| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Department of Society, Human Development & Health, Harvard School of Public Health, Boston, Massachusetts (E.B.L., L.F.B.); and the Division of Geriatrics, Department of Medicine, UCLA School of Medicine, Los Angeles, California (T.L.G., T.E.S.).
Address correspondence and reprint requests to Eric B. Loucks, PhD, Department of Society, Human Development and Health, Harvard School of Public Health, 677 Huntington Ave., Boston, MA. E-mail: eloucks{at}hsph.harvard.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Methods: Participants (ages 70–79; n = 375 men and 425 women) were part of the MacArthur Successful Aging Study, a longitudinal study from three community-based cohorts in the United States, who have relatively high physical and cognitive functioning. Social integration was measured using a social integration score, which assessed marital status, number of contacts with family and friends, frequency of religious service attendance, and participation in voluntary organizations. Fibrinogen concentrations were measured in plasma using an automated clot-rate assay. Cross-sectional multivariate logistic regression analyses were performed.
Results: Social integration was significantly associated with elevated concentrations of fibrinogen (>336 mg/dL) in men after adjusting for smoking, alcohol consumption, physical activity, body mass index, comorbidity, physical functioning, depression, age, race, and education (odds ratio [OR] = 2.29, 95% confidence interval [CI] = 1.07–4.89, p = .03 for having elevated fibrinogen in the least integrated quartile versus the most integrated quartile). There was no significant association between social integration and fibrinogen in women (multivariate-adjusted OR = 0.57, 95% CI = 0.27–1.21, p = .15).
Conclusions: Social integration is associated with fibrinogen concentrations in elderly men. This provides evidence of a biologic mechanism that may help explain the observed associations between social integration and mortality in men. There may be gender differences in the physiological pathways by which social integration influences health.
Key Words: fibrinogen • social integration • social isolation • elderly
Abbreviations: OR = odds ratio; CI = confidence interval.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Acute mental stress and chronic stress have been associated with elevated concentrations of fibrinogen in a number of studies (4). Fibrinogen is a soluble protein that is normally circulated in plasma of the blood. It is an important component of the blood-clotting pathway and has a number of functions, including to increase the speed of platelet aggregation and thrombus formation. Elevated fibrinogen levels predict larger thrombi within blood vessels and formation of tight and rigid network structures within the thrombi themselves (5). Thrombosis (the formation of clots in blood vessels) is a major cause of myocardial infarction and stroke. In fact, concentrations of fibrinogen consistently predict coronary heart disease outcomes in prospective studies independently of standard cardiovascular risk factors such as smoking, cholesterol, and blood pressure (6). A metaanalysis of 18 prospective studies demonstrated that elevated fibrinogen concentrations predicted coronary heart disease with an odds ratio of 1.8 (95% confidence interval [CI] = 1.6–2.0) for the upper tertile of fibrinogen compared with the lower tertile (6). This work identifies a potential biologic mechanism by which social integration might influence cardiovascular disease morbidity and mortality. Some studies have investigated the association between social integration and fibrinogen concentration in younger populations (
65 years) (7–11); however, to our knowledge, the relation between social integration and fibrinogen remains unknown in the elderly. Furthermore, the independent association between fibrinogen and social integration is not well understood. Specifically, in younger populations, the association between social integration and fibrinogen has been found in some but not all studies after adjustment for covariates (7–11).
In this study, we assessed the association between social integration and plasma concentrations of fibrinogen in a relatively high functioning cohort of 800 men and women 70 to 79 years of age who were part of the MacArthur Successful Aging Study. Because gender differences have been reported for the influence of social integration on mortality (12), we tested the association between fibrinogen concentration and social networks in gender-specific analyses. Furthermore, to determine if the association between social integration and fibrinogen is independent of potentially confounding or mediating pathways, we perform multivariate analyses adjusting for smoking status, alcohol consumption, physical activity, body mass index, comorbidity, physical functioning, socioeconomic status, and depression.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Measures
Social Networks
A social network score was computed by summing six dichotomous indicators, creating a possible score range of 0 to 6: 1) presence of a spouse (absent = 0, present = 1), 2) number of close relatives (zero or one relatives = 0, two or more relatives = 1), 3) number of close friends (zero or one friend = 0, two or more friends = 1), 4) frequency of participation in religious services (never or less than monthly = 0, monthly or more frequently = 1), 5) frequency of participation in religious activities other than religious services (never = 0, sometimes or more frequently = 1), and 6) frequency of participation in clubs and other voluntary associations (never = 0, sometimes or more frequently = 1). This score was divided into quartiles (lowest quartile = 0–2, second quartile = 3, third quartile = 4, highest quartile = 5–6). The measure was adapted from the social network index in the Alameda County Study, which has been shown to predict mortality in a number of cohorts (2,14,15).
Fibrinogen
Fibrinogen was measured in blood samples collected by phlebotomists in participants’ homes at approximately 8 am on the day after the in-home interview. Samples were drawn in heparinized tubes and plasma was prepared and frozen for storage at –80°C within 4 hours of blood collection. Stored plasma samples were analyzed by Dr. Russel Tracy’s laboratory (University of Vermont at Burlington) in duplicate, with fibrinogen concentration averages being reported. Fibrinogen was measured by automated clot-rate assay based on the original method of Clauss using the ST4 instrument (Diagnostica Stago Inc., Parsippany, NJ) with standardization using the College of American Pathologists (CAP) reference material (16). Analytical coefficient of variation for the assay is 1.7%; the expected normal range is 163 to 363 mg/dL.
Covariates
Covariates included age, race, smoking, alcohol consumption, physical activity, body mass index, depression, education, comorbidity, and physical functioning. Age was measured as of interview date. Race was coded black or white. Socioeconomic status was measured as educational level, which was coded 
high school degree versus <high school degree. Depression was assessed using the Hopkins Symptom Checklist depression subscale. Body mass index (kg/m2) was measured as a continuous variable. Smoking was coded as never smoked, exsmoker, or current smoker. Alcohol consumption was coded as nondrinker, light drinker, or moderate drinker and assessed using two dummy coded variables. Cut points were as follows: nondrinker: did not drink in the last year; light drinker: drank within the last year, but less than 1 oz in the last month (mean = 0.6 oz, standard deviation [SD] = 0.13); or moderate drinker: drank greater than 1 oz within the last month (mean = 12.2 oz, SD = 12.6). Self-reported physical activity was assessed using questions adapted from the Yale Physical Activity Survey (17) assessing frequency of current leisure- and work-related activity. Each activity mentioned was classified as light, moderate, or strenuous based on intensity codes (kcal/min) adapted from Paffenbarger et al. (18) and Taylor et al. (19). Summary scales were derived by multiplying the frequency of activity (five categories, ranging from never to 3+ times per week) by the intensity code and summing over all activities within a given category of intensity. Data are presented for summary indices of strenuous and moderate physical activity. Comorbidity was measured as the number of chronic conditions (four groups: 0 conditions, 1, 2, or 3 conditions). Physical functioning was measured by performance scores in five physical domains (balance, lower extremity strength, gait, lower extremity dexterity, hand dexterity). A summary measure of physical performance was developed by summing scores for the five subscales as described in detail elsewhere (20).
Statistical Analyses
The association between social network level and fibrinogen was examined with separate logistic regression models stratified by gender, predicting elevated concentrations of fibrinogen in the upper quartile of fibrinogen concentration (>336 mg/dL). Three models were examined: 1) social network quartiles as predictors; 2) sociodemographic (age, education, race), comorbid health conditions, and physical functioning covariates on the first step, social network quartiles entered on the second step; and 3) additional psychologic, behavioral, or physical covariates, including depression, smoking status, alcohol consumption, body mass index, and physical activity entered at the second step of the model, and social network quartiles entered on the third step. Model (3) enabled identification of the extent to which depression as well as behavioral and physical covariates (including smoking, alcohol consumption, body mass index, and physical activity) accounted for the strength of association between social integration and fibrinogen. Relative odds of being in the highest quartile of fibrinogen concentration were estimated comparing those in each of the three lowest social network quartiles with the reference group (i.e., those in the highest social network quartile). The proportions of variation explained by social integration and the covariates in each model were calculated using the Nagelkerke R Square calculation (21).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
|
Descriptive statistics for the association of the covariates (age, smoking, alcohol consumption, physical activity, body mass index, comorbidity, physical functioning, race, depression, and education) with elevated fibrinogen concentration are shown in Table 2 (derived using a multivariate model including all covariates and social network score). Smoking, body mass index, comorbidity, and physical function in women, and being an ex-smoker in men, were significantly (p < .05) associated with elevated fibrinogen concentrations.
|
Figure 1 demonstrates mean concentrations of fibrinogen according to social network level. In men, fibrinogen levels fall with increasing levels of social integration above the lowest two social network quartiles. In women, there was no graded association.
|
Logistic regression analyses demonstrated that males in the two most socially isolated quartiles were more than twice as likely to have elevated fibrinogen concentrations compared with the most socially integrated quartile (Table 3). Adjusting for age, race, education, comorbidity, and physical functioning (model 2) slightly increased the strength of association. Inclusion of the covariates physical activity, body mass index, smoking status, alcohol consumption, and depression into the final model (model 3) reduced the strength of these associations only slightly. The lowest two social networks quartiles remained significantly more likely to have elevated fibrinogen concentration (odds ratio [OR] = 2.29, 95% CI = 1.07–4.89, p = .03, and OR = 2.25, 95% CI = 1.09–4.69, p = .03). In women, there was no significant association between social network level and the odds of having elevated fibrinogen concentrations (Table 3).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Several studies have reported that social integration is significantly linked to fibrinogen concentration in younger populations (65 years) (7–11), which is consistent with findings of the present study for older men (results for women are discussed subsequently). After adjusting for cardiovascular risk factors, however, two studies found that the relationship was not statistically significant (8,9). Other studies have found the association to be independent of covariates (7,11), although the particular covariates adjusted for differ between studies. We found two potential explanations for differences in findings compared with the present study. One study in which investigators used a valid but different measure of social integration, which focused more on social participation, showed no significant effects after adjustment for covariates. Authors statistically adjusted for number of people in the household and did not include ties with spouse, which can reduce the statistical effect of having social support at home compared with the present study (8). Thus, this study likely measured less intimate types of social relationships compared with the present report. If it is close personal relationships that are strong predictors of fibrinogen concentration, this could result in different predicted concentrations of fibrinogen in men between the studies. Second, differences between studies may be related to the age of the participants in that all previous studies focused on younger populations (65 years) (7–11). The association between social integration and fibrinogen may be a stronger independent risk in older men compared with younger populations. Future studies using cohorts with both younger and elderly populations will be better able to test this hypothesis.
The present study found no association between social integration and fibrinogen among women in unadjusted or multivariate adjusted analyses. To our knowledge, three other analyses have been performed examining the association between social integration and fibrinogen in women (7,10). One study in 46 women recruited from a community (mean age 34.5 ± 5.7 years) and 42 women who were university students (mean age 21.4 ± 4.1 year) found no significant association between social integration and fibrinogen for the community women but a significant association in the university students (7). Another study in 300 middle-aged Swedish women (aged 31–65 years) found a significant association between social integration and fibrinogen (10). Differences in findings may be related to the different measures of social integration between studies in that one study (7) focused on sources of social support (such as socializing, confiding personal problems, and help when they are ill), whereas the other study used a shortened version of The Interview Schedule for Social Interaction (ISSI), which measures the availability and perceived adequacy of a wide range of social contacts and relationships. These measures address more qualitative aspects of social support as compared with the current study, and those constructs may better predict fibrinogen concentrations in women than the social network score used in the present study. A second source of differences in the findings among women may relate again to age. The present study is the first to our knowledge that investigates the association between social integration and fibrinogen in women over the age of 65 years. Future studies in cohorts with a wide range of ages will be able to better determine the effects of age on the relation between social integration and fibrinogen in women.
There are a number of potential reasons for the gender differences that were observed in this study, with fibrinogen concentration associated with social integration in men and not women. First, certain characteristics of social relationships tend to differ between genders (12,22). For example, it has been shown that women are more likely to serve as caregivers, providers of support, and exercise social control over network members’ health behaviors (22,23) as compared with men. Outcomes, including negative psychologic and physical health, have been shown to occur in those who experience strain as a result of provision of care (24,25). In this study, social integration was measured using an index of social ties that assessed the number of close personal relationships with friends and family, marital status, as well as participation in social groups, clubs, and religious organizations. Our measure of social network size does not measure the balance of positive versus negative, burdensome, or conflict-laden relationships, which may be different for men and women. Future studies investigating such constructs of social integration should provide further insight into potential gender differences. Second, there may be gender-specific physiological pathways through which social integration influences health. A growing body of literature suggests that there are gender differences in the association between social experiences and biomarkers such as blood pressure, norepinephrine, and epinephrine (26,27). Gender differences in fibrinogen concentration have been found in many but not all study populations (28,29).
Social integration may influence fibrinogen through a number of mediating mechanisms, including health behaviors and psychosocial responses, including strain and depression. Evidence suggests that social integration is related to health behaviors; however, our data do not provide evidence that the health behaviors smoking, alcohol consumption, and physical activity (or body mass index, which is in part related to health behaviors) substantively mediate or confound the relation between social integration and fibrinogen. With regard to psychosocial factors as mediating variables between social integration and fibrinogen, social relationships can affect health by regulating emotions and cognitions, which can lead to depressive symptoms and perceived stress (30,31). In turn, stress and depression have been associated with raised fibrinogen concentration, as described in a review by von Känel et al. (4). Findings in this study did not demonstrate depression to account for a substantial amount of the statistical strength of association between social networks and fibrinogen; however, stress remains a possible mediating pathway. This is consistent with the "fight or flight" stress response theory, in that during stressful and potentially harmful situations, factors that promote blood coagulation would be protective from excessive bleeding resulting from injury. Furthermore, studies have demonstrated that fibrinolytic factors (which have the opposite effects of fibrinogen in that they are responsible for dissolving and removing clots, e.g., tissue-type plasminogen activator [t-PA]) are actually impaired during chronic stress as measured by job strain (4,32). This should further augment the potentially injurious prothrombotic effects of fibrinogen during chronic stress. Interestingly, there is also evidence that fibrinolysis is increased during acute stress, which should have the reverse effect and inhibit the prothrombotic effects of fibrinogen during those situations (4). Knowledge on the potential injurious association between social isolation and fibrinogen should be aided by future studies that measure thrombolytic factors (such as t-PA) in addition to prothrombotic factors such as fibrinogen.
It is important to note that behaviors such as smoking and alcohol consumption, and other risk factors such as depression, may be considered as confounders (influencing both social integration and fibrinogen) or as mediators (causal intermediates in the pathway from social integration to fibrinogen response). The data used in these analyses are cross-sectional, so we cannot identify the temporal ordering of the conditions. However, our models have treated the conditions uniformly as confounders, which is a conservative approach to this issue.
In conclusion, this study found that social integration is significantly associated with fibrinogen concentrations in elderly men and not women after adjusting for age, smoking, alcohol consumption, physical activity, body mass index, comorbidity, physical functioning, depression, race, and education. This evidence suggests a biologic mechanism that may help explain the observed associations between social integration and mortality in the elderly.
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
This research was support by NIA grants AG-17056 and AG-17265, grants from the John D. and Catherine T. MacArthur Foundation, a research fellowship from the Harvard Center for Society and Health, and a personnel award from the Heart and Stroke Foundation of Canada.
DOI:10.1097/01.psy.0000160482.89163.e8
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. M. White, G. S. Philogene, L. Fine, and S. Sinha Social Support and Self-Reported Health Status of Older Adults in the United States Am J Public Health, October 1, 2009; 99(10): 1872 - 1878. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. H. Wirtz, L. S. Redwine, U. Ehlert, and R. von Kanel Independent Association Between Lower Level of Social Support and Higher Coagulation Activity Before and After Acute Psychosocial Stress Psychosom Med, January 1, 2009; 71(1): 30 - 37. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Maselko, L. Kubzansky, I. Kawachi, T. Seeman, and L. Berkman Religious Service Attendance and Allostatic Load Among High-Functioning Elderly Psychosom Med, June 1, 2007; 69(5): 464 - 472. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. P. Rieker and C. E. Bird Rethinking Gender Differences in Health: Why We Need to Integrate Social and Biological Perspectives J. Gerontol. B. Psychol. Sci. Soc. Sci., October 1, 2005; 60(suppl_Special_Issue_2): S40 - S47. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
