American Journal of Epidemiology Advance Access originally published online on January 10, 2007
American Journal of Epidemiology 2007 165(6):684-695; doi:10.1093/aje/kwk057
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
ORIGINAL CONTRIBUTIONS |
Inflammation Biomarkers and Near-Term Death in Older Men
1 Department of Pathology, College of Medicine, University of Vermont, Burlington, VT
2 Department of Biostatistics, University of Washington, Seattle, WA
3 Department of Epidemiology, School of Public Health and Community Medicine, University of Washington, Seattle, WA
4 Department of Health Services, School of Public Health and Community Medicine, University of Washington, Seattle, WA
5 Department of Medicine, School of Medicine, University of Washington, Seattle, WA
6 Department of Epidemiology, University of Pittsburgh, Pittsburgh, PA
7 Department of Medicine, University of California Davis Medical Center, Sacramento, CA
8 Department of Biochemistry, College of Medicine, University of Vermont, Burlington, VT
Correspondence to Dr. Russell P. Tracy, Department of Pathology, University of Vermont, Colchester Research Facility, 208 South Park Drive, Suite 2, Colchester, VT 05446 (e-mail: Russell.Tracy{at}uvm.edu (cc: Nancy.Jenny{at}uvm.edu)).
Received for publication November 4, 2005. Accepted for publication August 28, 2006.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Associations of C-reactive protein (CRP) and fibrinogen with death may weaken over time. Combining both markers may improve prediction of death in older adults. In 5,828 Cardiovascular Health Study participants (United States, 1989–2000), 383 deaths (183 cardiovascular disease (CVD)) in years 1–3 (early) and 914 deaths (396 CVD) in years 4–8 (late) occurred. For men, when comparing highest to lowest quartiles, hazard ratios for early death were 4.1 (95% confidence interval (CI): 2.7, 6.3) for CRP and 4.1 (95% CI: 2.7, 6.4) for fibrinogen in models adjusted for CVD risk. For early CVD death, hazard ratios were 4.3 (95% CI: 2.2, 8.4) and 3.4 (95% CI: 1.8, 6.3), respectively. When comparing men in the highest quartiles of both biomarkers with those in the lowest, hazard ratios were 9.6 (95% CI: 4.3, 21.1) for early death and 13.5 (95% CI: 3.2, 56.5) for early CVD death. Associations were weaker for late deaths. For women, CRP (hazard ratio = 2.3, 95% CI: 1.4, 3.9), but not fibrinogen (hazard ratio = 1.3, 95% CI: 0.8, 2.2), was associated with early death. Results were similar for CVD death. Neither was associated with late deaths. CRP and fibrinogen were more strongly associated with death in older men than women and more strongly associated with early than late death. Combining both markers may identify older men at greatest risk of near-term death.
aged; C-reactive protein; cardiovascular diseases; death; fibrinogen
Abbreviations: CI, confidence interval; CRP, C-reactive protein; CVD, cardiovascular disease; SD, standard deviation
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Inflammatory biomarkers such as C-reactive protein (CRP) and fibrinogen independently predict cardiovascular disease (CVD) and death in both middle-aged and older adults (1–5). In the Physicians Health Study and the Honolulu Heart Program, CRP predicted myocardial infarction up to and beyond 10 years after measurement, with little change in risk over time (6, 7). However, in other studies primarily in older people, the prognostic value of CRP appeared to be more pronounced for CVD events and death that occurred within 2–3 years of measurement than for later events (1, 8, 9). Similar results were reported for fibrinogen in men in the Cardiovascular Health Study. Fibrinogen was more strongly associated with death within 2 years of measurement than with later death (10).
In addition, sex-specific associations of inflammatory markers with cardiovascular events may exist. Fibrinogen independently predicted incident coronary heart disease in men in the Cardiovascular Health Study but was not associated with coronary heart disease or death in women (10). There are both age- and sex-dependent differences in morbidity and mortality for CVD and sex differences in cytokine production after acute injury (11) as well as intrinsic sex differences in cardiac muscle physiology and biochemistry (12). Other inflammatory conditions such as sepsis and trauma are also associated with sex-dependent outcomes (11).
Most studies of inflammatory markers have examined individual markers. However, the use of a single biomarker may not represent true inflammatory burden, especially in older adults. Many inflammatory molecules are involved in atherosclerosis, and not all biomarkers reflect the same underlying pathophysiology. A combination of biomarkers may therefore improve risk prediction. Both biovariability and reproducibility of assay method also need to be considered. Very little is known about the biovariability of many inflammatory cytokines, and many assays are not highly reproducible. Of the markers studied to date, only CRP meets both criteria (13). Fibrinogen levels show large interindividual variability, although the clinical fibrinogen assay is highly reproducible, eliminating some of the variation in the assay (13). Other markers such as interleukin-6 show less interindividual variability, but assays themselves are highly variable, with coefficients of variation in the range of 15 percent (13).
We examined the "proximate pathophysiology" of inflammatory burden (14) in older adults in the Cardiovascular Health Study by using CRP alone and combined with fibrinogen and other inflammatory markers (interleukin-6, albumin, and white blood cell count) to simultaneously assess prediction of total and CVD death 
3 years from measurement (early) and >3 years after measurement (late). We also examined potential sex-specific associations of markers with death. We hypothesized that individual markers would be more strongly associated with proximate deaths than deaths distant in time and that associations would be larger considering the joint effects of the two markers.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Cardiovascular Health Study
The Cardiovascular Health Study design has been reported previously (15). A total of 5,201 adults aged
65 years (original cohort) were recruited from 1989 to 1990. An additional 687 African Americans aged 65 years (minority cohort) were recruited between 1992 and 1993. Baseline examinations included lifestyle and medical history, physical examination, blood collection, resting 12-lead echocardiography, ankle-brachial blood pressure index, and carotid ultrasonography. Prevalence and extent of clinical CVD (confirmed angina, myocardial infarction, or stroke) were assessed (15). All subjects gave informed consent for participation in the study, and all procedures were conducted under institutionally approved protocols for use of human subjects.
Ascertainment of death
Original and minority cohorts were followed for 8 (mean, 6.9; standard deviation (SD), 1.7) and 5 (mean, 4.1; SD, 0.8) years, respectively. Deaths were ascertained by review of local obituaries and contact with proxies for those who died. Cause of death was adjudicated by using published criteria (16). CVD death was defined as death due to atherosclerotic coronary heart disease, cerebrovascular disease, atherosclerotic disease, or other CVD. There was 100 percent complete ascertainment of death status.
Definitions
Body mass index was defined as weight (kg)/height (m2). Smoking was classified as current, former (no smoking in the previous 30 days), or never. Diabetes was classified by American Diabetic Association guidelines (17). In those free of clinical CVD at baseline, subclinical CVD was defined as the presence of at least one of the following: ankle-brachial blood pressure index 0.9, maximum common or internal carotid artery intima media thickness 80th percentile, carotid artery stenosis 25 percent, major electrocardiography abnormalities, or positive response for angina or intermittent claudication on the Rose questionnaire (18). Early deaths were deaths that occurred within the first 3 years of follow-up. Deaths that occurred after 3 years of follow-up were defined as late.
Laboratory methods
Blood collection, laboratory methods, and fibrinogen measurement (coefficient of variation = 2.9 percent) have been reported (19). CRP was measured in a high-sensitivity assay developed in-house (coefficient of variation = 8.9 percent) (20). All assays were conducted by using baseline blood samples. CRP and fibrinogen were measured in 5,806 and 5,788 subjects, respectively. In analyses of CRP and fibrinogen, 5,828 subjects with at least one biomarker measurement were included. Other inflammatory markers were also measured in baseline blood samples. Albumin (19) (n = 5,807), white blood cell count (19) (n = 5,793), and interleukin-6 (21) (n = 5,382) levels were available for 5,828 participants selected for this study.
Analyses
Data were analyzed by using Stata software (22). To determine the combination of biomarkers that best predicted risk of death, each marker (CRP, fibrinogen, albumin, white blood cell count, and interleukin-6) was centered by its mean and was divided by its standard deviation to facilitate comparisons. Means were 3.64 (SD, 6.31) mg/liter for CRP, 324 (SD, 67) mg/dl for fibrinogen, 3.99 (SD, 1.89) g/liter for albumin, 6.31 x 109 (SD, 2.10) x 109cells/liter for white blood cell count, and 2.21 (SD, 1.89) pg/ml for interleukin-6. We used regression models, adjusted for age, sex, and race, to examine individual markers and combinations of markers and their associations with total and CVD death. Increased CRP, fibrinogen, white blood cell count, and interleukin-6 and decreased albumin individually were significantly associated with risk of total and CVD death in these Cardiovascular Health Study participants, as expected. When combinations of markers were examined, fibrinogen and interleukin-6 were mutually exclusive in all models tested. Decreased albumin was associated with risk of all-cause death but not significantly associated with risk of CVD death when CRP and fibrinogen were included in the models. Further examination of the biomarker combinations revealed that CRP, fibrinogen, and their interaction were significantly associated with both total and CVD death. There were no significant interactions between CRP and white blood cell count or between fibrinogen and white blood cell count for total or CVD death. CRP and fibrinogen were therefore chosen for additional analyses.
We fit Cox proportional hazards to model CRP and fibrinogen, both in quartiles and continuously, to examine associations of these biomarkers with death. CRP and fibrinogen were also modeled simultaneously to investigate joint effects of the two markers. We used extended Cox models with time-varying covariates to allow hazard ratios to differ for events prior to and after 3 years of follow-up and to formally test for differences. Hazard ratios were adjusted for age, body mass index, total cholesterol, diabetes, systolic blood pressure, smoking, race, and clinic site. Hazard ratios were also adjusted for clinical CVD at baseline and subclinical CVD among those free of clinical CVD. Results were reported as hazard ratios with 95 percent confidence intervals. Analyses were stratified by gender to investigate potential sex-specific interactions.
In secondary analyses, we examined the impact of hormone replacement therapy in women since hormone replacement therapy use is associated with significant changes in marker levels. We likewise examined the impact of aspirin and other nonsteroidal antiinflammatory drugs. Statin use was not included in the model because only 1.9 percent of the original cohort and 4.3 percent of the minority cohort were using statins at baseline (23).
To determine whether associations of biomarkers with total and/or CVD death were limited to younger or older Cardiovascular Health Study participants, analyses were also performed stratified on the median age of participants in this study (73 years). A total of 2,097 women and 1,456 men were aged 73 years at baseline; 1, 251 women and 1,024 men were aged 74 years at baseline.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Baseline risk factors and death during follow-up
Baseline characteristics by CRP and fibrinogen levels are shown in table 1. There were 1,297 total deaths (579 CVD-related deaths). A total of 4,531 participants were alive at the end of follow-up. In the 2,480 men in the Cardiovascular Health Study, 236 deaths occurred in the first 3 years (early) and 501 late deaths occurred. Of the 236 early deaths, 110 were CVD related; 13 were due to fatal myocardial infarction. Of the 501 late deaths, 223 were CVD related; 17 were due to fatal myocardial infarction. In the 3,348 women in the Cardiovascular Health Study, there were 147 early and 413 late deaths. Of the 147 early deaths, 73 were CVD related; five were due to fatal myocardial infarction. Of the 413 late deaths, 173 were CVD related; 18 were due to fatal myocardial infarction.
|
Associations of CRP and fibrinogen with early death
Among men, hazard ratios for total early death were similar for CRP and fibrinogen quartiles (table 2), and both were associated with increased risk of total early death in unadjusted and adjusted models. When modeled together to assess their joint effects, CRP, fibrinogen, and their interaction were significantly associated with total early death in both unadjusted and adjusted models. Hazard ratios were 1.39 (95 percent confidence interval (CI): 1.19, 1.62) for CRP and 1.43 (95 percent CI: 1.25, 1.64) for fibrinogen in the model adjusted for age, body mass index, total cholesterol, diabetes, systolic blood pressure, smoking, race, and clinic site. The interaction term was 0.91 (95 percent CI: 0.86, 0.96). Hazard ratios for early CVD death (table 2) were similar for both markers, and both remained significant predictors of CVD death following adjustment. Both markers and their interaction were significantly associated with early CVD death. Hazard ratios were 1.46 (95 percent CI: 1.18, 1.82) for CRP and 1.31 (95 percent CI: 1.17, 1.59) for fibrinogen in the fully adjusted model. The interaction term was 0.91 (95 percent CI: 0.85, 0.98).
|
Cardiovascular Health Study, United States, 1989–2000Among women, CRP and fibrinogen were associated with risk of total early death in unadjusted models (table 2). After adjustment, associations diminished (table 2). After our adjustments, women in the highest quartile of CRP, but not fibrinogen, were at increased risk of total early death (table 2). When modeled together, CRP, fibrinogen, and their interaction were significantly associated with total early death in unadjusted and adjusted models. Hazard ratios were 1.31 (95 percent CI: 1.10, 1.57) for CRP and 1.20 (95 percent CI: 1.00, 1.45) for fibrinogen in adjusted models. The interaction term was 0.85 (95 percent CI: 0.75, 0.96). Women in the highest quartile of CRP, but not fibrinogen, were at increased risk of early CVD death in adjusted models (table 2). CRP, fibrinogen, and their interaction were marginally associated with early CVD death after adjustment (p = 0.08). Hazard ratios were 1.31 (95 percent CI: 1.00, 1.71) for CRP and 1.23 (95 percent CI: 0.95, 1.61) for fibrinogen. The interaction term was 0.81 (95 percent CI: 0.66, 1.00).
Associations of CRP and fibrinogen with late death
Among men, hazard ratios for total late death were similar for CRP and fibrinogen quartiles, and both were associated with risk of total late death in unadjusted and adjusted models (table 3). Associations were weaker for late than for early total death. When modeled together, there was significant interaction between CRP and fibrinogen (p = 0.035) in the unadjusted model of late total death. The interaction did not remain significant upon adjustment (p = 0.078), indicating that the joint effect of the markers in the adjusted model was marginal. Hazard ratios were 1.15 (95 percent CI: 1.00, 1.31) for CRP and 1.09 (95 percent CI: 0.98, 1.20) for fibrinogen in the adjusted model. The interaction term was 0.96 (95 percent CI: 0.92, 1.00). Hazard ratios for late CVD death (table 3) were similar for both markers in quartile models, and both remained significant predictors following adjustment (table 3). When interactions between the markers were examined, hazard ratios were 1.25 (95 percent CI: 1.02, 1.53) for CRP and 1.21 (95 percent CI: 1.04, 1.40) for fibrinogen. The interaction term, 0.94 (95 percent CI: 0.88, 1.00), was marginally significant in the adjusted model.
|
Among women, CRP and fibrinogen were associated with increased risk of total late death in unadjusted models (table 3). After adjustment, only CRP was significant (table 3). In the adjusted interaction model, hazard ratios were 0.99 (95 percent CI: 0.87, 1.12) for CRP and 1.02 (95 percent CI: 0.92, 1.13) for fibrinogen. The interaction term, 1.05 (95 percent CI: 1.00, 1.11), was marginally significant. CRP and fibrinogen were associated with increased risk of late CVD death in unadjusted models (table 3). With adjustment, CRP, but not fibrinogen, remained significant. There was no significant interaction between markers in unadjusted (p = 0.853) or adjusted (p = 0.205) models. In the adjusted interaction model, hazard ratios were 1.00 (95 percent CI: 0.84, 1.19) for CRP and 1.05 (95 percent CI: 0.90, 1.24) for fibrinogen. The interaction term was 1.05 (95 percent CI: 0.97, 1.13).
Comparisons of the associations with early and late death
Using extended Cox models, we estimated hazard ratios for early and late events simultaneously for both CRP and fibrinogen. In adjusted models, increasing levels of markers were associated with increased risk of early total death and were weakly associated with late total death among men (figure 1). Hazard ratios for late total death for CRP and fibrinogen were significantly lower than those for early death (p = 0.002 and p < 0.001 for the respective comparisons). Increasing levels of CRP and fibrinogen were also associated with increased risk of early CVD death and weakly associated with increased risk of late CVD death (figure 2). Hazard ratios for late CVD death for CRP were significantly lower than those for early CVD death (p = 0.019). Hazard ratios for late CVD death for fibrinogen were marginally lower than those for early death (p = 0.052).
|
|
Among women, hazard ratios for total death for fibrinogen were not significantly different between early and late deaths (p = 0.377) (figure 3). Hazard ratios for late total death for CRP were marginally lower than those for early death (p = 0.059). Hazard ratios for CRP and fibrinogen were not significantly different when early and late CVD deaths were compared (p > 0.75) (figure 4). In models investigating all three effects, gender by early/late time by CRP and fibrinogen interactions, the omnibus tests of these interactions were highly significant (p < 0.001) for both total and CVD death.
|
|
Excluding women using hormone replacement therapy at baseline or adjusting for use of aspirin or other nonsteroidal antiinflammatory drugs in men and women did not significantly change our findings. The change in hazard ratios was
2 percent.
Combined association of CRP and fibrinogen with early death
Among men, hazard ratios for early total death increased with increasing CRP and fibrinogen, with a hazard ratio of 9.56 (95 percent CI: 4.34, 21.1) for those in the highest quartiles of both CRP and fibrinogen after adjusting for risk factors (figure 5). The hazard ratio for early CVD death associated with being in the highest quartiles of both markers compared with the lowest was 13.45 (95 percent CI: 3.20, 56.5; figure 6).
|
|
To determine whether the strong associations of CRP and fibrinogen with early death were limited to younger or older participants in the Cardiovascular Health Study, we performed analyses stratified on median baseline age (73 years). Among men aged 65–73 years at baseline, there were 96 early deaths (41 CVD related). Those in the highest quartiles of both CRP and fibrinogen compared with those in the lowest quartiles were at increased risk of early total and CVD death. Adjusted hazard ratios were 4.82 (95 percent CI: 1.60, 14.5) and 9.03 (95 percent CI: 1.14, 71.7), respectively. When we examined both markers individually and compared those in the highest with the lowest quartiles, adjusted hazard ratios for early total death were 4.43 (95 percent CI: 2.08, 9.43) for CRP and 3.08 (95 percent CI: 1.66, 5.71) for fibrinogen. Hazard ratios were 3.44 (95 percent CI: 1.11, 10.7) for CRP and 3.21 (95 percent CI: 1.23, 8.39) for fibrinogen for early CVD death.
Among men aged 74 years at baseline, there were 140 early deaths (69 CVD related). Adjusted hazard ratios for those in the highest compared with the lowest quartiles of both markers were 14.33 (95 percent CI: 4.42, 46.52) for early total death and 18.95 (95 percent CI: 2.52, 142) for early CVD death. For comparison, adjusted hazard ratios for those in the highest compared with the lowest quartiles of CRP alone and fibrinogen alone were 4.09 (95 percent CI: 2.38, 7.01) and 4.33 (95 percent CI: 2.35, 7.98), respectively, for early total death and 4.73 (95 percent CI: 2.04, 11.0) and 4.31 (95 percent CI: 1.76, 10.6), respectively, for early CVD death.
Analyses combining CRP and fibrinogen did not reveal consistent trends for either total death (figure 7) or CVD death (figure 8) in women. In adjusted analyses, the hazard ratios for early total and CVD death for women in the highest quartiles of both markers compared with the lowest were 1.50 (95 percent CI: 0.73, 3.10) and 1.44 (95 percent CI: 0.80, 2.59), respectively.
|
|
Similar trends were revealed in age-stratified analyses. Among women aged
73 years, there were 51 early deaths (24 CVD related). When we compared those in the highest quartiles of both CRP and fibrinogen with those in the lowest quartiles, adjusted hazard ratios were 1.31 (95 percent CI: 0.40, 4.34) for early total death and 1.94 (95 percent CI: 0.22, 17.2) for early CVD death. Adjusted hazard ratios for those in the highest versus the lowest quartiles of CRP alone and fibrinogen alone were 1.72 (95 percent CI: 0.75, 3.90) and 1.28 (95 percent CI: 0.60, 1.74) for early total death and 2.64 (95 percent CI: 0.55, 12.6) and 1.53 (95 percent CI: 0.50, 4.72) for early CVD death.
Among women aged 74 years at baseline, there were 96 early deaths (48 CVD related). Adjusted hazard ratios for those in the highest quartiles of both CRP and fibrinogen compared with those in the lowest were 1.25 (95 percent CI: 0.52, 3.02) for early total death and 0.78 (95 percent CI: 0.21, 2.87) for early CVD death. For those in the highest versus the lowest quartiles of CRP alone and fibrinogen alone, hazard ratios were 2.51 (95 percent CI: 1.33, 4.73) and 0.94 (95 percent CI: 0.51, 1.72) for early total death and 1.56 (95 percent CI: 0.67, 3.67) and 1.24 (95 percent CI: 0.49, 3.13) for early CVD death.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The major findings of this population-based study of older men and women were that 1) prediction of death, both total and CVD related, using CRP or fibrinogen was significantly stronger for early death occurring within 3 years of marker measurement than for later death; 2) among men, this association was largely independent of CVD risk factors while in women, adjustment tended to weaken the association; and 3) CRP and fibrinogen were independent predictors of death, and the joint effect of both markers provided enhanced ability to identify men at high risk of near-term death.
Our study confirmed the concept of an association of inflammation biomarkers with risk of death in older adults (1, 4, 9, 10, 24). However, the time dependency of death prediction in this study is striking and considerably greater than the relatively weak time-dependent prediction of CVD events reported previously (1, 8, 9, 25, 26). Other studies did not identify any time dependency (6, 24, 27). Differences in time dependency between studies may be due to the outcomes being assessed. Our findings are specific for death since, in the Cardiovascular Health Study, we did not observe a significant time dependency when the association of CRP with myocardial infarction was examined (28). Disparities among studies also may be due to differences in the age of participants, health status and behaviors, duration of follow-up, and/or extent or type of underlying atherosclerosis. For example, the Honolulu Heart Program studied elderly Japanese-American men only (4), and the Iowa 65+ Rural Health Promotion Project comprised very high-functioning, older men and women (1).
It is possible that our findings represent a survivor effect and not necessarily a change in inflammation burden over time. Through genes, environment, and/or their interactions, some persons may suffer more detrimental effects from a high inflammatory burden than others do. The diminished risk of death after 3 years may reflect a depletion of those with the highest biomarker levels who are most susceptible to inflammation-related morbidity and mortality.
Alternatively, there may be changes in inflammatory processes over time, with increased inflammation leading to increased risk of death. In support of the latter possibility, age likely modifies the associations of inflammatory markers with CVD and death. CRP and fibrinogen predict CVD events in both older and middle-aged men and women (1, 10, 28, 29); however, the relative risks for older people appear lower than for younger people and lower for older women than for older men (28). One reason for this finding may be that there are fundamental differences in underlying pathology between younger and older people. The frequency of finding a thrombus at autopsy decreases from greater than two thirds in younger men dying from sudden cardiac death to less than one third in older men (30).
In the face of autopsy and epidemiologic data, we hypothesize that, in younger individuals, in the absence of other major disease processes, inflammatory biomarkers reflect the progression toward vascular lesions and therefore yield both short- and long-term prediction, much as cholesterol does. However, in older people, inflammation biomarkers may reflect the degree of complex advanced vascular disease burden, as well as other morbidities, with the highest levels reflecting severe pathology associated with early risk. Even within the older cohort of the Cardiovascular Health Study, we saw an age-related difference in the joint effect of CRP and fibrinogen in the prediction of death. For men aged 65–73 years at baseline, combining CRP and fibrinogen enhanced prediction of near-term CVD death but not near-term all-cause death. For men who were 74 years or older at baseline, combining CRP and fibrinogen enhanced prediction of both near-term CVD and all-cause death.
The much weaker association of CRP and fibrinogen with death among women in the Cardiovascular Health Study may reflect different pathophysiology in men and women. There are gender-related differences in cardiac muscle physiology and biochemistry (12). Major risk factors are generally similar between men and women early in life, but they differ later in life. For example, women with obstructive coronary artery disease have more severe disease and disability than men do (31). There are also gender differences in presentation and detection of CVD (31) and differences in CVD morbidity and mortality (11). In addition, studies have reported gender differences in cytokine production after trauma (11) and differences in morbidity and mortality in inflammatory conditions other than CVD (12). Women may differ from men in their susceptibility to the effects of increased inflammation as well. Hormone replacement therapy use did not appear to account for the gender-specific prediction in the current study. Although hormone replacement therapy affects levels of fibrinogen and CRP (32), only 2.3 percent of women in this study were using hormone replacement therapy at baseline, and excluding these women from analyses did not significantly change the results.
Our study, like others, is potentially limited by the use of a single biomarker measurement that contains both analytical and within-subject (i.e., day-to-day) variance. However, this factor would likely bias results toward the null. Therefore, we attempted to more accurately place study participants in the proper strata by using a combination of biomarkers to represent inflammatory burden. The MacArthur Research Network on Successful Aging Community Study (14) used four biomarkers—total cholesterol, albumin, CRP, and interleukin-6—to reflect inflammatory burden. Study participants with abnormal levels of three or four of the markers were at substantially increased risk of death, while those with abnormal levels of only one or two of the markers were at a slightly increased risk of death compared with those whose levels of markers were in the healthy reference range (14). We examined five biomarkers for this study: CRP, fibrinogen, albumin, interleukin-6, and white cell count. Of these, CRP and fibrinogen provided independent and complimentary information and were the most significantly associated with risk of death. In addition, it is likely that CRP and fibrinogen represent different aspects of inflammation. CRP mediates foam cell uptake of plasma lipids (33), activates the complement cascade (34), and may impact monocyte and endothelial cell functions (35, 36). Elevated levels of fibrinogen are associated with plasma viscosity and may contribute to increased platelet cross-linking and increased fibrin formation (37). Both CRP and fibrinogen may be in the causal pathway of atherosclerosis, which, in turn, contributes to death through a variety of mechanisms.
Currently, a limited number of biomarkers are readily measured clinically (13). In the future, we anticipate using multianalyte panels to quickly and efficiently measure aspects of many different pathways involved in atherosclerosis and other disease processes. Measuring multiple markers by using well-characterized assays will likely improve our ability to predict events in women as well. For example, soluble intracellular adhesion molecule-1, a marker of endothelial perturbation, was more predictive of events in women than in men in the Cardiovascular Health Study cohort (38). Future studies may need to incorporate gender-specific marker panels to best assess risk of events.
Biomarkers that identify those at the greatest risk of near-term morbidity and mortality will be useful clinically in targeting aggressive interventions. In particular, interventions to reduce inflammatory burden will likely impact a variety of pathophysiologic processes. Much focus has been on pharmacologic interventions such as aspirin and statins. However, lifestyle interventions such as increased physical activity may also reduce inflammation (39–41). Large clinical trials, especially among women, are needed to determine appropriate interventions, measure effects of reducing inflammatory burden on clinical outcomes, and determine the necessary intensity and duration of treatment to maximize the benefit of specific interventions.
In summary, we report that the inflammatory biomarkers CRP and fibrinogen were associated with death, particularly death within 3 years of biomarker measurement, in older men and women. The associations were stronger among men than women, and prediction of death in men was improved when both markers were combined. Our data support further exploration of the use of multiple biomarkers in assessing inflammation status and risk of morbidity and mortality.
|
ACKNOWLEDGMENTS |
|---|
The research reported on in this article was supported by contracts N01-HC-85079 through N01-HC-85086, N01-HC-35129, and N01-HC-15103 from the National Heart, Lung, and Blood Institute and R01-HL-46696 (R. P. T., N. S. J.) from the National Institutes of Health.
Participating institutions and principal investigators, Cardiovascular Health Study—Steering Committee Chairman: Dr. Curt D. Furberg, Wake Forest University School of Medicine. National Heart, Lung, and Blood Institute Project Office: Dr. Jean Olson; Wake Forest University School of Medicine: Dr. Gregory L. Burke; Wake Forest University—Echocardiography Reading Center: Dr. Ronald Prineas; University of California, Davis: Dr. John Robbins; The Johns Hopkins University: Dr. Linda P. Fried; The Johns Hopkins University—Magnetic Resonance Imaging Reading Center: Dr. David Yousem; University of Pittsburgh: Dr. Lewis H. Kuller; University of California, Irvine—Echocardiography Reading Center (baseline): Dr. Julius M. Gardin; Georgetown Medical Center—Echocardiography Reading Center (follow-up): Dr. John S. Gottdiener; New England Medical Center, Boston—Ultrasound Reading Center: Dr. Daniel H. O'Leary; University of Vermont—Central Blood Analysis Laboratory: Dr. Russell P. Tracy; University of Arizona, Tucson—Pulmonary Reading Center: Dr. Paul Enright; Retinal Reading Center—University of Wisconsin: Dr. Ronald Klein; and University of Washington—Coordinating Center: Dr. Richard A. Kronmal.
Conflict of interest: none declared.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
- Tracy R, Lemaitre R, Psaty B, et al. (1997) Relationship of C-reactive protein to risk of cardiovascular disease in the elderly: results from the Cardiovascular Health Study and the Rural Health Promotion Project. Arterioscler Thromb Vasc Biol 17:1121–7.
[Abstract/Free Full Text] - Meade TW. (1995) Fibrinogen in ischaemic heart disease. Eur Heart J 16:suppl A, 31–4 discussion 34–5.
- Ernst E and Resch KL. (1993) Fibrinogen as a cardiovascular risk factor: a meta-analysis and review of the literature. Ann Intern Med 118:956–63.
[Abstract/Free Full Text] - Yano K, Grove J, Chen R, et al. (2001) Plasma fibrinogen as a predictor of total and cause-specific mortality in elderly Japanese-American men. Arterioscler Thromb Vasc Biol 21:1065–70.
[Abstract/Free Full Text] - Woodward M, Rumley A, Tunstall-Pedoe H, et al. (2003) Does sticky blood predict a sticky end? Associations of blood viscosity, haematocrit and fibrinogen with mortality in the West of Scotland. Br J Haematol 122:645–50.[CrossRef][Web of Science][Medline]
- Ridker PM, Cushman M, Stampfer MJ, et al. (1997) Inflammation, aspirin, and the risk of cardiovascular disease in apparently healthy men. N Engl J Med 336:973–9.
[Abstract/Free Full Text] - Sakkinen P, Abbott RD, Curb JD, et al. (2002) C-reactive protein and myocardial infarction. J Clin Epidemiol 55:445–51.[CrossRef][Web of Science][Medline]
- Pirro M, Bergeron J, Dagenais GR, et al. (2001) Age and duration of follow-up as modulators of the risk for ischemic heart disease associated with high plasma C-reactive protein levels in men. Arch Intern Med 161:2474–80.
[Abstract/Free Full Text] - Strandberg TE and Tilvis RS. (2000) C-reactive protein, cardiovascular risk factors, and mortality in a prospective study in the elderly. Arterioscler Thromb Vasc Biol 20:1057–60.
[Abstract/Free Full Text] - Tracy RP, Arnold AM, Ettinger W, et al. (1999) The relationship of fibrinogen and factors VII and VIII to incident cardiovascular disease and death in the elderly: results from the cardiovascular health study. Arterioscler Thromb Vasc Biol 19:1776–83.
[Abstract/Free Full Text] - Kher A, Wang M, Tsai BM, et al. (2005) Sex differences in the myocardial inflammatory response to acute injury. Shock 23:1–10.[Web of Science][Medline]
- Carroll JD, Carroll EP, Feldman T, et al. (1992) Sex-associated differences in left ventricular function in aortic stenosis of the elderly. Circulation 86:1099–107.
- Pearson TA, Mensah GA, Alexander RW, et al. (2003) Markers of inflammation and cardiovascular disease: application to clinical and public health practice: a statement for healthcare professionals from the Centers for Disease Control and Prevention and the American Heart Association. Circulation 107:499–511.
- Reuben DB, Cheh AI, Harris TB, et al. (2002) Peripheral blood markers of inflammation predict mortality and functional decline in high-functioning community-dwelling older persons. J Am Geriatr Soc 50:638–44.[CrossRef][Web of Science][Medline]
- Fried LP, Borhani NO, Enright P, et al. (1991) The Cardiovascular Health Study: design and rationale. Ann Epidemiol 1:263–76.[Medline]
- Ives DG, Fitzpatrick AL, Bild DE, et al. (1995) Surveillance and ascertainment of cardiovascular events. The Cardiovascular Health Study. Ann Epidemiol 5:278–85.[CrossRef][Medline]
- Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care (1997) 20:1183–97.[Web of Science][Medline]
- Kuller LH, Shemanski L, Psaty BM, et al. (1995) Subclinical disease as an independent risk factor for cardiovascular disease. Circulation 92:720–6.
- Cushman M, Cornell ES, Howard PR, et al. (1995) Laboratory methods and quality assurance in the Cardiovascular Health Study. Clin Chem 41:264–70.
[Abstract/Free Full Text] - Macy EM, Hayes TE, Tracy RP. (1997) Variability in the measurement of C-reactive protein in healthy subjects: implications for reference intervals and epidemiological applications. Clin Chem 43:52–8.
[Abstract/Free Full Text] - Jenny NS, Tracy RP, Ogg MS, et al. (2002) In the elderly, interleukin-6 plasma levels and the -174G>C polymorphism are associated with the development of cardiovascular disease. Arterioscler Thromb Vasc Biol 22:2066–71.
[Abstract/Free Full Text] - StataCorp. (2001) Stata Statistical Software. (Stata Corporation, College Station, TX).
- Lemaitre RN, Furberg CD, Newman AB, et al. (1998) Time trends in the use of cholesterol-lowering agents in older adults: the Cardiovascular Health Study. Arch Intern Med 158:1761–8.
[Abstract/Free Full Text] - Harris TB, Ferrucci L, Tracy RP, et al. (1999) Associations of elevated interleukin-6 and C-reactive protein levels with mortality in the elderly. Am J Med 106:506–12.[CrossRef][Web of Science][Medline]
- Danesh J, Wheeler JG, Hirschfield GM, et al. (2004) C-reactive protein and other circulating markers of inflammation in the prediction of coronary heart disease. N Engl J Med 350:1387–97.
[Abstract/Free Full Text] - Danesh J, Whincup P, Walker M, et al. (2000) Low grade inflammation and coronary heart disease: prospective study and updated meta-analyses. BMJ 321:199–204.
[Abstract/Free Full Text] - Kuller LH, Tracy RP, Shaten J, et al. (1996) Relation of C-reactive protein and coronary heart disease in the MRFIT nested case-control study. Multiple Risk Factor Intervention Trial. Am J Epidemiol 144:537–47.
[Abstract/Free Full Text] - Cushman M, Arnold AM, Psaty BM, et al. (2005) C-reactive protein and the 10-year incidence of coronary heart disease in older men and women: the Cardiovascular Health Study. Circulation 112:25–31.
- Danesh J, Collins R, Appleby P, et al. (1998) Association of fibrinogen, C-reactive protein, albumin, or leukocyte count with coronary heart disease: meta-analyses of prospective studies. JAMA 279:1477–82.
[Abstract/Free Full Text] - Burke AP, Farb A, Pestaner J, et al. (2002) Traditional risk factors and the incidence of sudden coronary death with and without coronary thrombosis in blacks. Circulation 105:419–24.
- Bairey Merz N, Bonow RO, Sopko G, et al. (2004) Women's Ischemic Syndrome Evaluation: current status and future research directions: report of the National Heart, Lung and Blood Institute workshop: October 2–4, 2002: executive summary. Circulation 109:805–7.
- Cushman M, Meilahn EN, Psaty BM, et al. (1999) Hormone replacement therapy, inflammation, and hemostasis in elderly women. Arterioscler Thromb Vasc Biol 19:893–9.
[Abstract/Free Full Text] - Zwaka TP, Hombach V, Torzewski J. (2001) C-reactive protein-mediated low density lipoprotein uptake by macrophages: implications for atherosclerosis. Circulation 103:1194–7.
- Lagrand WK, Visser CA, Hermens WT, et al. (1999) C-reactive protein as a cardiovascular risk factor: more than an epiphenomenon? Circulation 100:96–102.
- Cermak J, Key NS, Bach RR, et al. (1993) C-reactive protein induces human peripheral blood monocytes to synthesize tissue factor. Blood 82:513–20.
[Abstract/Free Full Text] - Pasceri V, Willerson JT, Yeh ET. (2000) Direct proinflammatory effect of C-reactive protein on human endothelial cells. Circulation 102:2165–8.
- Tracy R and Bovill E. (1992) Thrombosis and cardiovascular risk in the elderly. Arch Pathol Lab Med 116:1307–12.[Web of Science][Medline]
- Jenny NS, Arnold AM, Kuller LH, et al. (2006) Soluble intracellular adhesion molecule-1 is associated with cardiovascular disease risk and mortality in older adults. J Thromb Haemost 4:107–13.[Web of Science][Medline]
- Colbert LH, Visser M, Simonsick EM, et al. (2004) Physical activity, exercise, and inflammatory markers in older adults: findings from the Health, Aging and Body Composition Study. J Am Geriatr Soc 52:1098–104.[CrossRef][Web of Science][Medline]
- Geffken DF, Cushman M, Burke GL, et al. (2001) Association between physical activity and markers of inflammation in a healthy elderly population. Am J Epidemiol 153:242–50.
[Abstract/Free Full Text] - Reuben DB, Judd-Hamilton L, Harris TB, et al. (2003) The associations between physical activity and inflammatory markers in high-functioning older persons: MacArthur Studies of Successful Aging. J Am Geriatr Soc 51:1125–30.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. W McDade, J. N Rutherford, L. Adair, and C. Kuzawa Population differences in associations between C-reactive protein concentration and adiposity: comparison of young adults in the Philippines and the United States Am. J. Clinical Nutrition, April 1, 2009; 89(4): 1237 - 1245. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. S. Jenny, A. M. Arnold, L. H. Kuller, R. P. Tracy, and B. M. Psaty Associations of Pentraxin 3 With Cardiovascular Disease and All-Cause Death: The Cardiovascular Health Study Arterioscler Thromb Vasc Biol, April 1, 2009; 29(4): 594 - 599. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Goldman, D. A. Glei, Y.-H. Lin, and M. Weinstein Improving Mortality Prediction Using Biosocial Surveys Am. J. Epidemiol., March 15, 2009; 169(6): 769 - 779. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. W. McDade, J. N. Rutherford, L. Adair, and C. Kuzawa Adiposity and Pathogen Exposure Predict C-Reactive Protein in Filipino Women J. Nutr., December 1, 2008; 138(12): 2442 - 2447. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. M Ridker High-Sensitivity C-Reactive Protein as a Predictor of All-Cause Mortality: Implications for Research and Patient Care Clin. Chem., February 1, 2008; 54(2): 234 - 237. [Full Text] [PDF]
|
|
|
|
|
|
W. Koenig, N. Khuseyinova, J. Baumert, and C. Meisinger Prospective Study of High-Sensitivity C-Reactive Protein as a Determinant of Mortality: Results from the MONICA/KORA Augsburg Cohort Study, 1984-1998 Clin. Chem., February 1, 2008; 54(2): 335 - 342. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. A. Lange, E. M. Lange, R. P. Tracy, and A. P. Reine Polymorphisms in the CRP Gene and Cardiovascular Events--Reply JAMA, March 28, 2007; 297(12): 1317 - 1318. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||