|Home | About | Journals | Submit | Contact Us | Français|
Pentraxin-3 is an inflammatory marker thought to be more specific to vascular inflammation than C-reactive protein (CRP). Whether pentraxin-3 is independently associated with adverse events among persons with stable coronary heart disease (CHD), independently of CRP, and whether kidney dysfunction influences these associations, is not known.
We evaluated the associations of baseline pentraxin-3 levels with all-cause mortality, cardiovascular events (myocardial infarction, stroke or CHD death), and incident heart failure during 37 months among ambulatory persons with stable CHD participating in the Heart and Soul Study. Cox proportional hazards models were adjusted for age, sex, race, hypertension, diabetes, smoking, and CRP.
Among 986 persons with stable CHD, each one unit increase in log pentraxin-3 at baseline was associated with an 80% increased risk of all-cause mortality (HR 1.8, 95% CI 1.5–2.1), a 50% increased risk of cardiovascular events (HR 1.5, 95% CI, 1.2–1.9), and an 80% greater risk of incident heart failure (HR 1.8, 95% CI, 1.3–2.5). Further adjustment for estimated glomerular filtration rate (eGFR) attenuated these associations to 1.6 (1.3–1.9) for mortality, 1.3 (1.0–1.6) for cardiovascular events and 1.5 (1.1–2.1) for incident heart failure. Stratification by eGFR above or below 60 ml/min/1.73m2 did not affect these associations (p interaction >0.3 for all outcomes).
Among persons with stable CHD, higher pentraxin-3 concentrations were associated with increased risk for all-cause mortality, cardiovascular events and incident heart failure independently of systemic inflammation. Adjustment for eGFR modestly attenuated these associations, suggesting that future studies of pentraxin-3 should adjust for kidney function.
Pentraxin-3 (PTX3) has emerged as a novel marker thought to be more specific to vascular inflammation than other proteins in the pentraxin family such as C-reactive protein (CRP). Higher PTX3 levels are associated with worse cardiovascular outcomes after acute coronary syndromes, independently of CRP.1–2 PTX3 is also associated with increased risk of cardiovascular death among elderly persons without established cardiovascular disease (CVD).3 The biological plausibility of its role in CVD risk is supported by its localization in atherosclerotic plaques4 and the higher concentration of PTX3 but not CRP in the coronary sinus of patients with heart failure (HF).5 Whether PTX3 is a predictor of adverse outcomes among persons with stable coronary disease has not been well studied.
Kidney dysfunction may be an important factor in the association between PTX3 and adverse outcomes. Higher PTX3 concentrations are associated with lower estimated glomerular filtration rates (eGFR) across all stages of established chronic kidney disease (CKD).6–7 CKD is associated with endothelial dysfunction8–9 and, in subjects with CKD, PTX3 is independently associated with endothelial dysfunction.9 CKD is an independent risk factor for death, cardiovascular events and incident heart failure, an association that is present at earlier stages of kidney dysfunction when eGFR is measured by cystatin C.10–14 Thus, it is plausible that CKD may confound or modify the associations of PTX3 and adverse events. Understanding the influence of kidney dysfunction in these associations could elucidate pathways common to CKD and CVD, and could provide guidance for future studies about whether responsible mechanisms may be driven by PTX3 or inflammation, or rather through other mechanisms linked to kidney dysfunction.
We evaluated the associations of PTX3 with all-cause mortality, CVD events, and incident HF in ambulatory persons with stable coronary heart disease (CHD) enrolled in the Heart and Soul Study. We also evaluated whether the potential associations between PTX3 and the outcomes of interest are influenced by kidney function (eGFR measured by cystatin C).
As described previously, the Heart and Soul Study is a prospective cohort study initially designed to evaluate the influence of psychosocial factors on CVD events in ambulatory persons with stable CHD.15 Study participants were recruited from outpatient clinics in the San Francisco Bay Area using one or more of the following inclusion criteria: (1) history of myocardial infarction; (2) angiographic evidence of 50% stenosis in 1 or more coronary vessels; (3) evidence of exercise-induced ischemia by treadmill or nuclear testing; or (4) history of coronary revascularization. In addition, subjects who met the following criteria were excluded: myocardial infarction within the last 6 months, exercise tolerance less than one block, or likely to move out of the area within 3 years. The appropriate institutional review boards approved the protocol and all participants gave written, informed consent.
Overall, 1,024 study participants were recruited between September 2000 and December 2002. For every participant, baseline evaluation was conducted at a one-day appointment by trained research assistants, and included questionnaires related to social and medical history and physical examination. Phlebotomy was performed after a 12-hour fast, and serum was frozen at 70°C. For all-cause mortality and composite CV event outcomes, we included 986 patients with PTX3 measures; for HF, we excluded 172 patients with baseline HF by self-report. We created a missing category for covariates with missing observations.
PTX3 was measured at University of Vermont Laboratory for Clinical Biochemistry Research, Department of Pathology, from frozen samples collected at the baseline study visit by Human Pentraxin 3/TSG-14 Immunoassay (R&D Systems; Minneapolis, MN). Coefficients of variation (CV’s) ranged from 4.1–8.1% (average analytical CV 6.2%). The assay range is 0.31–20 ng/ml. Measurements were made in duplicate and averaged.
GFR was estimated from serum cystatin C. Cystatin C is an alternative filtration marker with stronger and more linear associations with cardiovascular outcomes than creatinine.11, 13 Cystatin C-based estimates may be more accurate among persons with higher levels of kidney function.16 Cystatin C was measured using a particle-enhanced immuno-nephelometric assay (N Latex Cystatin C; Siemens, formerly Dade Behring). Intra-assay CV’s are less than 3.1%. eGFR was calculated using the formula 76.7 x cys C−1.19, which was developed from pooled data from cohorts with eGFR measured by iothalamate.17
At the baseline visit, all participants completed questionnaires related to social and medical history, and underwent physical examinations. Blood pressure was measured by trained research assistants using calibrated blood pressure cuffs. Additional serological measurements included total cholesterol, high-density lipoprotein cholesterol, and triglyceride concentrations. The Friedewald equation was used to measure low-density lipoprotein cholesterol.18 High sensitivity C-reactive protein was measured as previously described.19
We considered three outcomes: all-cause mortality, CVD events (nonfatal myocardial infarction, stroke or death due to CHD), and incident HF. Outcomes were initially ascertained by annual phone interviews with study participants or family, and were then confirmed by two blinded adjudicators after review of medical records. CV death was defined as: (1) death during the same hospitalization in which an acute myocardial infarction was documented or (2) death not explained by other causes and that occurred within one hour of the onset of terminal symptoms. Nonfatal myocardial infarction was defined by the American Heart Association diagnostic criteria.20 Stroke was defined as a new neurological deficit not known to be secondary to brain trauma, tumor, infection, or other cause.21 Persons were deemed to have incident HF if they did not have a known history of HF but were hospitalized with a clinical diagnosis of HF involving at least two of the following new or changed symptoms: paroxysmal nocturnal dyspnea, orthopnea, elevated jugular venous pressure, pulmonary rales, third heart sound, and cardiomegaly or pulmonary edema on chest radiography.22 Quantitative evidence of new-onset heart failure (echocardiograms or invasively measured hemodynamics) was obtained when possible. For the analysis of incident HF, 172 persons with prevalent HF were excluded. Death certificates and autopsy reports were used to determine all-cause mortality. For mortality, the median follow-up time was 49 months; for HF, 50 months; and for composite CV event, 49 months. Among 986 patients with PTX3 measures, 4 (<1%) were lost to follow-up.
We compared sociodemographic and anthropometric data, and comorbidities by tertile of PTX3 using ANOVA or Chi Square. We evaluated the association between PTX3 and each outcome separately using Cox proportional hazards models. PTX3 was characterized as a linear variable, log transformed due to its skewed distribution and also categorized into tertiles. We used staged models to evaluate the importance of potential confounders or mediators in observed associations. We first adjusted for age, gender, and race/ethnicity. We further adjusted for well-established risk factors associated with the outcomes of interest including diabetes, hypertension, and smoking (past or present). Because the pathways by which PTX3 may be associated with these outcomes are unknown, we further adjusted for covariates that were significantly associated with PTX3 in bivariate analyses and also associated with the outcome at a p-value of <0.1. The variables evaluated included education, systolic blood pressure, fasting glucose, body mass index, LDL, HDL, triglycerides, and CRP. In order to understand the potential importance of eGFR in the associations of PTX3 and outcomes, we added eGFR to the fully adjusted model. Moreover, by adding a cross-term to models, we tested for interactions by the presence of CKD, defined as eGFR <60 ml/min/1.73m2 in order to understand whether CKD may modify the association of PTX3 with each outcome. In a secondary analysis, we also adjusted for LV mass and ejection fraction in order to test whether these may mediate observed associations between PTX3 and cardiovascular outcomes.
The Heart and Soul Study was supported by the Department of Veterans Affairs, the National Heart Lung and Blood Institute (R01 HL079235); the American Federation for Aging Research (Paul Beeson Scholars Program); the Robert Wood Johnson Foundation (Generalist Physician Faculty Scholars Program); and the Ischemia Research and Education Foundation, which were administered by the Northern California Institute for Research and Education, and with resources of the Veterans Affairs Medical Center, San Francisco, California. CP is currently supported by 1K23DK082793-01 (NIDDK) and the Robert Wood Johnson Harold Amos award. These funding sources had no involvement in the design or execution of this study. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper and its final contents.
Among 986 participants, mean (SD) age was 67 years (11), median (25th, 75th percentile) of PTX3 was 0.60 ng/ml (0.41, 0.96), and mean (SD) of eGFR was 70.8 (23.0). Persons with higher PTX3 were older, more likely to be of white race, more likely to have a history of CHF, and had lower eGFR, higher LV mass index and lower ejection fraction by echocardiography. Persons in the highest tertile of PTX3 were less likely to smoke, had lower triglycerides, and had lower levels of CRP and urine albumin to creatinine ratio (Table 1).
Higher PTX3 levels were associated with overall mortality independently of demographics, comorbidities, and CRP concentrations. For every unit increase in log PTX3, there was an 80% increased hazard for death after multivariate adjustment. Further adjustment for eGFR resulted in attenuation of the beta coefficient by 12%, and it remained statistically significant (Table 2). When we categorized PTX3 into tertiles, the age-adjusted death rate was more than 3-fold higher for persons in the highest tertile of PTX3 compared to the lowest (Figure 1). Compared to the lowest tertile, participants in the second and third tertile had a step-wise increase in the hazard for all-cause death after multivariate adjustment. The addition of eGFR to the models mildly attenuated the associations in the highest tertile and in the linear analysis (Table 2).
Higher PTX3 concentrations were also associated with a 50% increased hazard of CV events and an 80% increased hazard of incident HF in linear models after multivariate adjustment. The addition of eGFR to the model resulted in attenuation of the beta coefficient by 33% for CV events and by 24% for HF, but they remained statistically significant (Table 2). Comparing study participants in tertile three to tertile one by age-adjusted event rate per 1,000 person years, increased levels of PTX3 were associated with increased CV events (p= 0.018), incident HF (0.002) and overall mortality (<0.001) (Figure 1). Participants in the highest tertile of PTX3 had a 50% increased risk of CVD event (p=0.05) and 90% increased risk of incident HF (p=0.03) compared to the lowest tertile. Additional adjustment for eGFR attenuated these associations to 40% for CV events and 50% for incident HF (Table 2).
In an exploratory analysis, we found that further adjustment for LV mass index and ejection fraction had no meaningful effect on associations of PTX3 with cardiovascular outcomes. Hazard ratio (HR) for CVD events was 1.3 (p=0.04) adjusted for LV mass index, and 1.3 (p=0.05) when adjusted for ejection fraction; HR for incident HF was 1.5 (p=0.04) adjusted for LV mass index, and 1.4 (p=0.08) adjusted for ejection fraction. We also tested whether eGFR is an effect modifier in multivariate models using the product of eGFR and continuous log PTX3. Results were similar among persons with or without CKD; P-value for interaction was 0.4 for CV events, 0.3 for incident HF, and 0.9 for overall mortality. Point estimates for overall mortality in the multivariate model (not adjusted for eGFR), using log PTX3 as predictor, were identical for subjects with eGFR > and <60 ml/min/1.73m2: 1.7 (1.3–2.2) (p<0.001).
In this cohort of persons with stable CHD, we found that PTX3 is significantly associated with all-cause mortality, CV events, and incident HF independently of demographics, traditional CVD risk factors and systemic inflammation (CRP). Adjustment for eGFR modestly attenuated these associations. Our findings suggest that this novel marker of vascular inflammation may be an important mechanism involved in vascular injury and repair among persons with stable coronary heart disease.
To our knowledge, we are the first to report an independent association of PTX3 with overall mortality among persons with established and stable CVD, independent of systemic inflammation. PTX3 is produced by multiple cell types (peripheral leukocytes, vascular endothelial cells, smooth muscle cells, to name a few) and is produced in response to both inflammatory stimuli (IL-1, TNF-a, agonists of TLR, LPS) and anti-inflammatory stimuli (IL-10 and HDL).23 The finding that PTX3 is strongly associated with overall mortality in our cohort invites at least two interpretations. Vascular pathology may contribute to the overall risk of death in a cohort of subjects with CVD. Conversely, it is possible that this strong association with overall mortality relates to a non-vascular function of PTX3. The mechanism by which PTX3 is associated with overall mortality requires further study.
Our findings that PTX3 is associated with cardiovascular events and incident heart failure, independent of systemic inflammation are noteworthy. These findings are in accordance with prior reports suggesting that PTX3 may be more strongly associated with adverse outcomes after acute coronary ischemia than CRP and troponin1–2 and that it may be associated with cardiovascular death in an elderly cohort free of CVD.3 The association with incident heart failure is supported by previous findings that PTX3 is elevated in heart failure patients with normal ejection fraction.5 The mechanisms by which PTX3 is associated with cardiovascular outcomes remain unclear. Current investigations suggest that PTX3 may be part of a protective mechanism in vascular repair. PTX3 binds and inactivates FGF-2, an angiogenic growth factor responsible for smooth muscle proliferation involved in atherosclerosis.24–25 Knockout models in mice have been used to demonstrate that atherosclerotic lesions develop faster in PTX3 deficient mice.26 Cardiac ischemia-reperfusion injury was exacerbated in PTX3 knockout mice.27 Taken together, these findings suggest that PTX3 may be elevated in vascular injury as a protective mechanism, much like white blood cells are elevated during infection. Very high levels of PTX3 may indicate a more severe vascular disease state, explaining its ability to detect increased risk for adverse outcomes. Elucidating PTX3 pathways in vascular pathology may lead to a better understanding of risk for secondary events among persons with stable CHD.
Interestingly, the addition of eGFR as an adjustment variable resulted in some attenuation of associations of PTX3 with CV events and incident HF, albeit PTX3 remained statistically significantly associated with these outcomes even after adjustment for eGFR. Several possibilities may explain these findings. One is that vascular injury is a parallel process present in persons with kidney dysfunction and coronary heart disease. Another is that kidney dysfunction is a confounder due to filtration of PTX3 by the kidney. It is less likely that PTX3 is filtered in the glomerulus, given its molecular weight (42kDa), and our group’s findings that the association of PTX3 and eGFR differs by race/ethnicity.6 In the exploratory analysis, the addition of LV mass index and ejection fraction had little effect on cardiovascular outcomes. This would suggest that if PTX3 is related to vascular injury, these echocardiographic measures inadequately measure the effect of this injury in the heart. Future studies are needed to elucidate the mechanisms of these associations.
Strengths of the current study include the large sample size, the extended follow-up time, the comprehensive assessment of risk factors, and the availability of adjudicated outcome classifications. There are also several important limitations. Because the Heart and Soul cohort is recruited primarily from a veteran population and is comprised mainly of white men, our results may not generalize to more heterogeneous populations. Some misclassification may have occurred due to the fact that cardiovascular outcomes were assessed by review of medical records, and incident HF was defined by hospitalization. All participants had prevalent CHD, and results may differ in other settings.
In summary, we found that PTX3 is associated with overall mortality, CV events and incident HF among persons with stable coronary disease. These findings were independent of systemic inflammation, kidney dysfunction and traditional CV risk factors. The addition of eGFR to predictive models provided substantial attenuation, suggesting that future studies of PTX3 should take kidney function into account.
The Heart and Soul Study was supported by the Department of Veterans Affairs, the National Heart Lung and Blood Institute (R01 HL079235); the American Federation for Aging Research (Paul Beeson Scholars Program); the Robert Wood Johnson Foundation (Generalist Physician Faculty Scholars Program); and the Ischemia Research and Education Foundation, which were administered by the Northern California Institute for Research and Education, and with resources of the Veterans Affairs Medical Center, San Francisco, California. CP is currently supported by 1K23DK082793-01 (NIDDK) and the Robert Wood Johnson Harold Amos award. These funding sources had no involvement in the design or execution of this study.
Potential Conflicts of Interest
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.