Coping is undoubtedly one of the most extensively researched concepts in behavioral medicine. How people react to environmental challenges, as well as health-related hardship to reduce psychological distress, is a function of the type of the stressor and of an individual’s coping styles which may include thoughts, emotions, and behaviors [
Coping styles can be assessed in many ways [
The association between SSS and biological indicators of health may depend upon the quality of the received support [
The overarching aim of this study was to further elucidate the relationship between coping and biological indicators of health. For this purpose we investigated the relationship between coping strategies, assessed by the WCCL-R and circulating biomarkers of cardiovascular health in a community sample of elderly dementia caregivers and noncaregiving controls. An investigation of these associations in the elderly is clinically important since cardiovascular disease (CVD) becomes more prevalent with aging, and elderly dementia caregivers have a particularly increased risk of developing CVD, particularly CHD [
Increases in such biomarkers could be a function of the type of the stressor (e.g., caregiving, low socioeconomic status, life events, health-related problems) with which one needs to cope and of the unique emotional (e.g., negative affect), behavioral (e.g., exercise frequency), and physical (e.g., sympathetic activation) responses to such a stressor. For instance, if caregivers seek out for social support strongly, but will not receive it, this might elicit depressed mood that has been associated with a proinflammatory state [
The University of California San Diego (UCSD) Institutional Review Board approved the study protocol and all participants provided written consent. For the present study, we analyzed cross-sectional data obtained at study entry of the UCSD “Alzheimer’s Caregiver Study” investigating effects of dementia caregiving stress on health of elderly spousal caregivers. Participants were recruited through referrals from the UCSD Alzheimer’s Disease Research Center, community support groups and agencies serving caregivers, local senior citizen health fairs, and referrals from other participants. Inclusion criteria were being
Out of 186 enrolled subjects (126 caregivers, 60 noncaregivers), 151 had complete data for the 12 assessed biomarkers. One subject each missed data on coping, body mass index (BMI), and dyslipidemia, and 12 subjects missed data on norepinephrine (NEPI) levels. This yielded a sample of 136 subjects (93 caregivers, 43 noncaregivers) with a complete dataset for the present investigation allowing us to compute full linear regression approach. The 45 subjects with incomplete data did not significantly differ from the 136 subjects with complete data for sociodemographic factors, CVD risk factors, caregiving status, and ways of coping. All participants were interviewed in their homes using questionnaires to assess demographic factors, mood states, stressors, and health status.
We collected information on gender, age, and years of education to define socioeconomic status.
We asked participants for their weight and height to calculate BMI.
Plasma low-density lipoprotein cholesterol (LDL-C) and high-density lipoprotein cholesterol (HDL-C) were determined by standard methodology at the clinical chemistry laboratories at the UCSD medical center. We computed the LDL-C/HDL-C ratio with a higher ratio indicating greater dyslipidemia.
Using a noninvasive Microlife blood pressure (BP) monitor, three BP and heart rate measurements were collected by the research nurse over a 15-minute resting period. The participant’s mean systolic resting BP was used for the analysis because it confers higher CVD risk than diastolic BP in individuals over 50 years of age [
Smoking status was defined in terms of ever (i.e., former plus current) smoker versus never smoker (only one participant was a current smoker).
Participants were provided a list with 17 health problems they might currently have or that a doctor had informed them of having. Positive items were added to one
The amount of alcohol consumption was assessed using a score that considered the number of days subjects had at least one alcoholic drink and the number of alcoholic drinks they usually drank on these days considering the last month.
The Rapid Assessment of Physical Activity (RAPA) scale was used to assess the amount of light, moderate, and strenuous physical activities, including strength and flexibility exercises, in a typical week (total score 0–6) [
The Pittsburgh Sleep Quality Index was used to assess subjective sleep quality, sleep duration, sleep latency, sleep disturbances, sleep efficacy, use of sleep medication, daytime dysfunction (global score between 0 and 21; higher scores indicate poorer sleep quality) [
Participants completed the WCCL-R [
We used the Positive and Negative Affect Scale comprising 10 items per mood scale covering the past few weeks on a 5-point scale (
We used the Life Events Survey which assesses how many of a list of 34 events had occurred to the participant or a close relative or friend in the past year [
The 8-item Social Support Scale was used to assess help and support participants received from friends and relatives [
A highly sensitive catechol-o-methyltransferase (COMT)-based radioenzymatic assay was performed to determine plasma NEPI [
Blood was collected in the participant’s home at 10:30 AM. In order not to interfere with study participants daily routine, fasting state was not a prerequisite but was treated as a control variable. Plasma was stored at
We selected these 12 biomarkers because all of them are variously involved in the initiation, progression, and clinical manifestation of atherothrombotic CVD. They specifically cover atherosclerotic processes related to inflammation (proinflammatory cytokines: TNF-
Data were analyzed using PASW 18.0 statistical software package (SPSS Inc., Chicago, IL, USA) with
To test for a relation between coping and biomarkers, we first employed multivariate analysis of covariance (MANCOVA) to test whether each of the five coping scales would be significantly associated with the group of the 12 biomarkers as a whole applying Bonferroni correction (
Table
Participant characteristics.
Variables | All participants ( |
Caregivers ( |
Noncaregivers ( |
|
---|---|---|---|---|
Age (years) |
|
|
|
0.937 |
Women (%) | 67.6 | 67.7 | 67.4 | 0.972 |
Education (years) |
|
|
|
0.722 |
Body mass index (kg/m2) |
|
|
|
0.447 |
LDL-C/HDL-C ratio |
|
|
|
0.904 |
Systolic blood pressure (mm Hg) |
|
|
|
0.432 |
Heart rate (beats/min) |
|
|
|
0.802 |
Ever smoker (%) | 40.4 | 43.0 | 34.9 | 0.369 |
Diabetes (%) | 9.6 | 12.9 | 2.3 | 0.062 |
Any cardiovascular disease (%) | 14.7 | 19.4 | 4.7 | 0.035 |
Number of health problems |
|
|
|
0.001 |
Alcohol consumption (score) |
|
|
|
0.908 |
Rapid Assessment of Physical Activity (score) |
|
|
|
0.005 |
Pittsburgh Sleep Quality Index |
|
|
|
<0.001 |
Problem-focused coping (score) |
|
|
|
0.702 |
Seeks social support (score) |
|
|
|
0.027 |
Blamed self (score) |
|
|
|
0.626 |
Wishful thinking (score) |
|
|
|
0.020 |
Avoidance coping (score) |
|
|
|
0.098 |
Negative affect (score) |
|
|
|
0.001 |
Positive affect (score) |
|
|
|
<0.001 |
Number of life events |
|
|
|
0.108 |
Perceived social support (score) |
|
|
|
0.004 |
Fasting state (%) | 15.4 | 12.9 | 20.9 | 0.307 |
Norepinephrine (pg/mL) |
|
|
|
0.885 |
Tumor necrosis factor- |
5.76 (4.11–7.64) | 5.79 (4.12–7.70) | 5.41 (4.09–7.66) | 0.603 |
Interleukin-6 (pg/mL) | 1.12 (0.80–1.67) | 1.07 (0.84–1.53) | 1.30 (0.75–1.86) | 0.189 |
Interleukin-8 (pg/mL) | 6.91 (4.22–9.54) | 6.88 (4.15–9.11) | 7.07 (4.25–10.1) | 0.267 |
Interferon- |
1.61 (0.96–2.39) | 1.59 (1.06–2.41) | 1.74 (0.80–2.29) | 0.348 |
Serum amyloid A (mg/mL) | 2.44 (1.13–7.39) | 2.53 (1.14–7.30) | 2.09 (1.01–8.06) | 0.933 |
C-reactive protein (mg/mL) | 1.42 (0.81–3.92) | 1.49 (0.85–4.16) | 1.24 (0.69–3.05) | 0.422 |
Intercellular adhesion molecule-1 (ng/mL) | 295 (217–489) | 275 (208–486) | 316 (224–490) | 0.480 |
Vascular cellular adhesion molecule-1 (ng/mL) | 541 (369–950) | 545 (365–956) | 539 (397–930) | 0.842 |
Endothelin-1 (pg/mL) | 1.10 (0.81–1.40) | 1.08 (0.80–1.40) | 1.15 (0.84–1.42) | 0.514 |
von Willebrand Factor (%) | 158 (83–270) | 156 (92–269) | 144 (75–287) | 0.457 |
D-dimer (ng/mL) | 630 (452–954) | 671 (487–980) | 613 (415–957) | 0.300 |
Plasminogen activator inhibitor-1 (ng/mL) | 25.9 (15.6–43.3) | 29.1 (16.7–47.8) | 21.1 (12.6–33.6) | 0.149 |
Nonnormal distribution (even after log transformation)/all biomarkers (IVs) were log transformed/shown as median (IQR) Fischer Definition?
We first performed MANCOVA to test for a significant relationship between each of the five coping scales and the group of 12 biomarkers as a whole. All analyses controlled for age, gender, BMI, LDL-C/HDL-C ratio, systolic BP, smoking status, diabetes, any CVD, and fasting state. We found a significant relationship between SSS and the group of biomarkers as a whole (
In MANCOVA, SSS showed significant relationships with IL-8 (
The multivariate linear regression models for individual biomarkers are shown in Table
Multivariate linear regression model for the relationship between seeking social support and individual biomarkers.
Entered variables | Interleukin-8 | Serum amyloid A | C-reactive protein | Soluble VCAM-1 | D-dimer | |||||
---|---|---|---|---|---|---|---|---|---|---|
Partial corr. |
|
Partial corr. |
|
Partial corr. |
|
Partial corr. |
|
Partial corr. |
|
|
Age | 0.036 | 0.688 | −0.001 | 0.989 | 0.041 | 0.649 | 0.124 | 0.164 | 0.368 |
|
Female gender | 0.078 | 0.385 | 0.142 | 0.111 | 0.167 | 0.061 | −0.064 | 0.471 | 0.169 | 0.058 |
Body mass index | 0.053 | 0.552 | 0.203 |
|
0.215 |
|
0.123 | 0.169 | 0.014 | 0.873 |
LDL-C/HDL-C ratio | −0.017 | 0.846 | −0.008 | 0.925 | 0.111 | 0.214 | 0.026 | 0.774 | −0.006 | 0.951 |
Systolic blood pressure | −0.065 | 0.469 | 0.158 | 0.077 | 0.175 |
|
0.067 | 0.4522 | 0.046 | 0.604 |
Ever smoker | 0.016 | 0.860 | 0.009 | 0.924 | 0.062 | 0.492 | −0.013 | 0.884 | 0.059 | 0.508 |
Diabetes | 0.076 | 0.398 | −0.078 | 0.381 | −0.095 | 0.288 | −0.113 | 0.207 | 0.015 | 0.867 |
Any cardiovascular disease | −0.105 | 0.240 | −0.076 | 0.395 | −0.080 | 0.369 | 0.044 | 0.623 | 0.094 | 0.295 |
Fasting state | −0.206 |
|
0.033 | 0.716 | 0.013 | 0.881 | 0.097 | 0.280 | −0.051 | 0.566 |
Seeking social support | −0.264 |
|
0.289 |
|
0.223 |
|
0.204 |
|
0.190 |
|
| ||||||||||
Model statistic |
|
|
|
|
|
|||||
|
|
|
|
| ||||||
|
|
|
|
|
Partial corr.: partial correlation coefficient, significant
All meditational analyses were adjusted for age, gender, BMI, LDL-C/HDL-C ratio, systolic BP, smoking status, diabetes, any CVD, and fasting state. Chronic stress (caregiving, socioeconomic status, life events, health problems), affect (negative affect, positive affect, perceived social support), health behaviors (alcohol consumption, physical activity, sleep), and autonomic activity (NEPI, heart rate) did not emerge as significant mediators of the association of SSS with IL-8, SAA, CRP, sVCAM-1, and D-dimer (data not shown).
All interaction effects were adjusted for age, gender, BMI, LDL-C/HDL-C ratio, systolic BP, smoking status, diabetes, any CVD, fasting state, SSS (main effect), and the respective moderator variable (main effect).
There were significant interactions between SSS and caregiver status for D-dimer (
There were significant interactions between SSS and positive affect for CRP (
There were no significant interactions between SSS on the one hand and physical activity (
There was a significant interaction between NEPI and SSS for IL-8 (
We found that greater use of SSS was associated with elevated levels of several circulating biomarkers, all of which are proxy measures of an increased risk of atherothrombotic CVD in elderly community-dwelling subjects. This relationship was independent of sociodemographic and CVD risk factors that are known to affect biomarker levels. Specifically, we found greater use of SSS to be associated with increased levels of SAA, CRP, sVCAM-1, and D-dimer. We did not find that PFC, BS, WT, and AC were significantly related to biomarkers. The finding that greater use of SSS is associated with elevated levels of biomarkers of atherothrombotic risk is a novel one. Moreover, this observation is consistent with a previous meta-analysis showing that greater use of SSS was associated with negative physical health outcomes [
A variety of mechanisms might be involved in this pathway. The acute phase reactant SAA is expressed by human adipocytes and atherosclerotic lesions and may play a critical role in local and systemic inflammation by linking obesity and atherosclerosis [
To gain more insight into the mechanisms linking greater use of SSS with biomarkers, we performed a series of exploratory analyses in which we tested whether chronic stress, affect, health behaviour, and autonomic activity might possibly act as mediator or moderator variables. None of these variables were suggested as mediators; however, there may be other mediators that are relevant, but not tested in our study. For instance, a life-time approach to experienced psychosocial stress (i.e., considering stress burden for longer than one year), inclusion of daily hassles, account of dietary habits, and more thorough autonomic measures (e.g., heart rate variability) might have also be important to investigate.
We found that the seemingly counterintuitive inverse association between greater use of SSS and decreased levels of the proinflammatory marker IL-8 was moderated by life events and NEPI. Specifically, greater use of SSS was associated with decreased IL-8 in subjects with few life events, although it would seem that cardiovascular health might particularly draw benefit from such an effect if life stress is high. Moreover, greater use of SSS was also associated with decreased IL-8 when NEPI levels were high. As NEPI increases the IL-8 release from the endothelium [
These interaction effects suggest that help seeking is not uniformly bad for cardiovascular health when assessed through biomarkers, so elderly individuals clearly should not be discouraged from seeking help. In fact, entire interventions are based on the idea that for instance dementia caregivers need to seek out additional support from their social environment in order to better cope with caregiving burden [
From a clinical perspective, our data may imply that encouragement of elderly individuals by clinicians about seeking out support to possibly improve their cardiovascular health needs to consider that not all friends and family will be perceived as sources of support in which case greater use of SSS might not be helpful if not harmful. This might particularly be the case for those elderly low in positive affect and low in perceived social support. One recommendation could be that elderly individuals are advised to disengage in time from seeking support from certain members of their social network should they realize they will not receive the expected amount and quality of emotional and instrumental help. Whether such a strategy would be effective in improving cardiovascular health might be validated in an intervention study aimed at changing coping behavior where a decrease in the use of SSS should be associated with a concomitant decrease in SAA, CRP, and D-dimer levels. However, these implications must consider that the interactions between SSS and some covariates were not observed for all biomarkers. Moreover, the number of exploratory analyses conducted was substantial. Therefore, the clinical conclusions and recommendations form these secondary findings should be made with caution.
We discuss three limitations of our study. Firstly, cross-sectional investigations capture only a snapshot of biobehavioral processes. Coping processes are particularly actuated when challenge occurs at which time biomarker response would seem to be greatest. Therefore, experimental induction of acute stressful situations might be a more promising way to investigate which coping styles are either functional or dysfunctional in terms of moderating the stress response of biomarkers. Secondly, biomarkers are intermediate endpoints of CVD. Although similar to that explained by age and BMI, the variances in biomarkers that were explained by SSS ranged 4–9%. Whether this effect size translates into a clinically relevant increase in atherothrombotic CVD risk needs to be seen in a longitudinal study. Thirdly, our results stem from elderly with on average good physical and mental health, two thirds of whom were dementia caregivers. Hence, the findings may not be generalizable to younger populations or elderly with greater prevalence of CVD and frailty.
The findings from our study suggest that elderly community-dwelling individuals with increased use of SSS may have elevated circulating levels of a range of biomarkers that have been associated with an increased risk of atherothrombotic CVD. Several moderating variables relating to the type of chronic stress, affect, and sympathomedullary activity need to be considered in this relationship. If replicated, these results might offer promising avenues for interventions focused on healthy coping strategies that might also improve cardiovascular health in the elderly. These might include provision of better practical and emotional support, as well as cognitive-behavioral interventions to modify perceptions of burden and support.
This study was supported by Award AG 15301 from the National Institutes of Health/National Institute on Aging (NIH/NIA) to I. Grant, M.D. Additional support was provided through award AG 03090 to B. Mausbach, Ph.D., and AG 08415 to S. Ancoli-Israel, Ph.D. The authors wish to thank Susan Calleran, M.A., and Christine Gonzaga, R.N., for data collection.