Abstract
OBJECTIVE:
To test whether depressive symptoms are related to subsequent C-reactive protein (CRP) levels and/or whether CRP levels are related to subsequent depressive symptoms in mid-life women.
METHODS:
Women enrolled in the Study of Women's Health Across the Nation (SWAN) were followed for seven years and had measures of CES-Depression scores and CRP seven times during the follow-up period. Women were pre- or early peri-menopausal at study entry and were of Caucasian, African American, Hispanic, Japanese, or Chinese race/ethnicity. Analyses were restricted to initially healthy women.
RESULTS:
Longitudinal mixed linear regression models adjusting for age, race, site, time between exams, and outcome variable at year X showed that higher CES-D scores predicted higher subsequent CRP levels and vice versa over a 7-year period. Full multivariate models adjusting for body mass index, physical activity, medications, health conditions, and other covariates showed that higher CRP levels at year X predicted higher CES-D scores at year X+1, p = 0.03. Higher depressive symptoms predicted higher subsequent CRP levels at marginally significant levels, p=0.10.
CONCLUSIONS:
Higher CRP levels led to higher subsequent depressive symptoms, albeit the effect was small. The study demonstrates the importance of considering bi-directional relationships for depression and other psychosocial factors and risk for heart disease.
Keywords: Depression, inflammation, menopause, women, longitudinal, C-reactive protein
Depression is associated with an increased risk of mortality and recurrent morbidity in patients with coronary heart disease (CHD) and depression may increase the risk of the initial onset of CHD in healthy populations (Rozanski et al., 1999; Suls and Bunde, 2005; Wulsin and Singal, 2003). In reporting the findings from these studies, the standard risk factors for CHD are statistically controlled for, and therefore do not preclude alternative risk factors or other intermediate pathways as explanatory factors or mediators in the associations between depression and CHD.
Among the potential pathways that may be relevant to understanding depression's risk for CHD morbidity and mortality is chronic inflammation (Carney et al., 1995), which is thought to be important throughout the natural history of coronary atherosclerosis from the initial defect in endothelial function through plaque rupture (Ross, 1999). A marker of chronic inflammation that is a strong predictor of clinical coronary events is C-reactive protein (CRP) (Pearson et al., 2003).
A recent meta-analysis of 51 cross-sectional studies, published between 1967 and 2008, found that depression was associated concurrently with CRP levels (Howren et al., 2009). Subanalyses showed that (a) the relationship was significant in men but not in women in those studies that stratified by gender; (b) the effect was larger in clinic-based as opposed to community-based samples; and (c) the studies that did not adjust for body mass index (BMI) reported associations three times larger than those studies that did adjust for BMI. That obesity might also moderate the association between depression and some inflammatory markers was not examined in the meta-analysis. However, in men enrolled in MONICA (Ladwig et al., 2003), depression and anxiety were related to CRP but only among those who were obese, suggesting stratification by obesity status may be informative. In summary, depression is a reliable correlate of CRP levels, with weaker associations apparent in women, in community samples, and in analyses that consider BMI.
Recently it has been suggested that inflammation can produce symptoms resembling those of depression, including anhedonia, vegetative symptoms, and sleeping difficulties (Dantzer et al., 2009; Raison et al., 2006). Patients treated with interferon, which induces inflammation, develop depressive symptoms that can be ameliorated by antidepressants (Capuron et al., 2002). In experimental studies of healthy participants, endotoxin administration leads to short-term increases in anxiety, depressed mood, and poor memory performance (Reichenberg et al., 2001). These findings suggest that inflammation may not only be a consequence of depressive symptoms but also contribute to its presentation.
In the present report, we addressed the bi-directional and temporal relation of depressive symptoms and CRP levels across 7 years among women enrolled in the Study of Women's Health Across the Nation (SWAN). Specifically, we addressed two questions: Are women with higher depressive symptoms likely to have higher CRP levels at the next annual visit, adjusted for the CRP levels at the prior visit? And, are women with higher CRP levels likely to have higher depressive symptoms at the next annual visit, adjusted for depressive symptom levels at the prior visit? Secondarily, we explore the role of obesity in understanding any observed relationships because obesity is a strong correlate of CRP.
METHODS
Participants
SWAN is a longitudinal, community-based study of 3,302 pre- and early perimenopausal women followed through the menopausal transition and beyond (Sowers et al., 2000). Eligibility for enrollment was assessed between November 1995 and October 1997 as part of a telephone survey of health, reproductive, lifestyle, and demographic information of 16,065 women conducted by the seven study sites. From this sample, each site recruited approximately 450 eligible women composed of Caucasian women and women from one specified minority group (African Americans in Boston, MA, Chicago, IL, Pittsburgh, PA, and Detroit area, MI; Chinese in the Oakland/East Bay region, CA; Hispanic women in Newark, NJ, and Japanese in Los Angeles, CA). Eligibility criteria included: age 42 to 52 years, intact uterus, a menstrual period and no reproductive hormone use in the three months prior to the baseline interview, and self-identification with the site's designated race/ethnic groups. The Institutional Review Boards at all participating sites approved the study protocol.
The analytic cohort for the present study was based on 1,781 women from the SWAN longitudinal cohort (n=3302). We excluded 255 women with missing data on the following: the Center for Epidemiologic Studies Depression (CES-D) (n =2) scale (Radloff, 1977); CRP levels (N=8); or without data on at least two consecutive visits (including n=181 with only CES-D baseline data and n=64 with nonconsecutive visits). Having at least two consecutive visits is a requirement for lagged analyses to examine the influence of values at one exam on a subsequent exam's values. We also excluded data from 1,236 women who, at baseline, had one or more of the following health conditions: stroke/heart condition (including stroke, myocardial infarction, heart attack, angina, and medication for heart disease, antihypertensives and/or anticoagulants), diagnosis or treatment for diabetes, arthritis/osteoarthritis, and/or osteoporosis, and/or taking inhaler/steroids. Excluding women who already had chronic health conditions associated with elevated CRP at baseline provided a clearer estimate of the effect of depressive symptoms on change in CRP over time and vice versa. After the baseline examination, if women were diagnosed with any of these conditions or reported using the above medications, their data were retained in the analyses and these variables were treated as time-varying covariates. Finally, data from 30 women whose available CRP results were all >10 mg/L were excluded. The New Jersey site did not retain women beyond year 5 because of administrative reasons unrelated to the purpose of the study. Study retention rate at the end of the seventh follow-up examination was 73%.
Procedures
SWAN participants at all seven sites were assessed at study entry (baseline) and annually thereafter with a common protocol. All study forms and materials were available in English, Cantonese, Japanese, and Spanish, and bilingual staff was used, as appropriate. Baseline and annual assessments included self-and interviewer-administered questionnaires about health, psychosocial, and lifestyle factors. Standardized protocols were used for annual measurements of height, weight and waist circumference by certified staff. The fasting blood draw was targeted to the follicular phase of the menstrual cycle (days 2–5) in menstruating women and before 10 AM. All samples were maintained at 4°C until prepared for freezing at 80°C; they were later shipped on dry ice to a central laboratory.
Measures
C-Reactive Protein
CRP was quantified using an ultrasensitive rate immunonepholemetry method (Dade-Behring, Marburg, Germany). For analyses, blood samples were recorded as occurring between 8:00 AM and 10:00 AM (morning blood draw) versus not; and whether the participant had fasted 12 hours before the blood draw (fasting blood draw) versus not. These data were recorded because SWAN protocol was set up to accommodate the needs of all the analytes, some of which vary by fasting status and some by time of day. CRP was assayed annually with the exception of follow-up visit 2, due to financial constraints. Thus, participants had up to seven CRP values from years 0, 1, 3, 4, 5, 6, and 7 available for analysis.
Depressive Symptoms
Depressive symptoms were assessed at each year with the CES-D Scale, a 20-item measure that asks about the frequency of being bothered by depressive symptoms during the previous week on a scale of 0 (rarely) to 3 (most or all of the time) for a total range of 0 to 60 (Radloff, 1977). This scale was developed to assess depressive symptoms in epidemiological studies and has been used in multiethnic cohorts (Guarnaccia et al., 1989; Roberts, 1980; Ying, 1988). In our sample, Cronbach's α coefficient was 0.90.
Other Covariates / Variables
Information about race/ethnicity and education (≤ high school degree, some college/vocational training, college degree or more) was obtained at the baseline examination. Age and information on each of the following variables was obtained at each SWAN examination. Questions on medication use (verified by checking containers when possible) and self-reported health history since the previous examination were combined into four categories. For each health category, participant status was classified as having the condition, not having the condition or ‘unknown’ (i.e., having incomplete data for that condition). We identified four health conditions of interest: cardiovascular including stroke, heart attack, angina, high blood pressure or taking heart disease related medications (e.g., anticoagulants, antihypertensive medication); diabetes or use of insulin; arthritis/osteoarthritis or use of arthritis medication or over-the-counter pain medication; and thyroid condition or use of steroids. BMI (kg/m2) was calculated from measurements of weight and height, which were obtained using calibrated scales and stadiometers. Waist circumference was measured at the level of the natural waist or the narrowest part of the torso from the anterior aspect. Smoking status was classified as current versus not current smoker; and total physical activity was based on activity from housework and leisure activities (Sternfeld et al., 1999). Medication for nerves/depression was defined as use at least 2 times per week in the past month. Sleep problems (trouble falling asleep, waking up several times a night, and waking up earlier than planned and being unable to fall asleep again) was based on the number of problems that were reported as occurring at least three times per week (Kravitz et al., 2008). Hormone therapy use since the previous annual visit was self-reported. Menopausal status was based on menstrual bleeding patterns in the previous 12 months and was categorized as: a) premenopausal = menstrual period in the past 3 months with no change in regularity in the past 12 months; b) early perimenopausal = menstrual period in the past 3 months with some change in regularity over the previous 12 months; c) late perimenopausal = no menstrual period within the past 3 months, but some menstrual bleeding within the past 12 months; d) post menopausal = no menstrual period within the past 12 months; e) surgical menopause = hysterectomy or bilateral oophorectomy; f) indeterminant menopausal status = used hormone therapy before final menses or surgical menopause, thus, status could not be determined. Based on SWAN entry eligibility requirements, all women were pre- or early perimenopausal at baseline.
Data Analyses
CRP and CES-D scores were not normally distributed. Both were log transformed and scaled for a non-negative value by adding 1. Both depression scores and CRP levels showed good consistency across all visits (CES-D=Intraclass correlation coefficient (ICC and 95% confidence interval (CI)=0.48 (0.46, 0.50); CRP=ICC (95%CI)=0.62 (0.60, 0.64)) within the analytic sample. Analysis used CRP values ≤10 (6% of all CRP values were excluded), to minimize the possibility of including an acute infection (Pearson et al., 2003). As noted earlier, 30 women were excluded because all their values were >10; otherwise, only specific observations >10 at any particular SWAN visit were dropped, consistent with prior research (Stewart et al., 2009). Because waist circumference was highly correlated with BMI (r=0.91), a residualized waist circumference variable was defined from the linear regression model fitting waist to BMI.
Longitudinal associations between CES-D scores and CRP were evaluated using multivariable longitudinal linear mixed regression models with a (woman-specific) random intercept. Exploratory analyses and standard model fit statistics indicated that an auto-regressive moving average error correlation structure was more suitable than simpler alternatives (Yerbeke and Molenberghs, 2000). Model assumptions of normality were confirmed via diagnostic plots, while ordinal analyses and scatter plots verified the use of linear analyses.
Two sets of models were computed so that the association between depressive symptoms and CRP levels could be examined for temporal directionality. In the first set of models, the dependent variable was CRP levels at the subsequent visit (time X+1) with the CES-D score at the prior visit (time X) as the independent variable and CRP at time X included as a control variable. In the second set of models the associations were reversed such that CRP at time X was the independent variable and the CES-D score at the subsequent visit (time X+1) was the dependent variable, while controlling for CES-D score at time X. Thus, data from two consecutive visits per woman were required for inclusion in analyses. The same observations and covariates determined a priori from the literature were used for modeling in each direction to assure comparability of other covariates in the models. Covariates were taken from time X with the exceptions of morning blood draw and fasting blood draw, which were also taken from time X+1, and education, race/ethnicity and study site, which were not time-varying. Initially, models included covariates for study site, race/ethnicity, baseline age, and time between X and X+1. Then, fully adjusted models were computed that also included baseline education and time-varying covariates of fasting blood draw at X and X+1, morning blood draw at X and X+1, menopausal status, hormone use, physical activity, smoking, the disease status groups, BMI, waist circumference residual, and medication for nerves/depression. Thus, four models were computed – two initial and two full models in each direction. Because CRP levels were not available at follow-up 2, all data from this visit were censored and data from follow-up 1 were used to predict follow-up 3.
Partial correlation coefficients were estimated with each model to provide an indication of the strength of association when comparing models reflecting opposing directions. Models were computed separately, so beta coefficients and p values could not be directly compared. Exploratory analyses included testing for interactions between the predictor variable and BMI (<30 vs. ≥30), or visit number (to detect influence of aging). Number of sleep problems was added as covariate; it was not in the main models because sleep disturbance is a common symptom of depression and the CES-D scale includes one item about sleep. Additionally, comparisons were made between women who were excluded at baseline and those included in the analyses to evaluate whether CRP levels were elevated in the excluded women, as expected; models were also calculated in the women with baseline health conditions for exploratory purposes. Analyses were computed using SAS (Version 9.1, SAS Institute, Inc., Cary, NC).
RESULTS
Sample Characteristics
Table 1, column 2 shows the baseline characteristics of the analytic sample. Participants ranged in age from 42 to 52 years; almost half of the sample was overweight or obese. By design, the sample was composed of almost half Caucasians and one quarter African Americans, with smaller proportions of Chinese, Japanese, and Hispanic women. Forty-eight percent had a college degree or more and 21% had a high school degree or less; and 14% were current smokers. Baseline CRP and CES-D scores were positively associated though the association was small (Spearman's r=0.06, p=0.02).
Table 1.
Baseline Characteristics of Analytic Sample Compared to Women who were Excluded from Analysis
| Baseline Characteristics | Sample (n=1781) |
Excluded (n=1521) |
P-Value |
|---|---|---|---|
| Median (IQR) CES-D | 7 (3, 14) | 9 (4, 17) | <0.0001 |
| Mean (SD) Age (years) | 46.2 ± 2.6 | 46.5 ± 2.7 | 0.003 |
| Median (IQR) hs-CRP (mg/l) | 1.1 (0.5, 2.9) | 2.7 (0.9, 7.7) | <0.0001 |
| Mean (SD) Body Mass Index (kg/m2) | 26.4 ± 5.8 | 30.5 ± 8.1 | <0.0001 |
| N (%) Body Mass Index (kg/m2) | <0.0001 | ||
| Obese (≥30) | 391 (22) | 685 (46) | |
| Overweight (25-30) | 499 (28) | 377 (25) | |
| Normal/Underweight (<25) | 875 (50) | 433 (29) | |
| N (%) Race/Ethnicity | <0.0001 | ||
| African American | 393 (22) | 542 (36) | |
| Caucasian | 900 (51) | 650 (43) | |
| Chinese | 176 (10) | 74 (5) | |
| Hispanic | 104 (6) | 182 (12) | |
| Japanese | 208 (12) | 73 (5) | |
| N (%) Education | <0.0001 | ||
| Less than high school / high school | 370 (21) | 449 (30) | |
| Some college / technical school | 541(31) | 510 (34) | |
| College graduate / post graduate degree | 854 (48) | 547 (36) | |
| N (%) Current smoker | 252 (14) | 317 (21) | 0.0002 |
Women in the analytic sample had lower CES-D scores and CRP levels at baseline compared to those excluded from the analysis (Table 1). In addition, women in the analytic sample were thinner, better educated, less likely to smoke, and more likely to be Caucasian, Japanese, and Chinese and less likely to be African American or Hispanic.
The Bi-directional Influence of Depressive Symptoms and C-Reactive Protein
The initial model showed that higher CES-D scores at time X were positively associated with higher CRP at time X+1, β=0.009, p=0.02, partial r = 0.054. The same pattern of results was observed in the fully adjusted lagged model, although the association was no longer significant (p=0.10, partial r = 0.039, Table 2). Significant covariates in the fully adjusted model were CRP at time X, race, educational level, current smoker, taking medications for nerves/depression in past month, low physical activity, non-fasting X+1, increased age/greater time between visits, high BMI, and high waist circumference residual.
Table 2.
Multivariate longitudinal linear mixed models of the association predicting CRP or CES-D at time X + 1 (dependent variable) from CES-D or CRP, respectively at time X (independent variable).
| Covariate/estimated parameter | Outcome/dependent variable | |||
|---|---|---|---|---|
| CRP at time X + 1 | Depression at time X +1 |
|||
| b | p-Value | b | p-Value | |
| CES-D at time X | .008a | .10 | .717 | <.0001 |
| CRP at time X | .850 | <.0001 | .037a | .03 |
| Age (years) | .002 | .18 | _.001 | .70 |
| Time between visits (years) | .0002 | <.0001 | .0001 | .38 |
| Study site | .33 | <.0001 | ||
| Race/ethnicity (Caucasian as reference) | _.001 | .0004 | _.009 | .0001 |
| African American | _.038 | .92 | .019 | .71 |
| Chinese | .050 | .03 | _.042 | .59 |
| Hispanic | _.064 | .28 | .155 | .60 |
| Japanese | .104 | <.0001 | — | <.0001 |
| Fasting (for 12 h) at X + 1 | _.038 | .005 | .044 | — |
| Fasting (for 12 h) | _.006 | .30 | — | .45 |
| Morning blood draw (8–10AM) at X +1 | .027 | .68 | .015 | — |
| Morning blood draw (8–10AM) | .001 | .07 | .031 | .49 |
| Baseline education (college or more reference) |
.04 | .21 | ||
| High school education or less | .023 | .91 | .028 | .18 |
| Greater than high school/some college | _.031 | .01 | .079 | .13 |
| Status (premenopause reference) | .22 | .02 | ||
| Early perimenopause | _.013 | .02 | .043 | .0005 |
| Late perimenopause | _.010 | .58 | .049 | .29 |
| Post menopause | _.011 | .62 | .079 | .15 |
| Hysterectomy &/or oophorectomy | _.008 | .70 | .021 | .15 |
| Status unknown | .024 | .80 | .031 | .70 |
| Current hormone therapy user | .042 | .31 | .016 | .46 |
| Current smoker | .003 | .55 | ||
| Heart condition, medication (no reference) | .22 | .04 | ||
| Unknown | .008 | .65 | _.066 | .02 |
| Yes | .027 | .08 | .017 | .54 |
| Thyroid condition, steroid medication | .14 | .33 | ||
| Unknown | .147 | .16 | .147 | .38 |
| Yes | _.020 | .17 | .033 | .22 |
| Osteoarthritis, arthritis/OTC pain medication |
.67 | .004 | ||
| Unknown | .009 | .53 | .072 | .002 |
| Yes | _.001 | .93 | .072 | .007 |
| Diabetes, insulin medication | .51 | .42 | ||
| Unknown | _.106 | .65 | .120 | .75 |
| Yes | _.028 | .29 | .059 | .20 |
| Nerves/depression medication 2 _/week in past month |
.079 | .01 .04 |
.031 | .57 .59 |
| Physical activity | _.005 | _.003 | ||
| Body mass index | .009 | <.0001 | _.002 | .22 |
| Waist circumference residual | .013 | .009 | .018 | .07 |
All covariates are time-varying and from time X unless otherwise noted, N = 1714. CES-D, Center for Epidemiologic Studies-Depression; CRP, C-reactive protein; OTC, over-the-counter.
Predictor/independent variable; predictor and outcome variables are logged.
The initial model showed that higher CRP level at time X was positively associated with higher CES-D scores at time X+1, β=0.031, p=0.02, partial r = 0.054. The lagged relation remained significant in the fully adjusted model (p = .03, partial r = 0.053; Table 2). Significant covariates in the fully adjusted model were depressive symptoms at time X, race, study site, menopausal status, heart conditions/medications, and osteoarthritis/pain medications.
Additional Analyses
We repeated the initial and full models adjusted for number of sleep problems. CES-D was not a significant predictor of CRP level in either model, ps > 0.18; indication of sleep problems was not a significant covariate in the full model, β = 0.018, p = 0.09. CRP remained a significant predictor of CES-D scores in the initial model, β = 0.028, p = 0.04 and full models β = 0.036, p = 0.03; indication of sleep problems was a significant covariate in the full model, β = 0.080, p < 0.0001.
Tests for interactions between predictor variables and obese/nonobese were nonsignificant in the multivariate models, ps > 0.35. Tests for interactions between predictor variables and time, which examined whether later values (older age/additional visits) may have influenced the results, were also nonsignificant, ps > 0.35. Models conducted in only the women who had exclusionary health conditions at baseline showed no significant directional effects in initial or full models, ps > 0.30.
DISCUSSION
The purpose of the study was to ask whether higher depressive symptoms led subsequently to higher CRP levels, and whether higher CRP levels led subsequently to higher depressive symptoms in middle-aged women who were healthy at baseline We found evidence in our initial models for a bi-directional relationship between depressive symptoms and CRP levels over seven years of follow-up in an ethnically diverse sample of middle-aged women. In multivariate models adjusted for a substantial number of covariates known to influence CRP or depression or both, including health conditions, medications, ethnicity, age, menopausal status, and life style factors, we found that CRP remained a modest, albeit statistically significant, predictor of subsequent depressive symptoms, whereas the reverse association (CES-D→CRP) only approached conventional levels of significance. Adiposity is a strong correlate of CRP so any longitudinal influence of psychosocial factors in relation to CRP may be masked in obese women. However, formal tests for interactions with obesity were nonsignificant. We previously reported that among SWAN women followed for 5 years (with four observations of CRP, i.e. at years 0, 1, 3, and 5 and including women with health problems at baseline), CES-D scores did not predict concurrent CRP levels over time in a risk factor-adjusted model (Matthews et al., 2007). Note that depression and CRP were unrelated in women in cross-sectional studies that stratified by gender (Howren et al., 2009). Taken together, our lagged and concurrent analyses suggest that the temporal relationship and the direction of the influence are apparent from CRP levels to subsequent depressive symptoms.
To our knowledge, there are three longitudinal reports relevant to our study. In the Leiden 85-plus Study, CRP levels predicted increases in depressive symptoms over 5 years in those not initially depressed or cognitively impaired (van den Biggelaar et al., 2007). The Whitehall II study of over 3,000 civil servants tested for bi-directionality between CRP and a 4-item measure of cognitive symptoms of depression that were measured twice, 12 years apart (Gimeno et al., 2008). The cognitive symptom measure included feelings of hopelessness, worthlessness, life not worth living, and not able to do anything because of nerves. In fully adjusted models, which included health conditions, lifestyle factors, obesity, and other covariates, baseline CRP levels predicted subsequent cognitive symptoms of depression, adjusted for initial levels of depressive symptoms, but cognitive symptoms of depression did not predict subsequent levels of CRP or IL-6. Tests for interactions with gender were nonsignificant. Finally, among 263 healthy men and women, depressive symptoms predicted IL-6 six years later, but not CRP in structural equation models. In that sample depressive symptoms were not correlated with CRP measured concurrently at either assessment, although baseline CRP tended to predict change in depressive symptoms, p = 0.06 (Stewart et al., 2009). In sum, our results are consistent with most of the available literature suggesting that higher CRP levels lead to greater depressive symptoms.
The study also revealed several other interesting findings. Strong predictors of later depressive symptoms, adjusted for earlier symptoms, were sleep problems, osteoarthritis and pain medications, and ethnicity; strong predictors of later CRP levels, adjusted for earlier levels, were BMI, waist circumference (independent of BMI), current smoking, and ethnicity. Furthermore, women who were excluded from the primary analytic sample primarily because of health conditions present at baseline already had higher CRP and CES-D scores at baseline than had the women who did not have exclusionary health conditions; no directional effects of depressive symptoms to subsequent CRP or vice versa were observed in these women.
Our study has several limitations. The sampling frequency was driven by the overall study design as opposed to biological considerations of how long the duration of exposure to depression or elevated CRP would be required to have an influence. Second, we have not measured history of clinical depression in the full sample. Third, the findings are restricted to women around the time of the menopause. The study also has several strengths, including the multi-ethnic sample of women, thorough assessment of time-varying covariates, excellent retention of women across 7 years, and multiple measures of CRP and depressive symptoms.
In conclusion, our study shows that higher CRP levels are modestly related to higher subsequent depressive symptoms over a 7 year period, findings consistent with the Whitehall II study, Leiden 85-plus study, and experimental data. Given that our findings demonstrate that inflammation can influence depressive symptoms, which is a risk factor for CHD, efforts to evaluate the dynamic and bi-directional relationships between other psychosocial risk factors for CHD and putative pathways may also be worthwhile.
ACKNOWLEDGMENTS
The Study of Women's Health Across the Nation (SWAN) has grant support from the National Institutes of Health (NIH), DHHS, through the National Institute on Aging (NIA), the National Institute of Nursing Research (NINR) and the NIH Office of Research on Women's Health (ORWH) (Grants NR004061; AG012505, AG012535, AG012531, AG012539, AG012546, AG012553, AG012554, AG012495). The content of this paper is solely the responsibility of the authors and does not necessarily represent the official views of the NIA, NINR, ORWH or the NIH.
We thank the study staff at each site and all the women who participated in SWAN.
Glossary
- CHD
Coronary Heart Disease
- CRP
C-Reactive Protein
- BMI
Body Mass Index
- MONICA
Monitoring trends and determinants of cardiovascular diseases
- SWAN
Study of Women's Health Across the Nation
- CES-D
Center for Epidemiologic Studies-Depression
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
NIH Program Office: National Institute on Aging, Bethesda, MD - Marcia Ory 1994 – 2001; Sherry Sherman 1994 – present; National Institute of Nursing Research, Bethesda, MD – Program Officers.
Central Laboratory: University of Michigan, Ann Arbor - Daniel McConnell (Central Ligand Assay Satellite Services).
Coordinating Center: New England Research Institutes, Watertown, MA - Sonja McKinlay, PI 1995 – 2001; University of Pittsburgh, Pittsburgh, PA – Kim Sutton-Tyrrell, PI 2001 – present.
Steering Committee Chair: Chris Gallagher, Susan Johnson
All authors report no conflicts of interest.
Contributor Information
Karen A. Matthews, University of Pittsburgh, 3811 O'Hara Street, Pittsburgh, PA 15213, matthewska@upmc.edu
Laura L. Schott, University of Pittsburgh, 130 DeSoto Street, Pittsburgh, PA 15261, schottll@upmc.edu
Joyce T. Bromberger, University of Pittsburgh, 3811 O'Hara Street, Pittsburgh, PA 15213, brombergerjt@upmc.edu
Jill M. Cyranowski, University of Pittsburgh, 3811 O'Hara Street, Pittsburgh, PA 15213, cyranowskijm@upmc.edu
Susan A. Everson-Rose, University of Minnesota, 717 Delaware Street SE, Minneapolis, MN 55414, saer@umn.edu
MaryFran Sowers, University of Michigan, 109 Observatory, Ann Arbor, MI 48109-2205, mfsowers@umich.edu
References
- Capuron L, Gumnick JF, Musselman DL, Lawson DH, Reemsnyder A, Nemeroff CB, Miller AH. Neurobehavioral effects of interferon-alpha in cancer patients: phenomenology and paroxetine responsiveness of symptom dimensions. Neuropsychopharmacology. 2002;26:643–652. doi: 10.1016/S0893-133X(01)00407-9. [DOI] [PubMed] [Google Scholar]
- Carney RM, Freedland KE, Rich MW, Jaffe AS. Depression as a risk factor for cardiac events in established coronary heart disease: A review of possible mechanisms. Ann Behav Med. 1995;17:142–149. doi: 10.1007/BF02895063. [DOI] [PubMed] [Google Scholar]
- Dantzer R, O'Connor JC, Freund GG, Johnson RW, Kelley KW. From inflammation to sickness and depression: when the immune system subjugates the brain. Nat Rev Neurosci. 2008;9:46–56. doi: 10.1038/nrn2297. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gimeno D, Kivimaki M, Brunner EJ, Elovainio M, DeVogli R, Steptoe A, Kumari M, Lowe GD, Rumley A, Marmot MG, Ferrie JE. Associations of C-reactive protein and interleukin-6 with cognitive symptoms of depression: 12-year follow-up of the Whitehall II study. Psychol Med. 2009;3:413–423. doi: 10.1017/S0033291708003723. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Guarnaccia PJ, Angel R, Worobey JL. The factor structure of the CES-D in the Hispanic Health and Nutrition Examination Survey: the influences of ethnicity, gender and language. Soc Sci Med. 1989;29:85–94. doi: 10.1016/0277-9536(89)90131-7. [DOI] [PubMed] [Google Scholar]
- Howren MB, Lumkin DM, Suls J. Associations of depression with C-reactive protein, IL-1, and IL-6: A meta-analysis. Psychosom Med. 2009;71:171–186. doi: 10.1097/PSY.0b013e3181907c1b. [DOI] [PubMed] [Google Scholar]
- Kravitz HM, Zhao X, Bromberger JT, Gold EB, Hall MH, Matthews KA, Sowers MR. Sleep disturbance during the menopausal transition in a multi-ethnic community sample of women. Sleep. 2008;31:979–990. [PMC free article] [PubMed] [Google Scholar]
- Ladwig K, Marten-Mittag B, Lowel H, Doring A, Koenig W. Influence of depressive mood on the association of CRP and obesity in 3205 middle aged healthy men. Brain Behav Immun. 2003;17:268–275. doi: 10.1016/s0889-1591(03)00056-4. [DOI] [PubMed] [Google Scholar]
- Matthews KA, Schott LL, Bromberger J, Cyranowski J, Everson-Rose SA, Sowers MF. Associations between depressive symptoms and inflammatory/hemostatic markers in women during the menopausal transition. Psychosom Med. 2007;69:124–130. doi: 10.1097/01.psy.0000256574.30389.1b. [DOI] [PubMed] [Google Scholar]
- Pearson TA, Mensah GA, Alexander RW, Anderson JL, Cannon RO, Criqui M, Fadl YY, Fortman SP, Hong Y, Myers GL, Rifai N, Smith SC, Taubert K, Tracy RP, Vinicor F. Markers of inflammation and cardiovascular disease: Application to clinical and publich health practice. Circulation. 2003;107:499–511. doi: 10.1161/01.cir.0000052939.59093.45. [DOI] [PubMed] [Google Scholar]
- Radloff LS. The CES-D Scale: A self-report depression scale for research in the general population. Appl Psychol Meas. 1977;1:385–401. [Google Scholar]
- Raison CL, Capuron L, Miller AH. Cytokines sing the blues: inflammation and the pathogenesis of depression. Trends Immunol. 2006;27:24–31. doi: 10.1016/j.it.2005.11.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Reichenberg A, Yirmiya R, Schuld A, Kraus T, Haack M, Morag A, Pollmächer T. Cytokine-associated emotional and cognitive disturbances in humans. Arch Gen Psychiatry. 2001;58:445–452. doi: 10.1001/archpsyc.58.5.445. [DOI] [PubMed] [Google Scholar]
- Roberts RE. Reliability of the CES-D Scale in different ethnic contexts. Psychiatry Res. 1980;2:125–134. doi: 10.1016/0165-1781(80)90069-4. [DOI] [PubMed] [Google Scholar]
- Ross R. Atherosclerosis--an inflammatory disease. N Engl J Med. 1999;340:115–126. doi: 10.1056/NEJM199901143400207. [DOI] [PubMed] [Google Scholar]
- Rozanski A, Blumenthal JA, Kaplan J. Impact of psychological factors on the pathogenesis of cardiovascular disease and implications for therapy. Circulation. 1999;99:2192–2217. doi: 10.1161/01.cir.99.16.2192. [DOI] [PubMed] [Google Scholar]
- Sowers MF, Crawford S, Sternfeld B, Morganstein D, Gold E, Greendale G, Evans D, Neer R, Matthews KA, Sherman S, Lo A, Weiss G, Kelsey J. SWAN: a multi-center, multi-ethnic community-based cohort study of women and the menopausal transition. In: Lobo R, Marcus R, Kelsey J, editors. Menopause: Biology and Pathology. Academic Press; New York: 2000. [Google Scholar]
- Sternfeld B, Ainsworth BE, Quesenberry CP. Physical activity patterns in a diverse population of women. Prev Med. 1999;28:313–323. doi: 10.1006/pmed.1998.0470. [DOI] [PubMed] [Google Scholar]
- Stewart JC, Rand KL, Muldoon MF, Kamarck TW. A prospective evaluation of the directionality of the depression-inflammation relationship. Brain Behav Immun. 2009 doi: 10.1016/j.bbi.2009.04.011. in press. doi:10.1016/j.bbi.2009.04.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Suls J, Bunde J. Anger, anxiety, and depression as risk factors for cardiovascular disease: the problems and implications of overlapping affective dispositions. Psychol Bull. 2005;131:260–300. doi: 10.1037/0033-2909.131.2.260. [DOI] [PubMed] [Google Scholar]
- van den Biggelaar AHJ, Gussekloo J, de Craen AJM, Frolich M, Stek ML, van der Mast RC, Westerndeoprt RGJ. Inflammation and interleukin-1 signaling network contribute to depressive symptoms but not cognitive decline in old age. Exp Gerontol. 2007;42:693–701. doi: 10.1016/j.exger.2007.01.011. [DOI] [PubMed] [Google Scholar]
- Wulsin LR, Singal BM. Do depressive symptoms increase the risk for the onset of coronary disease? A systematic quantitative review. Psychosom Med. 2003;65:201–210. doi: 10.1097/01.psy.0000058371.50240.e3. [DOI] [PubMed] [Google Scholar]
- Yerbeke G, Molenberghs G. Linear mixed models for longitudinal data. Springer; New York: 2000. [Google Scholar]
- Ying YW. Depressive symptomatology among Chinese-Americans as measured by the CES-D. J Clin Psychol. 1988;44:739–746. doi: 10.1002/1097-4679(198809)44:5<739::aid-jclp2270440512>3.0.co;2-0. [DOI] [PubMed] [Google Scholar]