1.00 1.04 0.96 1.13
Seldom 1.00 0.97 1.03 0.98 0.93 1.03 1.02 0.95 1.09
Never Ref. Ref. Ref.
Interaction with gender 0.62 0.61 0.54
Interaction with social class 0.39 0.32 0.47
n = 5098 n = 5098 n = 5080
Perceived demands from other family
Always/often 0.97 0.90 1.04 0.69 0.91 0.81 1.03 0.36 0.90 0.77 1.04 0.04
Sometimes 1.01 0.97 1.05 0.99 0.93 1.06 1.09 1.01 1.19
Seldom 0.99 0.97 1.02 0.97 0.93 1.02 0.99 0.93 1.05
Never Ref. Ref. Ref.
Interaction with gender 0.65 0.36 0.83
Interaction with social class 0.42 0.28 0.75
n = 5115 n = 5115 n = 5097
Perceived demands from friends
Always/often 1.01 0.88 1.16 0.98 1.08 0.86 1.36 0.44 1.11 0.82 1.50 0.03
Sometimes 0.99 0.94 1.04 1.07 0.98 1.16 1.17 1.05 1.31
Seldom 1.00 0.97 1.02 1.00 0.96 1.04 1.04 0.98 1.10
Never Ref. Ref. Ref.
Interaction with gender 0.90 0.89 0.58
Interaction with social class 0.91 0.70 0.12
n = 4416 n = 4416 n = 4401
Perceived worries from partner
Always/often 1.00 0.94 1.06 1.00 0.97 0.87 1.07 0.75 0.95 0.83 1.09 0.82
Sometimes 1.00 0.96 1.04 0.98 0.92 1.04 1.00 0.93 1.09
Seldom 1.00 0.97 1.03 0.97 0.92 1.03 0.98 0.92 1.05
Never Ref. Ref. Ref.
Interaction with gender 0.48 0.17 0.13
Interaction with social class 0.85 0.25 0.62
n = 4679 n = 4679 n = 4662
Perceived worries from children
Always/often 1.02 0.96 1.07 0.87 0.96 0.88 1.05 0.44 0.94 0.84 1.06 0.66
Sometimes 1.00 0.96 1.03 0.95 0.89 1.01 1.00 0.92 1.08
Seldom 1.00 0.96 1.03 0.98 0.92 1.04 0.97 0.90 1.05
Never Ref. Ref. Ref.
Interaction with gender 0.47 0.40 0.76
Interaction with social class 0.51 0.52 0.91
n = 5091 n = 5091 n = 5073
Perceived worries from other family
Always/often 0.99 0.93 1.05 0.03 0.93 0.84 1.03 0.08 0.90 0.79 1.03 0.44
Sometimes 0.96 0.03 1.00 0.93 0.88 0.99 1.00 0.92 1.07
Seldom 0.96 0.93 0.99 0.95 0.90 1.00 1.00 0.93 1.07
Never Ref. Ref. Ref.
Interaction with gender 0.95 0.50 0.65
Interaction with social class 0.97 0.16 0.26
n = 5110 n = 5110 n = 5092
Perceived worries from friends
Always/often 0.98 0.85 1.12 0.32 0.97 0.77 1.23 0.44 1.25 0.92 1.70 0.42
Sometimes 1.00 0.96 1.04 0.99 0.92 1.05 1.05 0.96 1.14
Seldom 0.98 0.95 1.00 0.96 0.92 1.01 1.01 0.96 1.08
Never Ref. Ref. Ref.
Interaction with gender 0.70 0.56 0.94
Interaction with social class 0.79 0.65 0.87
n = 4414 n = 4414 n = 4414
Conflicts with partner
Always/often 0.99 0.92 1.06 0.84 0.94 0.84 1.06 0.61 0.94 0.81 1.10 0.65
Sometimes 1.01 0.97 1.05 0.99 0.92 1.05 0.98 0.90 1.07
Seldom 1.01 0.98 1.05 1.01 0.95 1.07 1.02 0.94 1.10
Never Ref. Ref. Ref.
Interaction with gender 0.41 0.64 0.15
Interaction with social class 0.32 0.10 0.88
n = 4678 n = 4678 n = 4661
Conflicts with children
Always/often 1.00 0.93 1.08 0.85 0.95 0.84 1.08 0.67 0.90 0.77 1.06 0.54
Sometimes 1.01 0.97 1.05 0.98 0.91 1.04 0.98 0.90 1.06
Seldom 1.01 0.98 1.05 1.00 0.95 1.06 1.01 0.94 1.08
Never Ref. Ref. Ref.
Interaction with gender 0.70 0.84 0.92
Interaction with social class 0.56 0.25 0.96
n = 5087 n = 5087 n = 5069
Conflicts with other family
Always/often 1.02 0.92 1.12 0.93 0.79 0.67 0.93 0.07 0.83 0.67 1.02 0.29
Sometimes 0.99 0.95 1.03 0.97 0.90 1.04 1.00 0.92 1.10
Seldom 1.00 0.97 1.02 0.94 0.90 0.99 0.97 0.92 1.03
Never Ref. Ref. Ref.
Interaction with gender 0.42 0.75 0.79
Interaction with social class 0.43 0.50 0.23
n = 5107 n = 5107 n = 5089
Conflicts with friends
Always/often 1.02 0.76 1.38 0.94 1.15 0.70 1.90 0.94 1.01 0.53 1.93 0.08
Sometimes 1.00 0.94 1.06 1.01 0.91 1.12 1.19 1.04 1.37
Seldom 0.99 0.97 1.02 1.01 0.97 1.05 1.03 0.98 1.09
Never Ref. Ref. Ref.
Interaction with gender 0.60 0.84 0.66
Interaction with social class 0.15 0.20 0.52
Open in a new tab

Estimates are from models adjusted for age, gender, social class, cohabitation status, major life events, morbidity, medication, acute inflammatory event, body mass index

Table 4.

Multiple linear regression analysis of structural aspects of social relations (cohabitation status, face-to-face contact, and contact (SMS, e-mail, phone) and TNF-alpha, IL-6 and hsCRP. Statistically significant estimates are presented in bold

TNF-alpha IL-6 hsCRP
Coefficient 95% CI P values F test Coefficient 95% CI P values F test Coefficient 95% CI P values F test
n = 5169 n = 5159 n = 5151
Cohabitation status
Living alone 1.01 0.97 1.04 1.11 1.04 1.17 1.12 1.04 1.21
Not living alone Ref. Ref. Ref.
Interaction with gender 0.27 0.01 0.11
Interaction with social class 0.24 0.94 0.28
n = 5113 n = 5113 n = 5095
Face-to-face contact with relatives
Several days a week Ref. 0.18 Ref. 0.76 Ref. 0.45
About once a week 1.05 1.00 1.11 1.05 0.96 1.15 1.09 0.97 1.22
One to three times a month 1.02 0.97 1.07 1.03 0.95 1.12 1.05 0.94 1.16
Less than once a month/never 1.02 0.97 1.07 1.03 0.95 1.12 1.05 0.94 1.17
Interaction with gender 0.03 0.13 0.09
Interaction with social class 0.59 0.58 0.81
n = 5124 n = 5124 n = 5106
Face-to-face contact with friends
Several days a week Ref. Ref. Ref.
About once a week 0.98 0.94 1.02 0.85 0.97 0.91 1.04 0.71 0.89 0.81 0.98 0.08
One to three times a month 0.99 0.95 1.03 0.99 0.92 1.06 0.92 0.84 1.00
Less than once a month/never 0.99 0.94 1.03 1.01 0.93 1.09 0.89 0.81 0.99
Interaction with gender 0.27 0.06 0.08
Interaction with social class 0.60 0.50 0.06
n = 5117 n = 5117 n = 5099
Contact with relatives
Several days a week Ref. Ref. Ref.
About once a week 1.00 0.97 1.04 0.94 1.01 0.95 1.06 0.53 1.06 0.99 1.14 0.34
One to three times a month 1.00 0.96 1.03 1.04 0.98 1.10 1.00 0.93 1.08
Less than once a month/never 0.99 0.95 1.03 1.03 0.96 1.10 1.04 0.95 1.13
Interaction with gender 0.47 0.04 0.39
Interaction with social class 0.87 0.16 0.97
n = 5132 n = 5132 n = 5114
Contact with friends
Several days a week Ref. Ref Ref.
About once a week 0.98 0.95 1.02 0.78 1.01 0.96 1.07 0.63 0.99 0.92 1.07 0.33
One to three times a month 0.99 0.96 1.03 1.04 0.98 1.09 1.00 0.94 1.08
Less than once a month/never 0.99 0.95 1.03 1.03 0.96 1.10 0.93 0.85 1.02
Interaction with gender 0.62 0.49 0.16
Interaction with social class 0.69 0.56 0.84
Open in a new tab

Estimates are from models adjusted for age, gender, social class, cohabitation status, major life events, morbidity, medication, acute inflammatory event, body mass index

Results

The distribution of functional and structural aspects of social relations, respectively, is shown in Online Resource 1. For this population especially partner and to some degree children were the most common source of emotional and instrumental support, but also of demands. Perceived worries were most common for children, and conflicts with partner or children were by far more common than conflicts with friends. About 14% were seldom in non-face-to-face contact with non-cohabiting family or friends (less than once a month or never), 28% and 17% had seldom face-to-face contact with non-cohabiting family or friends, respectively.

Table 2 shows the results from the multiple linear regression analyses of the association between positive functional aspects of social relations and TNF-α, IL-6 and hsCRP for all relational domains. Overall, we found only few and weak associations between perceived emotional or instrumental support from partner, children, other family or friends and inflammatory markers. In fully-adjusted analyses, compared to always perceiving support, seldom/never perceiving emotional support from friends or only sometimes perceiving instrumental support from children was associated with significantly higher level of IL-6, whereas the perception of often receiving instrumental support from partner was associated with significantly lower level of hsCRP. For hsCRP, the overall association with emotional support from other family was significant, and results suggest a protective effect of low levels of support. There were no significant interactions between gender and any of the items of social relations, whereas we found significant interactions between emotional support from other family and social class in relation to hsCRP, and between instrumental support from partner and social class in relation to all three inflammatory markers. In analyses stratified by social class (data not shown), the significant interaction between emotional support from family and social class in relation to hsCRP appeared to originate from significantly higher hsCRP among those reporting often perceived support in social class IV and V as the only significant estimates. The significant interactions between instrumental support from partner and social class appeared to originate from significantly lower IL-6 and hsCRP in social class VI but no association in other social classes, thus indicating a tendency to a protective effect of being in a low social class, whereas a significantly higher TNF-α among those perceiving support often was found in social class III only.

Table 3 shows the multiple linear regression analyses of the association between negative functional aspects of social relations and TNF-α, IL-6 and hsCRP for all relational domains. Only very weak and inconsistent associations between perceived demands, worries or conflicts from/with partner, children, other family or friends and inflammatory markers were identified. In fully-adjusted analyses, compared to never perceiving strain, sometimes perceiving demands from other family or friends, and sometimes perceiving conflicts with friends were associated with significantly higher level of hsCRP. The perception of worries sometimes and seldom from other family was associated with significantly lower level of TNF-α and IL-6, and the perception of conflicts always/often and seldom with other family was associated with significantly lower level of IL-6. There were no significant interactions with gender. We found a significant interaction between perceived demands from partner and social class in relation to hsCRP, and in analyses stratified by social class (data not shown) we found a significant association between sometimes perceiving demands from partner and higher hsCRP in social class III only, whereas there was a significant association between sometimes perceiving demands from partner and lower hsCRP in social class VI only.

Table 4 shows the multiple linear regression analyses of the association between structural aspects of social relations and TNF-α, IL-6 and hsCRP. The associations were weak. Living alone was associated with significantly higher levels of IL-6 and hsCRP, whereas face-to-face contact with friends about once a week or less than once a month/never were associated with significantly lower level of hsCRP, compared to frequent contact. We found no significant interactions with social class, whereas we found significant interactions between cohabitation status and non-face-to-face contact with friends, respectively, and gender in relation to IL-6, and between face-to-face contact with relatives and gender in relation to TNF-α. In analyses stratified by gender (data not shown), we found that there was a significant association between living alone and higher levels of IL-6 and hsCRP among men only. Regarding face-to-face contact with relatives, there was a significant association between contact about once a week (and a similar tendency for the other contact frequencies compared to several days a week) and a higher level of TNF-α among men only. Finally, for non-face-to-face contact with family, the interaction appeared to originate from a significant association between contact about once a week and lower IL-6 among women and higher IL-6 among men.

Discussion

This study set out to assess the association between structural and positive as well as negative functional aspects of social relations, across four relational domains—partner/spouse, children, other family and friends, and the level of low-grade inflammation in a large population-based middle-aged cohort and to examine variation by gender and SEP in these associations. Overall, we found no clear, and mostly ambiguous, associations between social relations and low-grade inflammation, and no considerable variation by SEP. Concluding cautiously, men appear to be more vulnerable toward living alone and low contact frequency with family compared to women as regards high level of low-grade inflammation.

Rook (2015) argue that negative social interactions occur with lower frequency than positive interactions, but that health-effects of negative social interactions often outweigh the effects of positive interactions. We could not identify any clear association with inflammatory markers of neither structural nor positive or negative aspects of functional social relations. This might have several explanations, including the measurement of social relations and inflammatory cytokines, and issues regarding the age and representativeness of the study population, which will be discussed below.

In their meta-analysis Uchino et al. (2018) argue that effect sizes for the association between markers of social relations and cytokines were equal, but with a tendency toward being largest for social integration, followed by perceived support and smallest for received support. That effect sizes might vary according to measure of social relations is supported by a recent study by Elliot et al. (2018) who suggest cautiously, also based on previous literature that measures of structural social relations are more strongly associated with CRP in older men, whereas measures of functional social relations might be more strongly related to IL-6 among older women. However, in a study on US black women aged 24–34 there was only a significant association between higher perceived relationship quality and lower hs-CRP, and not between social integration and hs-CRP (Ford et al. 2019). We applied measures of social support as well as strain from four distinct social domains, thus looking into more detailed nuances of the functional aspects of social relations than previous studies. Perhaps structural measures of contact frequency or social activities within these different social domains would have showed stronger effect sizes. Despite the findings of the meta-analysis (Uchino et al. 2018), this study is not the first to present an overall lack of a clear association between social relations and inflammation (Bajaj et al. 2016; Brown et al. 2018; Loucks et al. 2006b; Uchino et al. 2016, 2018), highlighting the continuous need for studies on this matter, preferably in large population samples with markers of social relations allowing comparison with previous studies, as well as more detailed knowledge on different aspects of social relations.

As mentioned, only few previous studies have assessed positive as well as negative aspects of social relations and their association with inflammatory markers. Yang et al. (2014) studied perceived social support and social strain from family, friends and spouse in a US midlife population and found that social support from all social domains summarized together modestly protected against subsequent overall inflammation, whereas strain from family, friends and total strain increased the risk of inflammation. Looking into the specific inflammatory markers also used in the present study, they found that family support was significantly related to lower levels of IL-6, and family strain was significantly related to higher levels of CRP and IL-6, but not in fully-adjusted models (Yang et al. 2014). Assessing variations by age and gender in the same population, there was a tendency to higher IL-6 among those below age 45 reporting social strain, which attenuated to non-significance in fully-adjusted models (Elliot et al. 2018). Social strain was not associated with CRP (Elliot et al. 2018). Yang et al. (2014) identified a substantial association between family strain and inflammation and called for further studies of strain from family and inflammation. The authors suggested that family strain might be particularly detrimental regarding level of inflammation, compared to other social domains, as family relations are by nature long-standing and not discretionary (Yang et al. 2014). In the present study, it is thus a major strength that we distinguished between children and other family members. We found no significant associations between the three different markers of negative social interaction with children and the three inflammatory markers. In addition, we did see significant, but contradictory, associations between negative social interaction with other family and inflammation. This might suggest that a certain amount of worries, demands and conflict with children is in line with expectations of everyday life and thus less harmful for level of stress and subsequent health. However, a longitudinal study by Lund et al. (2014a) reported that worries and demands from partner and children, compared to more distant social relations, were more strongly associated with mortality.

Donoho et al. (2013) studied marital support and marital strain in relation to IL-6 and CRP and found that low level of spousal support was only significantly associated with inflammation among women, whereas marital strain was not related to inflammation. It appears that the association between negative aspects of social relations with low-grade inflammation is weaker than the association between, e.g., social support and structural aspects, which with caution could also be concluded from the present study.

The very modest associations between social relations and inflammation found in the present study might also be explained by differences according to age in the effect on low-grade inflammation by social relations. This is supported by Eisenberger et al. (2017), who argue that both objective and subjective social isolation increase inflammation typically in older samples, and that other stressors might have to be present to elicit a similar inflammatory response in younger samples. Using cohorts of different ages Yang et al. (2016) found that social integration was associated with significantly lower odds of high CRP in young and old adulthood, but found no association in midlife. This finding confirms previous findings by Ford et al. (2006). By contrast perceived support and perceived strain were particularly related to CRP and other physiological markers in midlife; however, associations were generally more modest compared to social integration in all age groups (Yang et al. 2016). Elliot et al. (2018) pointed to the association between social strain and IL-6 being larger in younger (below age 45) compared to older age-groups, and also Bajaj et al. (2016) found different associations according to age of the study population. Recently, it has also been shown that loneliness, a marker of perceived lack of social relations and support, is significantly related to higher CRP, fibrinogen and IL-6 in midlife (Nersesian et al. 2018), supporting the conclusion of a recent review on loneliness and acute stress reactivity (Brown et al. 2018). Other studies suggested no interactions by age in the association between positive as well as negative aspects of social relations and inflammation (Yang et al. 2014). All in all, the ambiguous findings of the present study might be explained by the age of the study sample and the available markers of social relations.

The present study found weak support for gender and SEP differences in the association between social relations and low-grade inflammation. Previous studies in this field have yielded diverging findings. Social integration has been found to be associated with lower levels of inflammatory markers only among men (Ford et al. 2006; Loucks et al. 2006a; b), among both genders (Elliot et al. 2018) and among neither men nor women (Glei et al. 2012), whereas measures of functional social relations have been found to be associated with higher levels of CRP in men (Mezuk et al. 2010), lower level of CRP among both genders (Elliot et al. 2018; Glei et al. 2012), lower level of IL-6 and/or CRP among women only (Donoho et al. 2013; Elliot et al. 2018) and no association with CRP among neither men nor women (Kamiya et al. 2010). Also no gender differences in the impact of positive and negative aspects of social relations on inflammation have been shown (Yang et al. 2014). To the authors’ knowledge, this is the first study on positive and negative markers of social relations in association with low-grade inflammation that assess variation by SEP. The rationale for this is that level of inflammatory markers is significantly related to educational level (Loucks et al. 2006c; McDade et al. 2011; Stepanikova et al. 2017), wealth (McDade et al. 2011) and income (Friedman and Herd 2010), and social relations are also unequally distributed according to SEP (Cerhan and Wallace 1993; Heap et al. 2017; Krause and Borawski-Clark 1995; Weyers et al. 2008; Aartsen et al. 2017). Our findings from these analyses were inconclusive.

The present study has important strengths and limitations. A limitation of most studies on social relations and inflammation, shared by the present study, is the cross-sectional study design, which does not allow for assessment of any causal relation between social relations and inflammation (Eisenberger et al. 2017). In order to overcome issues of causality, we adjusted analyses for a range of chronic conditions, pharmaceutical intake and acute inflammatory conditions, and furthermore adjusted for eMLE (Dich et al. 2015) and BMI (Pedersen et al. 2016). It is important to separate symptoms of depression and anxiety from the perception of especially demands and worries and lack of social support, as studies have found depression as well as anxiety associated with elevated inflammatory markers (Puustinen et al. 2011).

A strength of the present study is the use of a large sample of the background-population in terms of health conditions and health-behaviors (Avlund et al. 2014). Compared to the previous similar study by Yang et al. (2014), it is a strength that unmarried individuals were not excluded. Non-participants in CAMB were more often men (in the DALWUH and CPC cohorts), unemployed or on transfer income, had lower educational level and a somewhat poorer health (Lund et al. 2015), which might have contributed to a selected study sample, where associations are underestimated. Yang et al. (2014) suggest that their lack of finding of a strong protective effect of social support against inflammation might be due to their national sample of community dwelling Americans that might be more heterogeneous than previous study samples, and that the effect of social support might appear stronger among individuals facing severe physiological stress. This might also apply to our findings.

The validated CSRQ tool for assessment of social relations (Lund et al. 2013), favoring structural and functional aspects of social relations from several social domains, is an important strength. However, other measures of social relations might be relevant for further studies. Bajaj et al. (2016) applied ecological momentary assessment with hourly assessments of social interactions and suggest that actual social events promote inflammatory responses, and these events will not be captured in commonly used measures of social support or social integration. Kim et al. (2016) argued that measuring social connectedness as the number of times an individual is nominated by others in the social network as friend or family member, is a more valid marker of actual support received, than the number of social contacts named by the individual, which indicate perceived support. The association of the latter to fibrinogen level was weaker than the association between social connectedness defined as above and fibrinogen.

Another limitation of this study is that occupational social class was the only applied measure of SEP, as studies indicate that different measures of SEP are associated with inflammation and perhaps not to a similar degree (Friedman and Herd 2010; Loucks et al. 2006c; McDade et al. 2011; Stepanikova et al. 2017). Markers of SEP in childhood might also be relevant to consider in relation to adult inflammation (Liu et al. 2017). However, occupational social class is indeed a relevant marker of SEP in a midlife population, where most are in active employment (Christensen et al. 2014).

Some studies have argued that the lack of an inflammatory index is a methodological limitation (Nersesian et al. 2018; Yang et al. 2013), whereas the inclusion of TNF-α, which is rarely studied compared to CRP and IL-6, is a strength (Fagundes et al. 2011; Nersesian et al. 2018). TNF-α is a trigger of inflammatory processes and production of IL-6, which stimulate the production of CRP (Bruunsgaard et al. 2001). Uchino et al. (2018) argued that effect sizes of associations between social relations and inflammation were comparable between types of cytokines.

An alternative approach would be the inclusion of an inflammatory load index instead of relying on single biomarkers. However, our previous attempts to group inflammatory markers into an inflammatory index has not proven to work out better than including the inflammatory markers individually, which is why we decided to study the associations to the inflammatory markers separately (Weddell-Neergaard et al. 2018).

In conclusion, this study find no strong associations between structural and positive as well as negative aspects of functional social relations, respectively, and low-grade inflammation in a large Danish midlife population study. The present study is an important contribution to the field of research elaborating mechanisms behind the impact of social relations on health. The findings suggest that low-grade inflammation is not a primary mechanism between social relations and health.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary material 1 (PDF 169 kb) (169.8KB, pdf)

Acknowledgements

The Copenhagen Aging and Midlife Biobank was Funded by generous Grants from the VELUX FOUNDATION (VELUX26145 and 31539). The authors thank the staff at Department of Public Health and National Research Centre for the Working Environment, who undertook the data collection. Further thanks to Kirsten Avlund†, Helle Bruunsgaard, Nils-Erik Fiehn, Åse Marie Hansen, Poul Holm-Pedersen, Rikke Lund, Erik Lykke Mortensen and Merete Osler, who initiated and established the Copenhagen Aging and Midlife Biobank from 2009 to 2011. The authors acknowledge the crucial role of the initiators and steering groups of The Metropolit Cohort, The Copenhagen Perinatal Cohort and The Danish Longitudinal Study on Work, Unemployment and Health. The Centre for Physical Activity Research is supported by a grant from TrygFonden.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

Footnotes

Responsible editor: Susanne Iwarsson.

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. Aartsen M, Veenstra M, Hansen T. Social pathways to health: on the mediating role of the social network in the relation between socio-economic position and health. SSM Popul. Health. 2017;3:419–426. doi: 10.1016/j.ssmph.2017.05.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Avlund K, et al. Copenhagen Aging and Midlife Biobank (CAMB): an introduction. J Aging Health. 2014;26(1):5–20. doi: 10.1177/0898264313509277. [DOI] [PubMed] [Google Scholar]
  3. Bajaj A, John-Henderson NA, Cundiff JM, Marsland AL, Manuck SB, Kamarck TW. Daily social interactions, close relationships, and systemic inflammation in two samples: healthy middle-aged and older adults. Brain Behav. Immun. 2016;58:152–164. doi: 10.1016/j.bbi.2016.06.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Brown EG, Gallagher S, Creaven AM. Loneliness and acute stress reactivity: a systematic review of psychophysiological studies. Psychophysiology. 2018;55(5):e13031. doi: 10.1111/psyp.13031. [DOI] [PubMed] [Google Scholar]
  5. Bruunsgaard H, Pedersen M, Pedersen BK. Aging and proinflammatory cytokines. Curr. Opin. Hematol. 2001;8(3):131–136. doi: 10.1097/00062752-200105000-00001. [DOI] [PubMed] [Google Scholar]
  6. Cerhan JR, Wallace RB. Predictors of decline in social relationships in the rural elderly. Am. J. Epidemiol. 1993;137(8):870–880. doi: 10.1093/oxfordjournals.aje.a116748. [DOI] [PubMed] [Google Scholar]
  7. Cevenini E, Monti D, Franceschi C. Inflamm-ageing. Clin. Nutr. Metabol. Care. 2013;16(1):14–20. doi: 10.1097/MCO.0b013e32835ada13. [DOI] [PubMed] [Google Scholar]
  8. Chiang JJ, Eisenberger NI, Seeman TE, Taylor SE. Negative and competitive social interactions are related to heightened proinflammatory cytokine activity. Proc. Natl. Acad. Sci. USA. 2012;109(6):1878–1882. doi: 10.1073/pnas.1120972109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Christensen U, et al. Addressing social inequality in aging by the Danish occupational social class measurement. J. Aging Health. 2014;26(1):106–127. doi: 10.1177/0898264314522894. [DOI] [PubMed] [Google Scholar]
  10. Dich N, et al. Early life adversity potentiates the effects of later life stress on cumulative physiological dysregulation. Anxiety Stress Coping. 2015;28(4):372–390. doi: 10.1080/10615806.2014.969720. [DOI] [PubMed] [Google Scholar]
  11. Donoho CJ, Crimmins EM, Seeman TE. Marital quality, gender, and markers of inflammation in the MIDUS Cohort. J. Marriage Fam. 2013;75(1):127–141. doi: 10.1111/j.1741-3737.2012.01023.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Due P, Holstein B, Lund R, Modvig J, Avlund K. Social relations: network, support and relational strain. Soc. Sci. Med. 1999;48(5):661–673. doi: 10.1016/S0277-9536(98)00381-5. [DOI] [PubMed] [Google Scholar]
  13. Eisenberger NI, Moieni M, Inagaki TK, Muscatell KA, Irwin MR. In sickness and in health: the co-regulation of inflammation and social behavior. Neuropsychopharmacology. 2017;42(1):242–253. doi: 10.1038/npp.2016.141. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Ekmann A, Avlund K, Osler M, Lund R. Do negative aspects of social relations influence fatigue? A cross-sectional study on a non-clinical sample of middle-aged Danish men. J. Psychosom. Res. 2012;73(4):277–282. doi: 10.1016/j.jpsychores.2012.08.005. [DOI] [PubMed] [Google Scholar]
  15. Elliot AJ, Heffner KL, Mooney CJ, Moynihan JA, Chapman BP. Social relationships and inflammatory markers in the Midus cohort: the role of age and gender differences. J. Aging Health. 2018;30(6):904–923. doi: 10.1177/0898264317698551. [DOI] [PubMed] [Google Scholar]
  16. Ershler WB, Keller ET. Age-associated increased interleukin-6 gene expression, late-life diseases, and frailty. Ann. Rev. Med. 2000;51:245–270. doi: 10.1146/annurev.med.51.1.245. [DOI] [PubMed] [Google Scholar]
  17. Fagundes CP, Bennett JM, Derry HM, Kiecolt-Glaser JK. Relationships and inflammation across the lifespan: social developmental pathways to disease. Soc. Pers. Psychol. Compass. 2011;5(11):891–903. doi: 10.1111/j.1751-9004.2011.00392.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Ford ES, Loucks EB, Berkman LF. Social integration and concentrations of C-reactive protein among US adults. Ann. Epidemiol. 2006;16(2):78–84. doi: 10.1016/j.annepidem.2005.08.005. [DOI] [PubMed] [Google Scholar]
  19. Ford J, et al. Social integration and quality of social relationships as protective factors for inflammation in a nationally representative sample of black women. J. Urb. Health Bull. New York Acad. Med. 2019;96(Suppl 1):35–43. doi: 10.1007/s11524-018-00337-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Friedman EM, Herd P. Income, education, and inflammation: differential associations in a national probability sample (The MIDUS study) Psychosom. Med. 2010;72(3):290–300. doi: 10.1097/PSY.0b013e3181cfe4c2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Glei DA, Goldman N, Ryff CD, Lin YH, Weinstein M. Social relationships and inflammatory markers: an analysis of Taiwan and the US. Soc. Sci. Med. 2012;74(12):1891–1899. doi: 10.1016/j.socscimed.2012.02.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Heap J, Fors S, Lennartsson C. Coexisting disadvantages in later life: demographic and socio-economic inequalities. J. Popul. Ageing. 2017;10(3):247–267. doi: 10.1007/s12062-016-9158-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Holt-Lunstad J. Why social relationships are important for physical health: a systems approach to understanding and modifying risk and protection. Ann. Rev. Psychol. 2018;69:437–458. doi: 10.1146/annurev-psych-122216-011902. [DOI] [PubMed] [Google Scholar]
  24. Holt-Lunstad J, Smith TB, Layton JB. Social relationships and mortality risk: a meta-analytic review. PLoSMed. 2010;7(7):e1000316. doi: 10.1371/journal.pmed.1000316. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kamiya Y, Whelan B, Timonen V, Kenny RA. The differential impact of subjective and objective aspects of social engagement on cardiovascular risk factors. BMC Geriatr. 2010;10:81. doi: 10.1186/1471-2318-10-81. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Kiecolt-Glaser JK, Gouin JP, Hantsoo L. Close relationships, inflammation, and health. Neurosci. Biobehav. Rev. 2010;35(1):33–38. doi: 10.1016/j.neubiorev.2009.09.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kim DA, Benjamin EJ, Fowler JH, Christakis NA (2016) Social connectedness is associated with fibrinogen level in a human social network. In: Proceedings biological sciences, vol 283(1837), 10.1098/rspb.2016.0958 [DOI] [PMC free article] [PubMed]
  28. Krause N, Borawski-Clark E. Social class differences in social support among older adults. Gerontologist. 1995;35(4):498–508. doi: 10.1093/geront/35.4.498. [DOI] [PubMed] [Google Scholar]
  29. Lara J, et al. A proposed panel of biomarkers of healthy aging. BMC Med. 2015;13:222. doi: 10.1186/s12916-015-0470-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Liu RS, et al. Socioeconomic status in childhood and C reactive protein in adulthood: a systematic review and meta-analysis. J. Epidemiol. Commun. Health. 2017;71(8):817–826. doi: 10.1136/jech-2016-208646. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Loucks EB, Berkman LF, Gruenewald TL, Seeman TE. Relation of social integration to inflammatory marker concentrations in men and women 70 to 79 years. Am. J. Cardiol. 2006;97(7):1010–1016. doi: 10.1016/j.amjcard.2005.10.043. [DOI] [PubMed] [Google Scholar]
  32. Loucks EB, Sullivan LM, D’Agostino RB, Sr, Larson MG, Berkman LF, Benjamin EJ. Social networks and inflammatory markers in the Framingham Heart Study. J. Biosoc. Sci. 2006;38(6):835–842. doi: 10.1017/s0021932005001203. [DOI] [PubMed] [Google Scholar]
  33. Loucks EB, et al. Association of educational level with inflammatory markers in the Framingham Offspring Study. Am. J. Epidemiol. 2006;163(7):622–628. doi: 10.1093/aje/kwj076. [DOI] [PubMed] [Google Scholar]
  34. Lund R, Rod NH, Christensen U. Are negative aspects of social relations predictive of angina pectoris? A 6-year follow-up study of middle-aged Danish women and men. J. Epidemiol. Commun. Health. 2010 doi: 10.1136/jech.2009.106153. [DOI] [PubMed] [Google Scholar]
  35. Lund R, Nielsen LS, Henriksen PW, Schmidt L, Avlund K, Christensen U. Content validity and reliability of the Copenhagen Social Relations Questionnaire (CSRQ) J. Aging Health. 2013;26(1):128–150. doi: 10.1177/0898264313510033. [DOI] [PubMed] [Google Scholar]
  36. Lund R, Christensen U, Nilsson CJ, Kriegbaum M, Hulvej Rod N. Stressful social relations and mortality: a prospective cohort study. J. Epidemiol. Commun. Health. 2014;68(8):720–727. doi: 10.1136/jech-2013-203675. [DOI] [PubMed] [Google Scholar]
  37. Lund R, Rod NH, Thielen K, Nilsson CJ, Christensen U. Negative aspects of close social relations and 10-year incident ischaemic heart disease hospitalization among middle-aged Danes. Eur. J. Prev. Cardiol. 2014;21(10):1249–1256. doi: 10.1177/2047487313486041. [DOI] [PubMed] [Google Scholar]
  38. Lund R, et al. Cohort profile: the Copenhagen aging and midlife Biobank (CAMB) Int. J. Epidemiol. 2015 doi: 10.1093/ije/dyv149. [DOI] [PubMed] [Google Scholar]
  39. McDade TW, Lindau ST, Wroblewski K. Predictors of C-reactive protein in the national social life, health, and aging project. J. Gerontol. Ser. B Psychol. Sci. Soc. Sci. 2011;66(1):129–136. doi: 10.1093/geronb/gbq008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Mezuk B, Diez Roux AV, Seeman T. Evaluating the buffering vs direct effects hypotheses of emotional social support on inflammatory markers: the multi-ethnic study of atherosclerosis. Brain Behav. Immun. 2010;24(8):1294–1300. doi: 10.1016/j.bbi.2010.06.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Nakata A, Irie M, Takahashi M. Source-specific social support and circulating inflammatory markers among white-collar employees. Ann. Behav. Med. Publ. Soc. Behav. Med. 2014;47(3):335–346. doi: 10.1007/s12160-013-9555-x. [DOI] [PubMed] [Google Scholar]
  42. Nersesian PV, et al. Loneliness in middle age and biomarkers of systemic inflammation: findings from Midlife in the United States. Soc. Sci. Med. 2018;209:174–181. doi: 10.1016/j.socscimed.2018.04.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Pedersen JM, et al. Late midlife C-reactive protein and interleukin-6 in middle aged danish men in relation to body size history within and across generations. Obesity (Silver Spring, Md) 2016;24(2):461–468. doi: 10.1002/oby.21311. [DOI] [PubMed] [Google Scholar]
  44. Puustinen PJ, Koponen H, Kautiainen H, Mantyselka P, Vanhala M. Psychological distress and C-reactive protein: do health behaviours and pathophysiological factors modify the association? Eur. Arch. Psychiatry Clin. Neurosci. 2011;261(4):277–284. doi: 10.1007/s00406-010-0134-x. [DOI] [PubMed] [Google Scholar]
  45. Rook KS. Social networks in later life: weighing positive and negative effects on health and well-being. Curr. Dir. Psychol. Sci. 2015;24(1):45–51. doi: 10.1177/0963721414551364. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Stepanikova I, Bateman LB, Oates GR. Systemic inflammation in midlife: race, socioeconomic status, and perceived discrimination. Am. J. Prev. Med. 2017;52(1s1):S63–S76. doi: 10.1016/j.amepre.2016.09.026. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Uchino BN, et al. Ethnic/racial differences in the association between social support and levels of C-reactive proteins in the North Texas Heart Study. Psychophysiology. 2016;53(1):64–70. doi: 10.1111/psyp.12499. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Uchino BN, Trettevik R, Kent de Grey RG, Cronan S, Hogan J, Baucom BRW. Social support, social integration, and inflammatory cytokines: a meta-analysis. Health Psychol. Off. J. Div. Health Psychol. Am. Psychol. Assoc. 2018;37(5):462–471. doi: 10.1037/hea0000594. [DOI] [PubMed] [Google Scholar]
  49. Walen HR, Lachman ME. Social support and strain from partner, family, and friends: costs and benefits for men and women in adulthood. J. Soc. Pers. Relationsh. 2000;17(1):5–30. doi: 10.1177/0265407500171001. [DOI] [Google Scholar]
  50. Wang X, et al. Inflammatory markers and risk of type 2 diabetes: a systematic review and meta-analysis. Diabetes Care. 2013;36(1):166–175. doi: 10.2337/dc12-0702. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Weddell-Neergaard, et al. Cardiorespiratory fitness and the metabolic syndrome: roles of inflammation and abdominal obesity. PLoS One. 2018;13(3):e0194991. doi: 10.1371/journal.pone.0194991. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Weyers S, et al. Low socio-economic position is associated with poor social networks and social support: results from the Heinz Nixdorf Recall Study. Int. J. Equity Health. 2008;7:13. doi: 10.1186/1475-9276-7-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Wilson RS, Boyle PA, James BD, Leurgans SE, Buchman AS, Bennett DA. Negative social interactions and risk of mild cognitive impairment in old age. Neuropsychology. 2015;29(4):561–570. doi: 10.1037/neu0000154. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Yang YC, McClintock MK, Kozloski M, Li T. Social isolation and adult mortality: the role of chronic inflammation and sex differences. J. Health Soc. Behav. 2013;54(2):183–203. doi: 10.1177/0022146513485244. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Yang YC, Schorpp K, Harris KM. Social support, social strain and inflammation: evidence from a national longitudinal study of U.S. adults. Soc. Sci. Med. 2014;107:124–135. doi: 10.1016/j.socscimed.2014.02.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Yang YC, Boen C, Gerken K, Li T, Schorpp K, Harris KM. Social relationships and physiological determinants of longevity across the human life span. Proc. Natl. Acad. Sci. USA. 2016;113(3):578–583. doi: 10.1073/pnas.1511085112. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary material 1 (PDF 169 kb) (169.8KB, pdf)

Articles from European Journal of Ageing are provided here courtesy of Springer

RESOURCES