hospitalized for heart failure Rongcheng Zhang1, Yuhui Zhang1, Tao An1, Xiao Guo1, Shijie Yin1, Yunhong Wang1, James L Januzzi2, Thomas P Cappola3 & Jian Zhang*,1 1State Key Laboratory of Cardiovascular Disease, Heart Failure Center,Fuwai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences & Peking Medical College, Beijing, 100037,
China 2Division of Cardiology, Massachusetts General Hospital, 55 Fruit St, Boston,MA 02114, USA 3Penn Cardiovascular Institute, University of Pennsylvania School of Medicine, 415 Curie Blvd, Philadelphia, PA 19104,USA
Aim: To evaluate the associations of soluble ST2 (sST2) and galectin-3 with death relative to renal function in patients with heart failure (HF).
Methods: Eleven-hundredand-sixty-one patients hospitalized for HF with 1-year follow up were enrolled for biomarkers analysis.
Results: Patients were divided into two groups based on eGFR of either > or ≤60 ml/min/1.73 m2. sST2 was independently associated with death in both categories of renal function, while galectin-3 lost this significance after addition of NT-proBNP to the model of patients with eGFR ≤60 ml/min/1.73 m2.
Conclusion:In patients with HF, sST2 improved prediction for death beyond risk factors without
being influenced by renal function, however, the prognostic value of galectin-3 is less clear below an eGFR of 60 ml/min/1.73 m2.
Keywords: biomarker • cardiac remodeling • death • eGFR • galectin-3 • heart failure
• NT-proBNP • prognosis • renal function • sST2
Individuals with heart failure (HF) often experience deterioration in renal function, which also predicts more serious clinical outcome [1,2]. Biological markers reflecting pathophysiological process of HF have the potential to predict adverse events. However, the expressions of many biomarkers are not only associated with HF, but also related to kidney dysfunction. The influence of this comorbidity so frequently associated with HF should be considered when biomarkers
are used for risk prediction.
Currently, soluble ST2 (sST2) and galectin-3 are two promising biomarkers that have
been shown to be associated with adverse events in patients with HF . As risk predictors, both biomarkers have been reported to be correlated with renal function [4,5]. The relative effect of renal failure on each of these two risk markers remains unclear, relative to their comparative ability to prognosticate. Bayes-Genis et al. recently showed that the prognostic value of sST2 did not appear to be influenced by renal function , yet they did not compare the predictive ability of sST2 in specified groups of renal dysfunction, nor did they compare sST2 to galectin-3. Therefore, the purpose of this study was to investigate the interaction between prognostic value of sST2 and galectin-3 and renal function in patients hospitalized for HF, and compared the result of these two biomarkers in relation to clinical risk factors with or without N-terminal pro-B type natriuretic peptide (NT-proBNP) after considering patients according to accepted estimated glomerular filtration rate (eGFR) categories.
Materials & methods
Study population & design
We consecutively enrolled patients admitted to FuWai Hospital HF center, Beijing (China) from March 2009 to June 2012 with HF as their primary diagnosis. As reported in our previous studies [7,8], hospitalized with HF was defined as a de novo presentation of HF or worsening of previously chronic stable HF requiring unplanned hospitalization. Patients were enrolled if they were 18 years or older and had venous blood sample available for biomarker analysis. Patients with a diagnosis of acute coronary syndrome, cancer and acute pulmonary embolism were excluded from this analysis. Clinical data including demographic characteristics, New York Heart Association (NYHA) functional class, primary HF etiologies, physical examination, pre-existing comorbidities and medical history were obtained at the time of the hospitalization. Creatinine obtained from laboratory examinations within 12 h of hospitalization was used to calculate eGFR
with the Chronic Kidney Disease Epidemiology Collaboration equation . All patients had received intravenous loop diuretics at least during the first 24 h of admission. Left ventricular ejection fraction (LVEF) was assessed and interpreted by specialists trained in cardiac ultrasonagraphy by using echocardiography within 48 h after admission. Echocardiographers and
other study staff were blinded to sST2 and galectin-3 values. The first results of chest radiography during hospitalization were also collected. For patients with multiple admissions, only the first admission was included in this study.
All-cause death was ascertained every 3 months via electronic hospital records follow-up or conversations with patients or patients’ families by telephone. All patients provided written informed consent and the ethics committee of FuWai Hospital approved the study.
Fasting venous blood sample was collected within 12 h of hospitalization, immediately centrifuged and stored at -80°C in plasma using EDTA. All biomarkers were determined by blood samples with no more than one freeze-thaw cycle. Professional laboratory technicians who performed the measurement of biomarkers were blind to demographic characteristics of patients and follow-up results. sST2 was measured by a high-sensitive sandwich immunoassay (Critical Diagnostics, CA, USA). Galectin-3 was measured by enzyme-linked immunosorbant assay in a microtiter plate format (BG Medicine, MA, USA). NT-proBNP was measured by the fluorescence immunoassay using the Triage® Meter (Alere, Inc., CA, USA). Detailed information of these
three assays has been reported previously [7,8].
Continuous variables were tested for normal distribution by using Kolmogorov–Smirnov test and were described as means ± SD for normally distributed variables and medians and interquartile range (IQR) for variables with skewed distribution. Categorical variables were described as percentages. Comparisons of biomarkers between two groups were performed by Mann–Whitney U test, and Kruskal–Wallis H testing was used to compare more than two groups. Ln transformation was performed to normalize the distribution of NT-proBNP, sST2 and galectin-3.
The relationships between studied biomarkers and eGFR were analyzed using Spearman correlation. Multivariable linear regression analysis was performed with stepwise method to see if renal function is an independent predictor of either ST2 or gal-3 concentrations.
Two sets of models were predefined: model l (clinical risk factors including sex, age, diabetes mellitus,ischemic heart disease, prior heart failure, systolic blood pressure, NYHA, LVEF, angiotensin-converting enzyme inhibitor or angiotensin II receptor blocker [ACEI/ARB] treatment, β-blocker treatment,hemoglobin and eGFR); and model 2 (model 1 + NTproBNP). In this study, patients were divided into two groups (eGFR ≥60 ml/min/1.73 m2 and eGFR<60 ml/min/1.73 m2) based on National Kidney Foundation Kidney Disease Outcomes Quality Initiative (NKF KDOQI) recommendations, which define‘moderate or more’ renal dysfunction on the basis of
an eGFR <60 ml/min/1.73 m2 . Receiver operating characteristic (ROC) curves were performed to determine the prognostic ability of biomarkers for 1-year mortality, and the optimum cut-off points were identified using the Youden approach. Then,the influence of renal function on prognostic value of sST2 and galectin-3 was analyzed by Cox regression in each group. Hosmer–Lemeshow statistic was used to evaluate model calibration. Differences in Harrell’s
C-statistic, and integrated discrimination improvement (IDI) were performed to evaluate the added predictive value of biomarkers to reference model. Confidence intervals and p-values for IDI were determined by bootstrapping with 1000 repetitions. All p-values of less than 0.05 from two-sided tests were accepted as statistically significant. Statistical analyses were conducted using SPSS version 19.0 (SPSS, Inc., IL, USA) and Stata version 11.2 (StataCorp LP,TX, USA).
Supplementary Figure 1 (see online at http://www.futuremedicine.com/doi/full/10.2217/BMM.15.12) details the study flow: a total of 1540 patients with diagnosis of HF were admitted between March 2009 and July 2012. A total of 167 patients were excluded and 212 patients had no blood sample for all biomarker analysis (18 patients did not provide the written informed consent, and 194 patients did not have enough blood sample for all biomarker measurement).Out of the remaining 1161 patients, 175 patients died and 15 patients were lost during 1-year follow-up. Base- line characteristics of the included patients are shown in Table 1. The mean age was 58 ± 15 years, most of patients were male (70.6％), the mean eGFR was 75.7± 33.4 ml/min/1.73 m2, 55％ of the patient had LVEF >40％. Approximately a half of patients had a history of ischemic heart disease. The median (IQR) sST2 and galectin-3 concentration was 36.9 ng/ml (25.2–55.0) and 21.6 ng/ml (16.4–29.5), respectively.
Association of sST2 & galectin-3 with renal
When dichotomized renal function according to eGFR of 60 ml/min/1.73 m2, across NYHA grouping,sST2 was only significantly higher in NYHA Class IV patients with impaired renal function (Figure 1A),whereas galectin-3 concentrations were routinely significantly higher in patients with lower value of eGFR across all NYHA groups (Figure 1B).
In unadjusted correlation, baseline concentrations of both biomarkers were significantly correlated with the severity of renal function: r-value between eGFR and sST2 was -0.10 (p < 0.001) (Figure 2A), while for galectin-3 it was -0.46 (p < 0.001) (Figure 2B). In comparison, NT-proBNP had an r value of -0.21 (p < 0.001) when correlated to renal function in this population. Multivariable linear regression analysis including clinical risk factors and NT-proBNP shown renal function only is an independent predictor of gal-3 concentrations (β = -0.006; p < 0.001) (Supplementary Table 1).
Interaction between renal function
& prognostic value of biomarkers
In Cox regression analysis, both markers were prognostic across the group as a whole, regardless of adjustment for NT-proBNP (Table 2). We tested if renal function has interactions with prognostic value of sST2 and galec tin-3 for predicting mortality. Considering the effect of
renal function (eGFR >60 ml/min/1.73 m2 vs eGFR≤60 ml/min/1.73 m2), we found no significant interaction term between eGFR and the prognostic value of sST2 (p = 0.761), whereas the interaction between eGFR and galectin-3 was significant (p = 0.007).
Association of biomarkers with death
according to renal function
sST2 was independently associated with 1-year allcause mortality in hospitalized patients with HF when adjusted by clinical risk factors and NT-proBNP, irrespective of eGFR (for patients with eGFR ≥60 ml/min/1.73 m2, per log unit, hazard ratio [HR]: 2.08, 95％ CI: 1.58–2.73, p < 0.001; and for patients with eGFR <60 ml/min/1.73 m2, per log unit, HR: 2.32, 95％ CI:1.71–3.14, p < 0.001). In contrast, galectin-3 lost its statistically significant link with all-cause mortality when NT-proBNP was included in multivariable analysis in patients with GFR <60 ml/min/1.73 m2 (per log unit,
HR: 1.73, 95％ CI: 0.97–3.06, p = 0.063) (Table 2).
Performance of biomarkers as discriminators of
death according to renal function
ROC was performed to evaluate the risk discrimination and optimal prognostic value of sST2 and galectin-3 as a function of eGFR. The AUC for sST2 was 0.79 (95％ CI: 0.73–0.84), which was not different from that of galectin-3 (AUC = 0.74; 95％ CI: 0.69–0.79) in patients with eGFR ≥60 ml/min/1.73 m2 (p = 0.175). Both of them were also not different from NT-proBNP (AUC = 0.77; 95％ CI: 0.71–0.82; p = 0.481 vs sST2; p = 0.485 vs galectin-3) (Figure 3A).With respect to patients with eGFR <60 ml/min/1.73 m2, baseline sST2 concentrations had a similar AUC (AUC = 0.74; 95％ CI: 0.68–0.81) to that of NTproBNP (AUC = 0.75; 95％ CI: 0.68–0.81; p = 0.918), but galectin-3 had a lower AUC (AUC = 0.61; 95％ CI: 0.53–0.68) compared with other two biomarkers (p = 0.004 vs sST2; p = 0.003 vs NT-proBNP) (Figure 3B). The optimal cut-off points for predicting 1-year mortality according to eGFR are shown in Supplementary Table 2; worsening renal function had a
negligible effect on optimal sST2 risk thresholds.
Discrimination & reclassification of biomarkers365医学网 转载请注明
to model according to renal function365医学网 转载请注明
The addition of sST2 to model with clinical risk factors significantly improved C-statistic and IDI for the prediction of death in hospitalized patients with HF, regardless of eGFR. This significant result even remained robust for sST2 when incorporation of NTproBNP to the model (Table 3). However, galectin-3 only provided additional prognostic value to model in patients eGFR ≥60 ml/min/1.73 m2 (Table 3): the C-statistic and IDI was improved after inclusion of galectin-3 into model with clinical risk factors; only C-statistic increased when NT-proBNP was included to the model.
This study focuses on the interaction between renal function and prognostic value of sST2 and galectin-3 with regard to 1-year all-cause death in a large cohort of hospitalized patients with HF. After dichotomizing patients according to eGFR of 60 ml/min/1.73 m2, we found sST2 was strongly associated with death,regardless of eGFR or NT-proBNP, whereas the asso ciation between galectin-3 and death was not significant when NT-proBNP was entered in the model in365医学网 转载请注明
patients with eGFR <60 ml/min/1.73 m2. Both sST2 and galectin-3 significantly improved prognostic yield of prediction models in patients with eGFR ≥60 ml/ min/1.73 m2, but only sST2 improved discrimination and reclassification of models while galectin-3 did not show any significant result in this context. Most curiously, we found that galectin-3 had a strong interaction term with renal function, which explains the loss of statistical accuracy in those with more than moderate renal insufficiency. Given that galectin-3 might be a potential biomarker of renal function in patients with HF.
sST2, a member of the IL-1 receptor family, competitively inhibits the cardioprotective function of the membrane-bound ST2 ligand (ST2L) by acting as a decoy receptor for circulating IL-33 . Abnormalities of the ST2 system (including high levels of sST2) are tightly associated with HF risk across a wide range of patient types. Galectin-3 has also been shown to be overexpressed by activated macrophages and associated with cardiac remodeling and occurrence of HF . Although these two biomarkers expressed in a wide range of tissues [13,14], sST2 has been mainly reported to be associated with fibrosis of heart and blood vessels while galectin-3 had been shown to promote fibrosis of many organs, such as heart, pancreas, liver and kidneys [15–17]. In this context, concentrations of galectin-3 could theoretically be more easily affected by renal function than sST2, and higher galectin-3 concentrations might not totally reflect the presence of cardiac dysfunction in patients with renal dysfunction. Dieplinger et al. reported that sST2 concentrations were similar in HF patients with and without renal dysfunction , indicating renal function might not significantly affect the expression of sST2. Meanwhile,365医学网 转载请注明
Gopal et al. found galectin-3 elevated in patients with renal dysfunction but no HF . In this line, we found galectin-3 concentrations were significantly higher in patients with eGFR <60 ml/min/1.73 m2 regardless of NYHA functional class whereas sST2 was only significantly higher in those with NYHA functional class IV in the face of renal insufficiency. In the balance of results, the higher sST2 values in class IV patients with worse renal function presumably reflects more severe HF in this class of patients, rather than effects of impaired clearance and/or renal fibrosis.
We found rather substantial links between galectin-3 and renal insufficiency, along with a statistical interaction term between galectin-3 and renal function. We suggest that values of galectin-3 may reflect a burden of renal fibrosis and dysfunction. In this vein, O’Seaghdha et al. found that higher levels of galectin-3 were associated with rapid decline in eGFR and a higher risk of incident chronic kidney disease (CKD) in 2450 Framingham offspring participants during 10.1 years follow-up . In this regard, with worsening renal function, it is reasonable to assert that365医学网 转载请注明
higher galectin-3 values might be seen, driven possibly by renal fibrosis, and in this manner, the prognostic value of galectin-3 would be lost in HF patients that have concomitant renal dysfunction, much as we have observed. Indeed, De Boer et al. showed that the predictive power of galectin-3 declined in HF patients when corrected by eGFR , as did Lopez-Andres et al. .365医学网 转载请注明
In contrast, no reports link vulnerability of sST2 in patients with moderate or more renal failure; BayesGenis et al. recently reported that sST2 was superior to galectin-3 in risk prediction in patients with chronic HF . The low median of eGFR (43.2 ml/min/1.73 m2 [29.7–59.8]) in their study might contribute to this result.
Our study has several limitations. First, 15 patients were lost during follow-up period, but all of them completed the first 6 months follow-up, and provided useful information for this analysis. Second, we only predefined death as the endpoint of this study. Except for all-cause death, current risk prediction of this two bio markers have been focus on cause-specific death [8,23,24]
and hospitalization due to worsening HF [23,25]. Third,because the number of patients with severe renal function in our study was small (only 80 patients with eGFR <30 ml/min/1.73 m2), we did not evaluate the prognostic value of sST2 and galectin-3 according to complete NKF KDOQI categories.
Moderate or worse renal insufficiency is extraordinarily common among patients with the diagnosis of HF and changes in renal function are a hallmark of worsening HF, it calls into question whether galectin-3 is a reliable biomarker for use in this fragile, medically365医学网 转载请注明
complex patient population. Based on our data, use of galectin-3 as a risk predictor for patients with both HF and renal dysfunction should only be undertaken with this consideration in mind. In contrast, sST2 appears considerably less affected by renal function.
sST2 is a strong prognostic biochemical markers in patients with heart failure. Based on the result of our study, the use of sST2 as a risk predictor was not affected by impaired renal function in patients hospitalized with HF. With the development of point-ofcare testing (POCT), sST2 POCT has the potential to help physicians prompt identification of risk more quickly and accurately and optimize management of HF. We expect that over the next 5–10 years, sST2 POCT will probably extend from outpatient to emergency department and primary care in risk prediction. In addition, the role of sST2 in cardiac fibrosis was only reported in animal model and was not validated in human. Based on the result of our another study, sST2 was negatively correlated with relative
wall thickness, which indicated that sST2 might be associated with the process of cardiac apoptosis. The association between sST2 and cardiac fibrosis need to be addressed clearly in future study. Furthermore, the utility of galectin-3 as a biomarker is mostly developed for HF. We validated that it is a valuable biomarker of HF mortality in patients with preserved renal function. In addition, it was strongly correlated with renal function across all NYHA classes examined suggesting that galectin-3 is a valuable biomarker of renal function in HF patients. Given that, more studies are needed to address whether galectin-3 was advantageous for detection of kidney injury in patients with HF. Better identifying the biological characteristic of cardiac biomarker might provide the therapeutic potential and pharmaceutical interest for the treatment of HF.
Papers of special note have been highlighted as: • of interest;
•• of considerable interest
1 Hillege HL, Nitsch D, Pfeffer MA et al. Renal function as axpredictor of outcome in a broad spectrum of patients with heart failure. Circulation 113(5), 671–678 (2006).
2 McAlister FA, Ezekowitz J, Tonelli M, Armstrong PW. Renal insufficiency and heart failure: prognostic and therapeutic implications from a prospective cohort study. Circulation 109(8), 1004–1009 (2004).
3 Shah RV, Januzzi JL Jr. Soluble ST2 and Galectin-3 in Heart Failure. Clin. Lab. Med. 34(1), 87–97 (2014).
4 Bao YS, Na SP, Zhang P et al. Characterization of future science group www.futuremedicine.com 441 Prognostic value of sST2 & galectin-3 relative to renal function in heart failure Research Article interleukin-33 and soluble ST2 in serum and their association with disease severity in patients with chronic kidney disease. J. Clin. Immunol. 32(3), 587–594 (2012).
5 Gopal DM, Kommineni M, Ayalon N et al. Relationship of plasma galectin-3 to renal function in patients with heart failure: effects of clinical status, pathophysiology of heart failure, and presence or absence of heart failure. J. Am. Heart Assoc. 1(5), e000760 (2012).
•• First study to investigate whether the relationship between renal function and galectin-3 is influenced by the presence or absence of clinical heart failure.
6 Bayes-Genis A, Zamora E, de Antonio M et al. Soluble ST2 serum concentration and renal function in heart failure. J. Card. Fail. 19(11), 768–775 (2013).
• First study to evaluate the association between prognostic value of sST2 and renal function.
7 Zhang R, Zhang Y, Zhang J et al. the prognostic value of plasma soluble ST2 in hospitalized Chinese patients with heart failure. PLoS ONE 9(10), e110976
8 Zhang Y, Zhang R, An T et al. The utility of galectin-3 for predicting cause-specific death in hospitalized patients with heart failure. J. Card. Fail. 21(1), 51–59 (2015).
9 Levey AS, Stevens LA, Schmid CH et al. A new equation to estimate glomerular filtration rate. Ann. Intern. Med. 150(9), 604–612 (2009).
• A new equation more accurately to estimate glomerular filtration rate than Modification of Diet in Renal Disease (MDRD) equation.
10 Levey AS, Coresh J, Balk E et al. National Kidney Foundation practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Ann. Intern. Med. 139(2), 137–147 (2003).
11 Sanada S, Hakuno D, Higgins LJ et al. IL-33 and ST2 comprise a critical biomechanically induced and cardioprotective signaling system. J. Clin. Invest. 117(6), 1538–1549 (2007).
•• First study to explore the role of ST2 in the myocardium.12 de Boer RA, Yu L, van Veldhuisen DJ. Galectin-3 in cardiac remodeling and heart failure. Curr. Heart Fail. Rep. 7(1), 1–8(2010).
13 Kumar S, Tzimas MN, Griswold DE, Young PR. Expression of ST2, an interleukin-1 receptor homologue, is induced by proinflammatory stimuli. Biochem. Biophys. Res. Commun. 235(3), 474–478 (1997).
14 Barondes SH, Cooper DN, Gitt MA, Leffler H. Galectins. Structure and function of a large family of animal lectins. J. Biol. Chem. 269(33), 20807–20810 (1994).
15 Henderson NC, Mackinnon AC, Farnworth SL et al. Galectin-3 regulates myofibroblast activation and hepatic fibrosis. Proc. Natl Acad. Sci. USA 103(13), 5060–5065(2006).
16 Henderson NC, Mackinnon AC, Farnworth SL et al. Galectin-3 expression and secretion links macrophages to the promotion of renal fibrosis. Am. J. Pathol. 172(2), 288–298(2008).
• Demonstrates for the first time that galectin-3 expression and secretion by macrophages is a major mechanism to the promotion of renal fibrosis.
17 Wang L, Friess H, Zhu Z et al. Galectin-1 and galectin-3 in chronic pancreatitis. Lab. Invest. 80(8), 1233–1241 (2000).
18 Dieplinger B, Januzzi JL Jr, Steinmair M et al. Analytical and clinical evaluation of a novel high-sensitivity assay for measurement of soluble ST2 in human plasma—the Presage ST2 assay. Clin. Chim. Acta. 409(1–2), 33–40 (2009).
• Evaluates the determination of sST2 with a novel Presage ST2 assay.
19 O’Seaghdha CM, Hwang SJ, Ho JE et al. Elevated galectin-3 precedes the development of CKD. J. Am. Soc. Nephrol.24(9), 1470–1477 (2013).
• Found galectin-3 are associated with increased risks of chronic kidney disease incident in the community.
20 de Boer RA, Lok DJ, Jaarsma T et al. Predictive value of plasma galectin-3 levels in heart failure with reduced and preserved ejection fraction. Ann. Med. 43(1), 60–68(2011).
21 Lopez-Andres N, Rossignol P, Iraqi W et al. Association of galectin-3 and fibrosis markers with long-term cardiovascular outcomes in patients with heart failure, left ventricular dysfunction, and dyssynchrony: insights from the CARE-HF (Cardiac Resynchronization in Heart Failure) trial. Eur. J.
Heart Fail. 14(1), 74–81 (2012).
22 Bayes-Genis A, de Antonio M, Vila J et al. Head-to-head comparison of 2 myocardial fibrosis biomarkers for long-term heart failure risk stratification: ST2 versus galectin-3. J. Am. Coll. Cardiol. 63(2), 158–166 (2014).
• First study to compare the predictive power of ST2 and galectin-3 in patients with heart failure.
23 Broch K, Ueland T, Nymo SH et al. Soluble ST2 is associated with adverse outcome in patients with heart failure of ischaemic aetiology. Eur. J. Heart Fail. 14(3), 268–277(2012).
24 Ahmad T, Fiuzat M, Neely B et al. Biomarkers of myocardial stress and fibrosis as predictors of mode of death in patients with chronic heart failure. JACC Heart Fail. 2(3), 260–268(2014).
25 Meijers WC, Januzzi JL, deFilippi C et al. Elevated plasma galectin-3 is associated with near-term rehospitalization in heart failure: a pooled analysis of 3 clinical trials. Am.Heart J. 167(6), 853–860, e854 (2014).