| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Department of Welfare Promotion and Epidemiology, Toyama Medical and Pharmaceutical University, Toyama, Japan.
Address correspondence and reprint requests to Ali Nasermoaddeli, Department of Welfare Promotion and Epidemiology, Faculty of Medicine, Toyama Medical and Pharmaceutical University, 2630 Sugitani, Toyama 930-0194, Japan. E-mail: moaddeli{at}ms.toyama-mpu.ac.jp
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Methods: Data were collected in 2003 and 2004 cross-sectionally from the two phases of a longitudinal study. Atherosclerotic risk factors were evaluated along with psychosocial characteristics for 1,339 apparently healthy civil servants (748 women and 591 men) aged 20 to 60 years. The Japanese version of the SOC-13 was used in the study.
Results: After controlling for traditional atherosclerotic risk factors, SOC was inversely associated with hsCRP levels only in men. Among the atherosclerotic risk factors, body mass index and high-density lipoprotein cholesterol had the strongest associations with hsCRP in both sexes. Mean arterial pressure and blood glucose were positively correlated with hsCRP in men, and serum triglyceride levels positively correlated with hsCRP in women.
Conclusion: The findings suggested that a higher SOC might be associated with lower serum CRP levels in men. Also, the association of traditional atherosclerotic risk factors with CRP was not equal in men and women.
Key Words: C-reactive protein • atherosclerotic risk factors • psychosocial characteristics • sense of coherence • cross-sectional studies
Abbreviations: hsCRP = high-sensitivity C-reactive protein; SOC = sense of coherence; BMI = body mass index; MAP = mean arterial pressure; HDL = high-density lipoprotein.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
In contrast to the pathogenic orientation that proposes an association between negative emotions and health outcomes (7,8), Antonovsky (9) proposed the concept "sense of coherence" (SOC), which focuses on personality dispositions that promote health. SOC focuses on the factors keeping individuals healthy rather than on factors making them ill. These resources, when strong, include the belief that what happens in life is predictable, structured, and understandable (comprehensibility); that adequate and sufficient resources are perceived to be available to help resolve stressful problems when they arise (manageability); and that these demands are challenges worthy of investment and engagement (meaningfulness) (10). A few prospective studies have reported that a strong SOC is associated with reduced incidence and mortality from cardiovascular diseases (11,12). However, the evaluation of the role of SOC in moderating cardiovascular events through biological mechanisms remains to be tested epidemiologically.
In this study, we address the relationship between SOC as a construct that measures outlook on life against stressful conditions and an inflammatory marker, hsCRP, in a substantial number of men and women. We hypothesized that a strong SOC would associate with a lower CRP.
The lifetime risk of developing coronary heart disease is not equally distributed among the genders, men being more affected than women (13). However, it is not completely clear whether this is only related to a high prevalence of the risk factors in men or to a different susceptibility to these risk factors between the genders, or a combination of both. Investigating the prevalence and associates of strong predictors of cardiovascular diseases in men and women could improve our knowledge on a gender-specific risk assessment. Previous reports have underlined the association of atherosclerotic risk factors and serum CRP, but the data were pooled for men and women in the analyses (14–16). This may underestimate or overestimate the associations due to natural differences in the risk factors between men and women. Gender differences have also been pointed out in the associations between SOC and health (17) and personality and health (18). For these reasons, we check whether gender modifies the association of SOC with CRP, and study the association of atherosclerotic risk factors and psychosocial characteristics with serum CRP in a substantial number of male and female subjects.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Biological Markers
Height, weight, blood pressure, total cholesterol, high-density lipoprotein (HDL) cholesterol, triglyceride, and blood sugar were measured during the annual health checkups of the civil servants by the Hokuriku Health Management Center, Toyama, Japan.
Body height was measured in stocking feet. Body weight was recorded with the subjects clothed, and 0.5 kg was deducted from the recorded weight. The body mass index (BMI) was calculated as weight (kg)/height (m2).
The systolic and diastolic blood pressures were measured with an automated oscillometric sphygmomanometer, placed on the right arm of the seated subject who had rested in a sitting position for at least 5 minutes before the measurement. Mean arterial pressure (MAP) was used for the analysis.
Blood samples were drawn from the antecubital vein of the seated subjects. Specimens were collected in vacuum glass tubes containing sodium fluoride for blood glucose or without additives for lipids. Tubes were centrifuged for 15 minutes. After separation, the serum samples were stored at 4°C for the analysis within 72 hours.
Blood glucose was measured by an enzymatic method (Denka Seiken, Tokyo, Japan). Serum triglyceride, total cholesterol, and HDL cholesterol were measured by an enzymatic method (Daiichi, Tokyo, Japan).
HsCRP, not an item measured in the annual health checkup, was assayed by nephelometry, a latex particle–enhanced immunoassay (NA Latex, Dade Behring, Tokyo, Japan). The serum samples of 40 subjects were divided and analyzed on separate days to assess reproducibility of the measurement. This function of the assay was found to be satisfactory, as the intra-assay coefficient of variation was 3.50%. The assay was sensitive enough to detect 0.02 mg/l of serum CRP.
Psychosocial Characteristics
SOC
We used the Japanese version of SOC-13 (19) for our survey. The items were rated on a 7-point Likert scale in which higher values indicate stronger SOC. SOC consists of the three dimensions of comprehensibility, manageability, and meaningfulness, which are equally weighted and together make an overall score. SOC was assessed at phase 1 (2001) and phase 3 (2003). The difference between the two assessments for the same subjects was less than 
2 x 1.96 x mean of the within-subject SD (reproducibility (20)) for 93.3% of the pairs of observations. The average intraclass correlation coefficient was 0.77 (95% confidence interval, 0.74–0.79). These values suggest good reproducibility for SOC during the 2-year period. SOC scores in phase 3 were used for our data analysis.
Psychological Work Characteristics
A questionnaire based on Karasek’s job strain model was adopted for this study to assess job control and demand (21). The job control section consisted of 15 items, whereas the job demand section was based on five items. We used job demand and control as continuous variables by rating on a 4-point Likert scale, in which higher values indicated higher control and demand at workplace. Internal consistency-reliability was satisfactory for the scale (job control, 0.77; and job demand, 0.68). We used the job control-demand ratio to provide the psychological work characteristic.
Information about the following variables was provided by the questionnaire: educational level (less than 12 years, more than 12 years), smoking habit (current smoker, nonsmoker), regular physical exercise (less than twice a week, two times a week or more) and alcohol drinking habit (drinker, nondrinker).
Statistical Analysis
To detect the differences in biological and psychosocial characteristics between the sexes, we applied Student t test for continuous variables and the 
2 test for categorical variables. As CRP distribution skewed to lower levels, the analyses were conducted on the logarithmic transformed values. To detect differences in the mean CRP levels between the age groups, we performed an analysis of variance. Analysis of covariance was also conducted to assess the difference in CRP level between men and women adjusted for age. The Pearson product-moment correlation coefficient was used to analyze the relationship between CRP and the study variables. We conducted multiple linear regression analysis with the natural log-transformed hsCRP value as the dependent variable, in order to evaluate the association of each variable with the serum CRP level. In the first model, we entered age, BMI, MAP, total cholesterol, HDL-cholesterol, triglyceride, blood glucose, job control/demand, education, and SOC as covariates. In the second model, we entered smoking status, alcohol drinking habit, and regular physical exercise, in addition to the covariates used in the first model. The ß weights provide an estimate of the relative contributions of the independent variables in predicting CRP level. To check the fitness of the models, a qualitative approach showed that the distributions of the residuals were normal. Plots of the residuals versus each independent variable were uniformly distributed in a symmetric band centered on a residual value of zero, which was also the case for the plot of the residual versus the fitted value. As a quantitative measure, we performed Cook’s distance statistic (22) and found no potential outlier, which reiterated the validity of the assumptions in the regression models. The variance inflation factor for each variable was checked for multicollinearity between the independent variables. The data were analyzed with SPSS (version 12.0J). All reported significance levels are p < .05 (two-tailed test).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Table 1 presents the distribution of mean serum CRP and natural log-transformed values of the subjects by sex and age group. Increasing age was associated with higher CRP levels in both sexes. The age-adjusted geometric mean of CRP in women (0.23 mg/l) was significantly (p < .001) lower than that for men (0.36 mg/l).
|
The basic characteristics of the subjects are summarized in Table 2. Except for total cholesterol and practicing regular exercise, there were substantial differences in covariates between men and women. SOC was higher in men, whereas the percentage of those with more than 12 years of education was higher in women.
|
Correlations of CRP with the study variables are displayed for men and women in Table 3. The overall pattern of correlations was similar for the two groups. A strongly positive relationship between CRP and BMI was observed for both men and women. Similarly, a highly negative relationship was found between CRP and HDL-cholesterol. However, SOC was inversely related with CRP in men only (r = –0.09, p < .05). No significant correlation was detected with respect to job control/demand ratio.
|
A multiple linear regression analysis was conducted to assess the importance of the risk factors, in which the dependent variable was natural log-transformed hsCRP. For all the participants, age, BMI, MAP, HDL cholesterol, triglyceride, blood sugar, SOC, and the interaction term for SOC x sex were statistically significant (Table 4). Because of the significant interaction term, a stratified analysis was performed to check the associations for men and women separately. As shown in Table 4, an inverse association between CRP and SOC was found to be significant in men after controlling for other atherosclerosis risk factors (standardized ß coefficient –0.10, p < .01). Among the atherosclerotic risk factors, HDL cholesterol and BMI had the strongest correlations with CRP in men and women, respectively. In contrast to the univariate correlations, the predictive importance of plasma triglyceride disappeared in men, and that for MAP, total cholesterol, and blood glucose disappeared in women. Overall, the linear regression models accounted for 16% and 22% of the variance in men and women, respectively. Interaction terms of the independent variables were not significant in predicting CRP levels in men and women.
|
The association between hsCRP and lifestyle behaviors (practicing regular exercise, smoking and drinking habits) was found to be borderline nonsignificant for smoking habit in men (standardized ß coefficient 0.06, p = .07). That is, CRP was slightly higher in male current smokers after adjusting for the effects of other variables. Among women, as only 16 subjects reported as current smokers, smoking status was not entered into the multivariate analysis.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Serum CRP at concentrations known to predict adverse vascular outcomes directly inhibits endothelial progenitor cell differentiation, survival, and function, key components of angiogenesis and the response to chronic ischemia (23). As CRP is also reported to be an even stronger predictor of cardiovascular events than low-density-lipoprotein cholesterol (3), an understanding of its relevancy to biological and psychological characteristics is important from the public health viewpoint.
Associations of atherosclerotic risk factors with serum CRP level stratified for sex has been less well investigated by multivariate analysis. Due to the remarkable differences in serum CRP levels and atherosclerotic risk factors between men and women, stratification and separate analysis would be an appropriate method for handling the sex effect. Previously, Yamada et al. (24) reported on this sex difference in the results of a univariate analysis on a Japanese sample but pooled data for men and women in the multivariate model.
Our data showing higher CRP in men are controversial. This result is, however, consistent with one cross-sectional study of 2,275 males and 3,832 Japanese females aged 30 years and over (24), whereas other studies on western samples show higher CRP levels in females (25,26), even after matching for age and BMI (25). The reason for these sex discrepancies is unclear but suggests that race and environmental factors should be considered in association with CRP.
In our study, the interaction term of SOC and sex in association with CRP was significant. In the sex-stratified analysis, SOC was inversely correlated with hsCRP in men. Due to the cross-sectional nature of the study, further discussion about causality would be premature, although current and previous findings imply that the observed association is clinically explicable. Previously, we reported on a positive correlation between SOC and parasympathetic tone of cardiac autonomic function (27). Lee et al. (28) have also hypothesized that symptoms of aging such as systemic inflammation may be partly due to loss of parasympathetic function. These findings permit us to make the assumption that the health promoting effect of SOC on serum CRP levels may be partly modulated by parasympathetic nervous function. Confirmation of this hypothesis will require further investigations.
SOC did not correlate with hsCRP in women. The low levels of hsCRP among women may restrict the ability to statistically detect an association among women. The lower levels of hsCRP in women may partly be explained by estrogen suppression of interleukin-1 and interleukin-6 production (main modulators of CRP biosynthesis) (29,30) and also through lower BMI and a reduced smoking rate in our female sample. Our findings show a higher association between CRP and BMI in women than in men, which is consistent with results observed in the NHANES III study on 20,050 individuals aged over 17 years (26). Sex differences relating to the health effects of SOC have already been reported. A higher SOC was associated with shorter spells of sickness absence in men only (31). The association of SOC and cancer mortality was also observed only in men (11). Whether the gender differential observed in this study is caused by mechanistic differences between the genders as it relates to atherosclerosis cannot be determined cross-sectionally. However, the fact that gender differences were noted for serum hsCRP levels, as well as other atherosclerotic risk factors and psychosocial characteristics, lends credence to a possible intrinsic role for gender.
The incidence of cardiovascular diseases increases progressively in women after menopause (32). Recent findings have considered CRP as a valuable prognostic marker for future risk of cardiovascular events in women as well (33). In our sample, women aged over 50 years have elevated CRP levels than those aged below 50 years. To check whether menopause was related to the observed CRP associations, we replicated the analyses for 44% of our female subjects aged over 50 years. However, the results remained unchanged, suggesting that the menopause did not affect the association between atherosclerotic risk factors and psychosocial characteristics with hsCRP. We did not exclude those who had received hormone replacement therapy, but this was not likely to have confounded our results because less than 10% of Japanese women reported using hormone replacement therapy (past or current) after menopause (34).
A few prospective studies have reported on the association between negative aspects of job environment (low job control and high demand) and cardiovascular disease and mortality (35–37). Because our participants were civil servants, we entered the psychological work characteristics in the analysis as a covariate for checking the association between SOC and CRP. However, our data do not provide insight into the association between psychological work characteristics and hsCRP, which is consistent with recently published data from a population of British civil servants (38). Education is a common socioeconomic indicator in health studies (39,40). CRP has been shown to be elevated in the lower socioeconomic groups (38). However, adjusting for educational attainment did not affect the association between SOC and hsCRP among men in our sample.
Among lifestyle behaviors, smoking is related to CRP levels in men. Because there were few female smokers in this study, the effect of smoking on serum CRP could not be ascertained among women. In our study, frequency of practicing regular exercise was not associated with serum CRP, which was compatible with recent findings regarding CRP and the frequency of practicing physical exercise (41). However, previous studies have provided evidence that placement in the higher quartiles of physical activity was associated with lower concentrations of serum CRP (42). Our questionnaire did not ask for information on the exercise quantity and intensity, and this might be the reason for the lack of compatibility with the previous findings.
In 2003, a set of clinical guidelines for hsCRP was issued as a part of a global risk-prediction strategy, which suggested that levels of hsCRP of <1, 1 to <3, and 
3 mg/l be used to represent low, moderate, and high vascular risk (43). Although approximately 25% of the US population has levels of CRP greater than 3 mg/l (44), our findings suggest that CRP values tend to be low in the Japanese population. Of the total population, 80.7% of men and 90.0% of women had CRP levels less than 1 mg/l, whereas only 4.8% of men and 2.2% of women had CRP levels above 3 mg/l. These results are consistent with previous findings regarding the distribution of CRP levels in the Japanese population (24) and may reflect lower risk for coronary events among the Japanese.
Our findings have important clinical implications in terms of preventive cardiovascular strategies. The findings suggest that gender variations exist in respect to hsCRP as a strong predictor of cardiovascular diseases. Depending on the level of risk, appropriate management should be instituted, and a differential approach toward both genders would have better efficiency if based on a gender-specific risk assessment.
A limitation of this study is its lack of availability of depression to examine as a covariate. Previous reports indicated that the mental state of angioplasty patients was associated with serological markers of inflammation (45,46). However, SOC appears to be an important response in alleviating depressive symptoms (47). Furthermore, there are clues suggesting that the association between SOC and mortality from cardiovascular diseases is not explicable in terms of neuroticism (11). Thus, we speculate that depression alone does not explain the relationship between psychosocial characteristics and hsCRP in men.
It should also be mentioned that in our study population, the number of subjects less than 40 years of age was in the minority. However, the numbers were more or less representative of the age distribution in Japanese civil servants.
In conclusion, possible pathophysiologic evidence for the health-promoting effects of SOC was observed, as stronger SOC correlated with lower CRP levels in men. In addition, serum CRP correlated with traditional atherosclerotic risk factors but with a different contribution in men and women. It appears that there are gender differences in both prevalence and susceptibility with regard to atherosclerotic risk factors.
The authors are very grateful to the Municipality of Departments in Toyama Prefecture, Japan, for their support and cooperation in this study, and all the study team, in particular Mrs. Yoriko Yoshida, for data collection and clerical support.
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Received for publication April 13, 2005; revision received August 24, 2005.
DOI:10.1097/01.psy.0000195882.00407.05
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
