Printer Friendly

Soluble ST2 is associated with all-cause and cardiovascular mortality in a population-based cohort: the Dallas heart study.

ST2, a member of the interleukin-1 (IL-1) [4] receptor family, is induced when cardiac myocytes and fibroblasts undergo mechanical strain (1). The ST2 gene encodes two isoforms: ST2 ligand (ST2L), the functionally active transmembrane form, and a shed, truncated soluble ST2 (sST2), a circulating protein detectable in human plasma (2). Both isoforms and their active ligand, IL-33, are upregulated in cardiomyocytes stimulated by mechanical stressors such as angiotensin II and phenylephrine (3). In vivo studies demonstrate that IL-33/ST2L signaling reduces cardiac myocyte hypertrophy, prevents apoptosis after ischemia/reperfusion injury, and inhibits cardiac fibrosis. These effects are reversed by administration of sST2 (3, 4).Onthe basis of these findings, sST2 is believed to be a decoy receptor, limiting the cardioprotective effects of IL-33/ ST2L activation.

sST2 has also been investigated as a potential circulating cardiac biomarker. Studies in patients with acute and chronic heart failure (HF) (5-8) and acute myocardial infarction (9, 10) have reported associations between higher plasma concentrations of sST2 and increased risk for mortality and nonfatal adverse cardiac events such as worsening heart failure, evolving infarct size, recurrent myocardial infarction, and stroke. The prognostic value of sST2 in these high-risk populations has been demonstrated to be independent of traditional markers of cardiac injury and neurohormonal activation. However, to our knowledge, no prior study has evaluated sST2 as a cardiac biomarker in the general population. In the present study, we performed a comprehensive evaluation of sST2 in the Dallas Heart Study, a large, multiethnic population-based cohort.

Materials and Methods

The Dallas Heart Study (DHS) is a probability-based population sample of Dallas County residents ages 18-65 (11). African Americans were intentionally oversampled to achieve a final cohort of approximately 50% African Americans and 50% women. This study includes 3294 DHS participants between ages 30 and 65 years with measured plasma sST2 concentrations.

Traditional risk factors were directly determined by physical exam, laboratory evaluation, or self-reported medication history, as previously described (11). History of HF was determined by self-report. Self-reported annual household income and education were used as a measurement of socioeconomic status, each categorized into 4 ordinal groups. Renal function was based on estimated glomerular filtration rate (eGFR).

We ascertained nonfatal cardiovascular events from quarterly tracking of the Dallas-Fort Worth Hospital Council Data Initiative database through 2009. This database is updated quarterly and consists of 100% of hospital admission and discharge data for 70 of 72 hospitals in the Dallas-Fort Worth metropolitan area. We requested primary records for all suspected cardiovascular events, and these events were each adjudicated separately by two cardiologists. Adjudicated events included nonfatal myocardial infarction, hospitalization for unstable angina, coronary revascularization (percutaneous revascularization or coronary artery bypass graft surgery), stroke or transient ischemic attack, peripheral artery revascularization, hospitalization due to congestive heart failure, and hospitalization for atrial fibrillation.

All-cause, cardiovascular, and cancer mortality were based on National Death Index Data through December 31,2009 (median follow-up 8.3 years; interquartile range 7.9-8.8 years). Cardiovascular death was defined with International Classification of Diseases, Revision 10 (ICD-10) codes I00-I99 (12). Noncardiovascular death excluded ICD codes I00-I99, and cancer deaths were defined with ICD-10 codes C00-C99.

Coronary artery calcification (CAC) was calculated as the average score on two consecutive electron-beam computed tomography scans (13). Aortic imaging was performed with a 1.5-Tesla whole-body MRI system (Intera, Philips Medical Systems) with a free-breathing, electrocardiograph-gated, T2-weighted turbo spin-echo (black-blood) sequence, as described previously (14). Abdominal aortic plaque, aortic wall thickness, aortic compliance, and aortic pulse wave velocity were all determined as previously described (14-18). Left ventricular mass and left ventricular ejection fraction (LVEF) were calculated by the cardiac 1.5-Tesla MRI system, as previously described (15). Personnel who performed imaging measurements were blinded to all participant data.

Samples were collected from venous blood in standard blood collection tubes containing citrate EDTA and maintained at 4[degrees]C for [less than or equal to] 4 h and then centrifuged (1430g for 15 min) at 4[degrees]C. Plasma was then removed, divided into aliquots, and frozen at -80[degrees]C until assays were performed. sST2 was measured from thawed frozen plasma at Alere with a sandwich assay on a Luminex 200 reader and modified paramagnetic Luminex beads from Radix Biosolutions with minimum and maximum detection limits of 0.40 and 200 [micro]g/L, respectively. For comparison, the limit of detection for the Presage assay (Critical Diagnostics) is 1.3 [micro]g/L. The intraassay and interassay CVs were 31% and 28% at the undetectable range and 13% and 12% in the detectable range. Personnel who performed the assays were blinded to all clinical data.

The following analytes were measured previously and the methods have been described: growth differentiation factor 15 (GDF-15) (16), cystatin C (17), high-sensitivity C-reactive protein (hsCRP) (19), pulmonary surfactant protein B (SP-B) (20), osteoprotegerin (21), N-terminal pro-B-type natriuretic peptide (NTproBNP) (22), cardiac troponin T (cTnT) measured with a highly sensitive assay (23), tumor necrosis factor-[alpha]1 receptor (TNFR1A) (20), lymphotoxin [beta] receptor (LTBR) (24), soluble vascular cell adhesion molecule 1 (sVCAM-1) (25), soluble intercellular adhesion molecule 1 (sICAM-1) (25), soluble receptor for advanced glycation end products (26), caspase-3 (18), fructosamine (22), adiponectin (27), monocyte chemoattractant protein 1 (MCP-1) (28), homeostasis model assessment-insulin resistance (25), and IL-18 (29).


Participants with sST2 concentrations below the detection limit of 0.40 [micro]g/L were categorized into the undetectable group. We grouped the remaining participants into quartiles. We compared demographic and clinical variables across increasing sST2 categories with the [chi square] trend test for categorical variables and the Jonckheere-Terpstra test for continuous variables. Correlations between selected biomarkers and sST2 were evaluated by Spearman rank correlation coefficients. We used Cox proportional hazards models to assess associations between sST2 categories and all-cause mortality, adjusted for traditional risk factors including age, sex, race, hypertension, diabetes, smoking, hypercholesterolemia, low HDL cholesterol, body mass index (BMI), and eGFR. Further adjustments were made for history of HF, log hsCRP, log NT-proBNP, log GDF-15, and quantiles of cTnT. Proportional hazards assumptions were met for all models. Metrics of biomarker performance included the Chambless c-statistic (30), time-dependent integrated discrimination index (IDI) (31), calibration with a modified Hosmer-Lemeshow [chi square] statistic (32), and category-less net reclassification index (NRI) (33). We conducted similar analyses for nonfatal cardiovascular events, cardiovascular mortality, noncardiovascular mortality, and cancer mortality.

We conducted sensitivity analyses for all models excluding participants with a history of HF. Additional sensitivity analyses were conducted with inclusion of annual household income and education as an index of socioeconomic status. All models included only subjects with complete data available for covariates and the phenotype of interest. Two-sided probability values (P value) <0.05 were considered statistically significant.



The plasma concentration of sST2 ranged from 0.40 to 28.6 [micro]g/L. The overall prevalence of detectable sST2 was 33%. The prevalence of detectable sST2 was similar by sex (P = 0.11) but was significantly higher in African American participants (44% vs 21% in non-African Americans, P < 0.0001) (Fig. 1). The proportion of individuals with detectable sST2 was higher among those with vs without a history of cardiovascular disease (CVD) (37% vs 32%, P = 0.02) and among those with vs without a history of HF (55% vs 32%, P < 0.0001), with a similar increment seen among African Americans with vs without a history of HF (60% vs 43%, P = 0.003).

Although the prevalence of several traditional cardiovascular risk factors increased among those with detectable sST2 concentrations compared to those with undetectable concentrations, there were no statistically significant trends of cardiovascular risk factors across increasing quartiles of detectable sST2, with the exception of male sex (Table 1). The proportion of individuals with detectable cTnT with the high-sensitivity cTnT assay increased significantly across sST2 categories, from 25% to 40% in the highest sST2 category (P < 0.0001). Concentrations of NT-proBNP did not significantly increase across sST2 categories, and history of HF did not correlate with increasing sST2 concentrations among those with detectable sST2 (Table 1). sST2 was positively correlated with other inflammatory and cardiovascular biomarkers in the general population, including TNFR1A, LTBR, SP-B, and GDF-15 (Table 2). sST2 was inversely correlated with sVCAM-1, sICAM-1, and caspase-3 (Table 2).


CAC and aortic wall thickness were marginally increased in those with detectable sST2 concentrations, but there was no graded relationship with increasing sST2 quartiles (Table 1). Similar findings were observed for LV mass and aortic pulse wave velocity (Table 1). sST2 was not associated with LVEF.

Traditional cardiovascular risk factors modestly contributed to variation in detectable sST2 concentrations (c-index = 0.68). Adding inflammatory markers increased the c-statistic for detectable sST2 to 0.78 (see Supplemental Table 1, which accompanies the online version of this article at content/vol59/issue3). The most influential variable was African American race, followed by TNFR1A and GDF-15 (see online Supplemental Table 1). Among participants with detectable sST2 concentrations, the multivariable risk factor model did not significantly explain meaningful variation in sST2 concentrations ([R.sup.2] from linear regression = 0.01). Moreover, in analyses restricted to African American participants, age, BMI, and left ventricular mass were only borderline associated with detectable sST2 concentrations in multivariable models, with the largest contribution from TNFR1A (see online Supplemental Table 1).


A total of 164 deaths, including 65 cardiovascular deaths and 29 deaths due to cancer, occurred over a median follow-up of 8.3 years. Unadjusted all-cause mortality increased significantly across sST2 categories, from 4.0% in the undetectable category to 11.3% in the highest sST2 category ([P.sub.trend] < 0.0001) (Table 3; Fig. 2A). Similar trends were seen for cardiovascular mortality ([P.sub.trend] = 0.0004), noncardiovascular mortality ([P.sub.trend] = 0.0003), and cancer mortality ([P.sub.trend] = 0.003) (Table 3). Sensitivity analyses excluding participants with a self-reported history of HF did not affect the trends for all-cause mortality across sST2 categories (P = 0.0001) (Fig. 2).

In unadjusted Cox proportional hazards models, participants in the highest sST2 category ([greater than or equal to] 92nd percentile) were at increased risk for all-cause mortality (hazard ratio 3.1, 95% CI 2.0-4.7; P < 0.0001) compared to those in the undetectable group (Table 4). The association with all-cause mortality was attenuated with adjustment for traditional risk factors and eGFR, but remained significant (hazard ratio 2.1,95% CI 1.4-3.2; P = 0.0009) (Table 4). Additional adjustment for history of HF (Table 4), annual income, or education levels (data not shown) did not alter the associations with all-cause death. However, additional adjustment for novel cardiovascular biomarkers, including log hsCRP, log NT-proBNP, log GDF-15, and cTnT, attenuated the associations with all-cause death (Table 4). Sensitivity analyses with detectable sST2 did not alter the above findings. sST2 concentrations were not significantly associated with all-cause death in non-African Americans (n = 35 deaths; P = 0.7). Formal testing for interaction revealed no statistical interaction between African American race and sST2 categories ([P.sub.interaction] = 0.5). Participants in the highest sST2 category were also at increased risk for cardiovascular death in unadjusted models (hazard ratio 3.3, 95% CI 1.7-6.5; P = 0.0004). Adjusted for traditional risk factors, the hazard for cardiovascular death was attenuated and borderline statistically significant (hazard ratio 1.9, 95% CI 0.99-3.9; P = 0.05). Participants in the highest sST2 category also had an increased risk for noncardiovascular death (P = 0.009) and cancer death (P = 0.018) after adjustment for traditional risk factors.

The ability of sST2 concentrations to improve prediction of all-cause death beyond traditional risk factors, including eGFR, was assessed with metrics of discrimination (c-index), calibration (Hosmer-Lemeshow P value), and reclassification (IDI and NRI) (see online Supplemental Table 2). The c-index did not change with addition of sST2 categories to traditional risk factors (c-index = 0.79 with and without sST2, P = 0.7 for comparison) but did significantly improve with the addition of log hsCRP, log NT-proBNP, log GDF15, and cTnT (c-index = 0.82, P = 0.0003 for comparison) (see online Supplemental Table 2). Adding sST2 categories to a model including traditional cardiovascular risk factors and these markers did not improve the c-index (0.82, P = 0.8 for comparison) (see online Supplemental Table 2). However, this model including sST2 was well calibrated and significantly improved reclassification (see online Supplemental Table 2). Inclusion of history of HF did not qualitatively alter these findings.


We analyzed 185 nonfatal cardiovascular events. Although the highest sST2 category was significantly associated with increased nonfatal cardiovascular events in unadjusted analyses (hazard ratio 1.9, 95% CI 1.2-3.0, P = 0.005) (Table 4), this association was attenuated and no longer significant in a fully adjusted model (HR for highest sST2 category 1.5, 95% CI 0.9-2.3, P = 0.10) (Table 4). Incident heart failure was the only nonfatal endpoint associated with sST2 (highest sST2 category: unadjusted P value = 0.007, adjusted P value = 0.06).


We report, in a large population-based study of middle-aged participants, that higher sST2 concentrations are independently associated with a 2-fold increased risk of all-cause death. However, this association was significantly attenuated after adjustment for cardiac-specific biomarkers. Unlike most cardiovascular biomarkers, sST2 concentrations do not correlate strongly with age or other traditional risk factors except for male sex, but are remarkably higher in African Americans. Our findings extend prior reported associations of higher sST2 concentrations with mortality within high-risk populations to a younger, low-risk population-based sample and are the first report of a race-specific association with sST2 concentrations within the general population.

ST2 is a member of the IL-1 receptor superfamily, which is mainly involved in regulating inflammatory responses in various chronic inflammatory conditions (34). Similar to other IL-1 receptor family members, ST2 has both a membrane-bound isoform, ST2L, and a soluble isoform, sST2 (2). In murine models, ST2 is transcriptionally activated in cardiac myocytes by mechanical strain (1). However, serum sST2 does not correlate with myocardial expression of ST2 in patients with chronic HF, a finding that contrasts notably with natriuretic peptides, which have a much closer correlation between myocardial expression and circulating concentrations (35). In addition, HF does not result in a transmyocardial gradient of soluble ST2 (35). ST2 is rapidly synthesized and released from endothelial cells through a secretory pathway in response to inflammation, supporting the notion that vascular endothelium is likely the major source of increases in circulating sST2 (35).

IL-33 is the major ligand for ST2. Within mechanically stimulated cardiomyocytes, IL-33/ST2L mediated nuclear factor (NF)-[kappa]B signaling pathway has an antihypertrophic effect on cardiomyocytes (1). However, this antihypertrophic effect is inhibited by IL-33 binding to sST2 in the circulation, suggesting that sST2 may serve to act as a decoy receptor for IL-33 signaling (1). Given its inhibitory effect on the IL-33/ST2L signaling pathway, sST2 has been studied as a potential adverse cardiovascular biomarker, particularly in heart failure and hypertensive heart disease.

Prior studies of sST2 in humans have mainly reported associations between sST2 concentrations and cardiovascular risk factors in patients with HF or other high-risk conditions (5-7, 9). In our population-based study, traditional cardiovascular risk factors were only modestly correlated with the presence vs absence of detectable sST2 concentrations and did not increase with increasing sST2 concentrations in a dose-dependent fashion. Although detectable concentrations were significantly more common among the small number of individuals with self-reported HF, this factor did not explain variation in sST2 in the multivariable models. Inflammatory markers and GDF-15 significantly associated with detectable sST2 concentrations in both the overall sample as well as in African Americans, suggesting that pathways reflected by these markers may be more important in determining circulating sST2 concentrations than traditional cardiovascular risk factors. The relatively modest association of sST2 with traditional risk factors is consistent with several prior studies in selected populations at high risk (7, 8, 10).

These findings support the notion that sST2 concentrations may be significantly influenced by genetic factors. Our observations that sST2 concentrations are markedly higher in African Americans support this concept. Detectable sST2 concentrations were present in 44% of African Americans in the Dallas Heart Study, double the prevalence in non-African Americans (21% of white and 21% of Hispanic participants), including those without self-reported HF. In addition, African American race was the strongest predictor of detectable sST2 concentrations, and traditional risk factors were not significantly associated with sST2 among African Americans. One prior study in patients with New York Heart Association class III-IV heart failure found that nonwhite patients had significantly higher baseline sST2 concentrations compared with white patients (8). We report the first race-specific association with sST2 concentrations in the general population free from HF.

Prior studies have shown that higher sST2 concentrations associate with both nonfatal cardiac events as well as mortality in high-risk patients (5-10). Our study extends these findings to a low-risk population-based sample: higher baseline sST2 concentrations are associated with increased all-cause mortality and are weakly associated with cardiovascular mortality in the general population. However, sST2 concentrations were not associated with subclinical CVD and only marginally associated with nonfatal cardiovascular events, suggesting that the increased risk of death with higher sST2 concentrations may be mediated through non-CVD phenotypes.

One putative mechanism of an increased mortality risk with increased sST2 may occur by mitigation of the antihypertrophic effect of IL-33/ST2L activation (1). This could lead to cardiac hypertrophy, a known risk factor for cardiovascular mortality (36). Inour population-based sample, detectable sST2 was correlated with LV mass/body surface area (BSA), but the association between sST2 concentration and LV mass/ BSA was no longer significant among subjects with detectable sST2 concentrations. Moreover, the adjusted hazard for mortality was not significantly attenuated with further adjustment for LVmass/BSA. Lastly, there was no association between left ventricular systolic function or NT-proBNP and sST2 concentrations. These findings suggest that the increased mortality risk with higher sST2 concentrations in the general population is not explained by LV hypertrophy or systolic dysfunction.

Another putative mechanism of increased mortality may be related to the role of IL-33/ST2L signaling in reducing atherosclerosis formation (4). In our study sample, although the prevalence of coronary calcium (Agatston score >10) was increased in those with detectable sST2 concentrations, there was no association in adjusted regression models. Similarly, subclinical measures of aortic atherosclerosis also did not correlate with sST2 concentrations.

Although sST2 concentrations were not significantly associated with subclinical cardiovascular phenotypes or cardiovascular events that confer increased mortality risk, complex and sometimes divergent associations were observed with several inflammatory markers. For example, sST2 positively associated with TNFR1, LTBR, and GDF-15, but inversely associated with sVCAM, sICAM, and MCP-1. IL-33 increases production of these endothelial adhesion molecules, supporting the inverse association with its decoy receptor, sST2 (37). IL-33 has been shown to induce both proinflammatory and anti-inflammatory cytokines, which might explain the discordant associations seen between sST2 and inflammatory markers in our study (38, 39). Clinical studies in humans have reported correlations between increased sST2 concentrations and inflammatory conditions (40), suggesting that inflammation may be one pathway by which sST2 confers increased mortality risk. In addition, in our study sST2 was significantly correlated with cTnT, a sensitive marker of cardiac injury (23), and GDF-15 (16), well-established markers of mortality not only in high risk cohorts but in healthier populations as well. The association between sST2 and mortality in our population-based sample was attenuated by adjustment for these markers (hsCRP, cTnT, and GDF-15). It is therefore likely that part of the increased mortality risk seen with increased sST2 concentrations in the general population may be explained via pathways reflected by these markers.

Our findings provide support for a potential pathophysiological role of sST2 in overall risk of death but not specifically CVD; in particular, the novel findings in African Americans suggest that the sST2 pathway may contribute to excess risk for mortality among African Americans. However, the present findings do not support a clinical role at this time for sST2 measurement to assess risk in the general population. Adjustment for cTnT (measured with the high-sensitivity cTnT assay), NT-proBNP, and the emerging cardiovascular biomarker GDF-15 significantly attenuated the associations between sST2 and mortality. Moreover, the improved reclassification (IDI and NRI) with addition of sST2 to traditional risk factors, although statistically significant, was quantitatively modest. It is important to note that the majority of individuals did not have detectable sST2 concentrations, which limited the ability to assess dose effects of sST2 with risk markers and fatal outcomes. The development of more sensitive assays is needed to fully explore the potential role of this biomarker for population screening.

Several limitations are noted. First, survival analyses were based on relatively small numbers of cardiovascular deaths in this study population, limiting statistical power. Second, the cardiovascular deaths were based on ICD-10 codes, which are prone to misclassification. Third, the main ligand of ST2, IL-33, was not measured, prohibiting complete assessment of the circulating ST2 system.

In conclusion, in the general population, plasma sST2 concentrations are significantly increased in African Americans and are associated with increased all-cause mortality and weakly associated with increased cardiovascular mortality. These associations are largely unexplained by left ventricular hypertrophy, heart failure, subclinical atherosclerosis, or nonfatal cardiac events and are likely related, in part, to increased chronic inflammation. In contrast to most other cardiovascular biomarkers, sST2 concentrations are not strongly influenced by age and other traditional risk factors, supporting the idea that sST2 concentrations may be influenced by genetic factors to a greater extent than other cardiac biomarkers. Further studies are warranted exploring race-specific associations with sST2 and the potential role of this biomarker for mortality risk assessment in the general population.

Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.

Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:

Employment or Leadership: None declared.

Consultant or Advisory Role: None declared.

Stock Ownership: None declared.

Honoraria: None declared.

Research Funding: J.A. de Lemos, Inverness.

Expert Testimony: None declared.

Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript.


(1.) Weinberg EO, Shimpo M, De Keulenaer GW, MacGillivray C, Tominaga S, Solomon SD, et al. Expression and regulation of ST2, an interleukin-1 receptor family member, in cardiomyocytes and myocardial infarction. Circulation 2002;106: 2961-6.

(2.) Iwahana H, Yanagisawa K, Ito-Kosaka A, Kuroiwa K, Tago K, Komatsu N, et al. Different promoter usage and multiple transcription initiation sites of the interleukin-1 receptor-related human ST2 gene in UT-7 and TM12 cells. Eur J Biochem 1999;264:397-406.

(3.) Sanada S, Hakuno D, Higgins LJ, Schreiter ER, McKenzie AN, Lee RT. IL-33 and ST2 comprise a critical biomechanically induced and cardioprotective signaling system. J Clin Invest 2007;117: 1538-49.

(4.) Miller AM, Xu D, Asquith DL, Denby L, Li Y, Sattar N, et al. IL-33 reduces the development of atherosclerosis. J Exp Med 2008;205:339-46.

(5.) Manzano-Fernandez S, Mueller T, Pascual-Figal D, Truong QA, Januzzi JL. Usefulness of soluble concentrations of interleukin family member ST2 as predictor of mortality in patients with acutely decompensated heart failure relative to left ventricular ejection fraction. Am J Cardiol 2011;107: 259-67.

(6.) Shah KB, Kop WJ, Christenson RH, Diercks DB, Henderson S, Hanson K, et al. Prognostic utility of ST2 in patients with acute dyspnea and preserved left ventricular ejection fraction. Clin Chem 2011; 57:874-82.

(7.) Ky B, French B, McCloskey K, Rame JE, McIntosh E, Shahi P, et al. High-sensitivity ST2 for prediction of adverse outcomes in chronic heart failure. Circ Heart Fail 2011;4:180-7.

(8.) Weinberg EO, Shimpo M, Hurwitz S, Tominaga S, Rouleau JL, Lee RT. Identification of serum soluble ST2 receptor as a novel heart failure biomarker. Circulation 2003;107:721-6.

(9.) Dhillon OS, Narayan HK, Quinn PA, Squire IB, Davies JE, Ng LL. Interleukin 33 and ST2 in non-ST-elevation myocardial infarction: Comparison with global registry of acute coronary events risk scoring and NT-proBNP. Am Heart J 2011;161: 1163-70.

(10.) Eggers KM, Armstrong PW, Califf RM, Simoons ML, Venge P, Wallentin L, James SK. ST2 and mortality in non-ST-segment elevation acute coronary syndrome. Am Heart J 2010;159:788-94.

(11.) Victor RG, Haley RW, Willett DL, Peshock RM, Vaeth PC, Leonard D, et al. The Dallas Heart Study: a population-based probability sample for the multidisciplinary study of ethnic differences in cardiovascular health. Am J Cardiol 2004;93: 1473-80.

(12.) Lloyd-Jones D, Adams RJ, Brown TM, Carnethon M, Dai S, De Simone G, et al. Heart disease and stroke statistics-2010 update: a report from the American Heart Association. Circulation 2010; 121:e46-e215.

(13.) Jain T, Peshock R, McGuire DK, Willett D, Yu Z, Vega GL, et al. African Americans and Caucasians have a similar prevalence of coronary calcium in the Dallas Heart Study. J Am Coll Cardiol 2004; 44:1011-7.

(14.) Rohatgi A, Ayers CR, Khera A, McGuire DK, Das SR, Matulevicius S, et al. The association between peptidoglycan recognition protein-1 and coronary and peripheral atherosclerosis: observations from the Dallas Heart Study. Atherosclerosis 2009;203: 569-75.

(15.) Drazner MH, Dries DL, Peshock RM, Cooper RS, Klassen C, Kazi F, et al. Left ventricular hypertrophy is more prevalent in blacks than whites in the general population: the Dallas Heart Study. Hypertension 2005;46:124-9.

(16.) Rohatgi A, Patel P, Das SR, Ayers CR, Khera A, Martinez-Rumayor A, et al. Association of growth differentiation factor-15 with coronary atherosclerosis and mortality in a young, multi-ethnic population: observations from the Dallas Heart Study. Clin Chem 2011.

(17.) Patel PC, Ayers CR, Murphy SA, Peshock R, Khera A, de Lemos JA, et al. Association of cystatin C with left ventricular structure and function: the Dallas Heart study. Circ Heart Fail 2009;2:98 104.

(18.) Matulevicius S, Rohatgi A, Khera A, Das SR, Owens A, Ayers CR, et al. The association between plasma caspase-3, atherosclerosis, and vascular function in the Dallas Heart Study. Apoptosis 2008;13:1281-9.

(19.) Khera A, McGuire DK, Murphy SA, Stanek HG, Das SR, Vongpatanasin W, et al. Race and gender differences in C-reactive protein levels. J Am Coll Cardiol 2005;46:464-9.

(20.) Nguyen AB, Rohatgi A, Garcia CK, Ayers CR, Das SR, Lakoski SG, et al. Interactions between smoking, pulmonary surfactant protein B, and atherosclerosis in the general population: the Dallas Heart Study. Arterioscler Thromb Vasc Biol 2011; 31:2136-43.

(21.) Abedin M, Omland T, Ueland T, Khera A, Aukrust P, Murphy SA, et al. Relation of osteoprotegerin to coronary calcium and aortic plaque (from the Dallas Heart Study). Am J Cardiol 2007;99: 513-8.

(22.) Abdullah SM, Khera A, Das SR, Stanek HG, Can ham RM, Chung AK, et al. Relation of coronary atherosclerosis determined by electron beam computed tomography and plasma levels of N-terminal pro-brain natriuretic peptide in a multiethnic population-based sample (the Dallas Heart Study). Am J Cardiol 2005;96:1284-9.

(23.) de Lemos JA, Drazner MH, Omland T, Ayers CR, Khera A, Rohatgi A, et al. Association of troponin T detected with a highly sensitive assay and cardiac structure and mortality risk in the general population. JAMA 2010;304:2503-12.

(24.) Owens AW, Matulevicius S, Rohatgi A, Ayers CR, Das SR, Khera A, et al. Circulating lymphotoxin beta receptor and atherosclerosis: observations from the Dallas Heart Study. Atherosclerosis 2010;212:601-6.

(25.) Rohatgi A, Owens AW, Khera A, Ayers CR, Banks K, Das SR, et al. Differential associations between soluble cellular adhesion molecules and atherosclerosis in the Dallas Heart Study: a distinct role for soluble endothelial cell-selective adhesion molecule. Arterioscler Thromb Vasc Biol 2009;29: 1684-90.

(26.) Lindsey JB, de Lemos JA, Cipollone F, Ayers CR, Rohatgi A, Morrow DA, et al. Association between circulating soluble receptor for advanced glycation end products and atherosclerosis: observations from the Dallas Heart Study. Diabetes Care 2009;32:1218-20.

(27.) Turer AT, Khera A, Ayers CR, Turer CB, Grundy SM, Vega GL, Scherer PE. Adipose tissue mass and location affect circulating adiponectin levels. Diabetologia 2011;54:2515-24.

(28.) de Lemos JA, Morrow DA, Blazing MA, Jarolim P, Wiviott SD, Sabatine MS, et al. Serial measurement of monocyte chemoattractant protein-1 after acute coronary syndromes: results from the A to Z trial. J Am Coll Cardiol 2007;50:2117-24.

(29.) Zirlik A, Abdullah SM, Gerdes N, MacFarlane L, Schonbeck U, Khera A, et al. Interleukin-18, the metabolic syndrome, and subclinical atherosclerosis: results from the Dallas Heart Study. Arterioscler Thromb Vasc Biol 2007;27:2043-9.

(30.) Chambless LE, Cummiskey CP, Cui G. Several methods to assess improvement in risk prediction models: extension to survival analysis. Stat Med 2011;30:22-38.

(31.) Pencina MJ, D'Agostino RB Sr, D'Agostino RB Jr, Vasan RS. Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond. Stat Med 2008; 27:157-72; discussion 207-12.

(32.) D'Agostino RB, Nam B-H. Evaluation of the performance of survival analysis models: discrimination and calibration measures. In: Balakrishnan N, Rao CR, eds. Handbook of statistics, Vol. 23. New York: Elsevier 2003;1-25.

(33.) Pencina MJ, D'Agostino RB Sr, Steyerberg EW. Extensions of net reclassification improvement calculations to measure usefulness of new biomarkers. Stat Med 2011;30:11-21.

(34.) Kakkar R, Lee RT. The IL-33/ST2 pathway: therapeutic target and novel biomarker. Nat Rev Drug Discov 2008;7:827-40.

(35.) Bartunek J, Delrue L, Van Durme F, Muller O, Casselman F, De Wiest B, et al. Nonmyocardial production of ST2 protein in human hypertrophy and failure is related to diastolic load. J Am Coll Cardiol 2008;52:2166-74.

(36.) Havranek EP, Froshaug DB, Emserman CD, Hanratty R, Krantz MJ, Masoudi FA, et al. Left ventricular hypertrophy and cardiovascular mortality by race and ethnicity. Am J Med 2008;121: 870-5.

(37.) Demyanets S, Konya V, Kastl SP, Kaun C, Rauscher S, Niessner A, et al. Interleukin-33 induces expression of adhesion molecules and inflammatory activation in human endothelial cells and in human atherosclerotic plaques. Arterioscler Thromb Vasc Biol 2011;31:2080-9.

(38.) Chackerian AA, Oldham ER, Murphy EE, Schmitz J, Pflanz S, Kastelein RA. IL-1 receptor accessory protein and ST2 comprise the IL-33 receptor complex. J Immunol 2007;179:2551-5.

(39.) Schmitz J, Owyang A, Oldham E, Song Y, Murphy E, McClanahan TK, et al. IL-33, an interleukin-1-like cytokine that signals via the iL-1 receptor-related protein ST2 and induces T helper type 2-associated cytokines. Immunity 2005;23:479 90.

(40.) Mok MY, Huang FP, Ip WK, Lo Y, Wong FY, Chan EY, et al. Serum levels of IL-33 and soluble ST2 and their association with disease activity in systemic lupus erythematosus. Rheumatology (Oxford) 2010;49:520-7.

Lu Q. Chen, [1] James A. de Lemos, [1,2] Sandeep R. Das, [1,2] Colby R. Ayers, [2,3] and Anand Rohatgi [1,2] *

[1] Department of Internal Medicine; [2] Division of Cardiology, Donald W. Reynolds Cardiovascular Research Center, Dallas, TX; [3] Department of Clinical Sciences, University of Texas Southwestern Medical Center, Dallas, TX.

* Address correspondence to this author at: UT Southwestern Medical Center, 5323 Harry Hines Blvd., E5.730, Dallas, TX 75390-8830. Fax 214-645-2480; e-mail:

Received June 12, 2012; accepted October 29, 2012.

Previously published online at DOI: 10.1373/clinchem.2012.191106

[4] Nonstandard abbreviations: IL-1, interleukin-1; ST2L, ST2 ligand; sST2, soluble ST2; HF, heart failure; DHS, Dallas Heart Study; eGFR, estimated glomerular filtration rate; CVD, cardiovascular disease; ICD-10, International Classification of Diseases, Revision 10; CAC, coronary artery calcification; LVEF, left ventricular ejection fraction; GDF-15, growth differentiation factor 15; hsCRP, high-sensitivity C-reactive protein; SP-B, pulmonary surfactant protein B; NT-proBNP, N-terminal pro-B-type natriuretic peptide; cTnT, cardiac troponin T; TNFR1A, tumor necrosis factor-[alpha]1 receptor; LTBR, lymphotoxin [beta] receptor; sVCAM-1, soluble vascular cell adhesion molecule 1; sICAM-1, soluble intercellular adhesion molecule 1; MCP-1, monocyte chemoattractant protein 1; BMI, body mass index; IDI, integrated discrimination index; NRI, net reclassification index; NF-[kappa]B, nuclear factor-KB; BSA, body surface area.
Table 1. Demographic characteristics, cardiovascular risk
factors, and cardiac phenotypes across increasing categories of

Variables                            Undetectable      Q1

n                                    2207           272
sST2, [micro]g/L                       <0.40        0.40-0.58
Age, years                             43            45
Male, %                                43            40
White, %                               35            24
African American, %                    43            65
Hispanic, %                            20             9
BMI, kg/[m.sup.2]                      28.2          28.8
Hypertension, %                        30            43
Diabetes, %                            11            15
Current smoker, %                      29            28
Hypercholesterolemia, %                12            17
History of HF, %                        2.4           7.7
Education, %
  Less than high school                20.2          23.8
  High school                          23.7          23.4
  Some college                         25.2          22.4
  College or more                      30.9          30.4
Income, %
  <$16K                                20.7          16.9
  $16-30K                              30.3          36
  $30-50K                              27.2          27.2
  [greater than or equal to] $50K      21.9          19.9
Cholesterol, mg/dL (a)
  Total                               176           178
  LDL                                 104           103
  HDL                                  47            49
Triglycerides, mg/dL                   98            92
LV mass/BSA, g/[m.sup.2]               78.7          81.9
LVEF, %                                73            73
LVEF <40%, %                            0.3           1.4
CAC >10 units, %                       20            25
Aortic compliance, mL/mmHg             24.2          20.7
Aortic wall thickness, mm               1.65          1.65
Pulse wave velocity, m/s                4.22          4.27
Aortic plaque prevalence, %            38            44
eGFR, mL x [min-.sup.1] x
  [(1.73 [m.sup.2]).sup.-1]            97.9          98.7

Variables                               Q2          Q3          Q4

n                                    271         272         272
sST2, [micro]g/L                     0.59-0.86   0.87-1.43   1.44-28.6
Age, years                            42          44          44
Male, %                               45          46          54
White, %                              19          19          14
African American, %                   66          70          76
Hispanic, %                           14          11           9
BMI, kg/[m.sup.2]                     28.9        29.8        28.1
Hypertension, %                       43          40          40
Diabetes, %                            9          17          16
Current smoker, %                     26          32          31
Hypercholesterolemia, %               17          14          14
History of HF, %                       5.2         4.4         7.4
Education, %
  Less than high school               20.5        26.8        28.4
  High school                         27.7        25          22.1
  Some college                        23.2        22.4        26.6
  College or more                     28.6        25.9        23
Income, %
  <$16K                               14.9        20.6        21.4
  $16-30K                             30.5        36.3        29.4
  $30-50K                             32.1        24.4        31.3
  [greater than or equal to] $50K     22.5        18.7        17.9
Cholesterol, mg/dL (a)
  Total                              177.5       179         176.5
  LDL                                103         104         102
  HDL                                 48          48          48.5
Triglycerides, mg/dL                  91.5        98          97
LV mass/BSA, g/[m.sup.2]              85.6        83.8        85.2
LVEF, %                               72          73          73
LVEF <40%, %                           0.5         0.5         1.0
CAC >10 units, %                      22          26          24
Aortic compliance, mL/mmHg            22.8        21.8        22.5
Aortic wall thickness, mm              1.7         1.63        1.69
Pulse wave velocity, m/s               4.38        4.69        4.50
Aortic plaque prevalence, %           38          38          40
eGFR, mL x [min-.sup.1] x
  [(1.73 [m.sup.2]).sup.-1]           93.5        97.4        98.7

                                     [P.sub.trend]   P (detectable vs
Variables                            (Q1-Q4)         undetectable)

sST2, [micro]g/L
Age, years                           0.52            0.02
Male, %                              0.001           0.13
White, %                             0.004           <0.0001
African American, %                  0.004           <0.0001
Hispanic, %                          0.69            <0.0001
BMI, kg/[m.sup.2]                    0.86            0.03
Hypertension, %                      0.27            <0.0001
Diabetes, %                          0.34            0.002
Current smoker, %                    0.15            0.86
Hypercholesterolemia, %              0.23            0.004
History of HF, %                     0.38            <0.0001
Education, %
  Less than high school              0.001           0.005
  High school                        0.87            0.64
  Some college                       0.72            0.38
  College or more                    0.007           0.04
Income, %
  <$16K                              0.64            0.14
  $16-30K                            0.39            0.11
  $30-50K                            0.31            0.35
  [greater than or equal to] $50K    0.10            0.16
Cholesterol, mg/dL (a)
  Total                              0.99            0.31
  LDL                                0.53            0.67
  HDL                                0.97            <0.0001
Triglycerides, mg/dL                 0.34            0.02
LV mass/BSA, g/[m.sup.2]             0.13            <0.0001
LVEF, %                              0.41            0.12
LVEF <40%, %                         0.68            0.05
CAC >10 units, %                     0.99            0.03
Aortic compliance, mL/mmHg           0.60            0.0004
Aortic wall thickness, mm            0.83            0.08
Pulse wave velocity, m/s             0.91            0.002
Aortic plaque prevalence, %          0.5             0.4
eGFR, mL x [min-.sup.1] x
  [(1.73 [m.sup.2]).sup.-1]          0.28            0.62

(a) To convert cholesterol concentrations to mmol/L,
multiply by 0.02586.

Table 2. Overall and race-specific Spearman correlations between
continuous markers and sST2.

                                 Overall         African American

                            [rho]    P value     [rho]    P value

TNFR1                        0.36    <0.0001     0.43     <0.0001
LTBR                         0.31    <0.0001     0.31     <0.0001
SP-B                         0.31    <0.0001     0.27     <0.0001
GDF-15                       0.30    <0.0001     0.33     <0.0001
sVCAM-1                      -0.25   <0.0001    -0.27     <0.0001
Cystatin C                   -0.24   <0.0001    -0.27     <0.0001
sICAM-1                      -0.24   <0.0001    -0.28     <0.0001
LV mass/BSA                  0.12    <0.0001     0.10      0.001
Systolic blood pressure      0.12    <0.0001     0.06      0.01
sRAGE (a)                    -0.10   <0.0001     0.04      0.12
eGFR                         -0.10    0.62      -0.07      0.04
LV concentricity             0.10    <0.001      0.05      0.07
OPG                          0.09    <0.0001     0.08      0.001
Caspase-3                    -0.08   <0.0001    -0.12     <0.0001
HDL cholesterol              0.08    <0.0001     0.04      0.07
cTnT                         0.08    <0.0001     0.07      0.006
CAC                          0.07     0.0003     0.03      0.23
Fructosamine                 0.07     0.0001     0.06      0.01
Adiponectin                  -0.06    0.0004     0.001     0.98
Aortic PWV                   0.06     0.003      0.09      0
MCP-1                        -0.05    0.002     -0.02      0.41
hsCRP                        0.05     0.003      0.04      0.07
HOMA-IR                      0.05     0.02       0.02      0.49
Insulin                      0.04     0.02       0.02      0.49
IL-18                        0.04     0.06       0.06      0.06
Triglycerides                -0.04    0.02       0.02      0.45
Total cholesterol            0.02     0.36       0.04      0.12
Age                          0.04     0.03       0.03      0.25
BMI                          0.04     0.03      -0.005     0.85
LDL cholesterol              0.00     0.86       0.01      0.58
NT-proBNP                    0.00     0.74       0.02     0.35

                           Non-African American

                            [rho]    P value

TNFR1                       0.33     <0.0001
LTBR                        0.36     <0.0001
SP-B                        0.27     <0.0001
GDF-15                      0.25     <0.0001
sVCAM-1                    -0.25     <0.0001
Cystatin C                 -0.24     <0.0001
sICAM-1                    -0.21     <0.0001
LV mass/BSA                 0.06      0.04
Systolic blood pressure     0.03      0.25
sRAGE (a)                  -0.06      0.03
eGFR                       -0.02      0.53
LV concentricity            0.04      0.16
OPG                         0.03      0.17
Caspase-3                  -0.11     <0.0001
HDL cholesterol             0.01      0.58
cTnT                        0.001     0.96
CAC                         0.05      0.07
Fructosamine                0.004     0.89
Adiponectin                -0.04      0.1
Aortic PWV                  0.05      0.07
MCP-1                      -0.05      0.03
hsCRP                      -0.01      0.66
HOMA-IR                     0.02      0.49
Insulin                     0.02      0.66
IL-18                       0.08      0.02
Triglycerides               0.02      0.36
Total cholesterol           0.03      0.26
Age                        -0.003     0.9
BMI                         0.03      0.32
LDL cholesterol             0.02      0.54
NT-proBNP                   0.00      0.80

(a) sRAGE, soluble receptor for advanced glycation end products;
OPG, osteoprotegerin; PWV, pulse wave velocity; HOMA-IR,
homeostasis model assessment-insulin resistance.

Table 3. Incident events (%) by sST2 categories in the DHS.

                             n    Undetectable

All-cause death            164        4.0
Cardiovascular death        65        1.5
Noncardiovascular death     99        2.4
Cancer death                29        0.6
Nonfatal cardiovascular
  events                   185        6.1

                                       Quartiles of detectable
                                          sST2, [micro]g/L

                          0.40-0.58  0.59-0.86  0.87-1.43  1.44-28.63

All-cause death              7.4        5.5        7.4        11.3
Cardiovascular death         3.1        1.6        3.5        4.7
Noncardiovascular death      4.3        3.9        3.9        6.6
Cancer death                 1.6        1.6        1.2        2.3
Nonfatal cardiovascular
  events                    10.6        5.9        7.4        10.7


All-cause death           <0.0001
Cardiovascular death       0.0004
Noncardiovascular death    0.0003
Cancer death               0.003
Nonfatal cardiovascular
  events                   0.02

Deaths ascertained from unadjudicated National Death Index ICD-10
codes over 8.3 years' follow-up through 2009: cardiovascular death
= 100-199; cancer = C00-C99. Nonfatal cardiovascular events
adjudicated from chart review through 2009: MI, coronary
revascularization, stroke, peripheral artery revascularization,
chronic HF hospitalization, and atrial fibrillation

Table 4. Association between sST2, mortality, and nonfatal
cardiovascular events. (a)

                                         Quartiles of detectable
                                             sST2, [micro]g/L

                          Undetectable          0.40-0.58
All-cause mortality
  Unadjusted              Reference      1.9 (1.2-3.2), P = 0.01
  Model 1                 Reference      1.5 (0.8-2.5), P = 0.13
  Model 2                 Reference      1.3 (0.8-2.3), P = 0.23
  Model 3                 Reference      1.4 (0.8-2.3), P = 0.19
Nonfatal cardiovascular
  Unadjusted              Reference      1.8 (1.0-2.8), P = 0.01
  Model 1                 Reference      1.2 (0.8-2.0), P = 0.37

                          Quartiles of detectable
                              sST2, [micro]g/L

All-cause mortality
  Unadjusted              1.4 (0.8-2.4), P = 0.25
  Model 1                 1.1 (0.6-2.0), P = 0.69
  Model 2                 1.1 (0.6-1.9), P = 0.76
  Model 3                 0.9 (0.5-1.7), P = 0.75
Nonfatal cardiovascular
  Unadjusted              1.0 (0.5-1.7), P = 0.92
  Model 1                 0.8 (0.5-1.5), P = 0.51

                          Quartiles of detectable
                              sST2, [micro]g/L

All-cause mortality
  Unadjusted              1.9 (1.2-3.1), P = 0.01
  Model 1                 1.6 (0.9-2.6), P = 0.09
  Model 2                 1.4 (0.8-2.3), P = 0.19
  Model 3                 1.2 (0.7-2.0), P = 0.48
Nonfatal cardiovascular
  Unadjusted              1.2 (0.7-2.1), P = 0.46
  Model 1                 1.0 (0.6-1.7), P = 0.94

                           Quartiles of detectable
                               sST2, [micro]g/L

All-cause mortality
  Unadjusted              3.1 (2.0-4.7), P <0.0001
  Model 1                 2.1 (1.4-3.2), P = 0.0009
  Model 2                 1.8 (1.2-2.8), P = 0.007
  Model 3                 1.4 (0.9-2.3), P = 0.13
Nonfatal cardiovascular
  Unadjusted              1.9 (1.2-3.0), P = 0.005
  Model 1                 1.5 (0.9-2.3), P = 0.10
(a) Data are hazard ratios (95% CI) for sST2 categories with
undetectable as referent derived from Cox proportional hazards
models for all-cause death and composite nonfatal events (MI,
coronary revascularization, unstable angina, stroke, peripheral
artery revascularization, CHF hospitalization, and atrial
fibrillation hospitalization). Proportional hazards assumptions
were met for all models. Model 1: age, sex, race, hypertension,
diabetes, smoking, hypercholesterolemia, low HDL cholesterol, BMI,
eGFR (164 deaths; 185 nonfatal cardiovascular events). Model 2:
model 1 plus history of HF. Model 3: model 2 plus log hsCRP, log
NT-proBNP, log GDF-15, and cTnT categories.

Fig. 1. Prevalence of detectable sST2 (>0.40 [micro]g/L) by
sex and race.

Prevalence of
detectable sST2 (%)

P value = 0.11

Men        35%
Women      32%

P value < 0.0001

Black      44%
White      21%

Hispanic   21%

Note: Table made from bar graph.

Fig. 2. Kaplan--Meier survival curve for sST2 and all-cause
mortality in the DHS (A, total number of deaths 164) and
population excluding patients with chronic HF (B, total number of
deaths 130).

A. Overall population

Numbers at risk    Year O    Year 2    Year 4    Year 6

Undetectable         2154      2140      2117      2093
Quartile 1            262       256       251       248
Quartile 2            262       260       254       252
Quartile 3            263       259       256       251
Quartile 4            262       256       245       239

B. Population excluding CHF

Numbers at risk    Year O    Year 2    Year 4    Year 6

Undetectable         2102      2089      2069      2047
Quartile 1            247       243       239       237
Quartile 2            247       245       240       239
Quartile 3            247       243       241       239
Quartile 4            247       242       236       230
COPYRIGHT 2013 American Association for Clinical Chemistry, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2013 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Lipids, Lipoproteins, and Cardiovascular Risk Factors
Author:Chen, Lu Q.; de Lemos, James A.; Das, Sandeep R.; Ayers, Colby R.; Rohatgi, Anand
Publication:Clinical Chemistry
Article Type:Report
Date:Mar 1, 2013
Previous Article:Diurnal Rhythm rather than Dietary Iron Mediates Daily hepcidin variations.
Next Article:Quantification of globotriaosylsphingosine in plasma and urine of fabry patients by stable isotope ultraperformance liquid chromatography-tandem mass...

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters