Serial sampling of st2 predicts 90-day mortality following destabilized heart failure

Journal of Cardiac Failure Vol. 14 No. 9 2008 Serial Sampling of ST2 Predicts 90-Day Mortality Following Destabilized Heart Failure SASKIA BOISOT, MD,JENNIFER BEEDE, SUSAN ISAKSON, BS,ALBERT CHIU, BS,PAUL CLOPTON, JAMES JANUZZI, MD,ALAN S. MAISEL, AND ROBERT L. FITZGERALD, PhD, San Diego, California; Boston, Massachusetts Background: To prospectively determine the prognostic utility of serial sampling of the interleukin-1receptor family member, ST2, for predicting 90-day mortality in patients with heart failure (HF) admittedto a Veteran Affairs Medical Center.
Methods and Results: A total 150 patients hospitalized with acutely destabilized HF were followed atthe Veteran Affairs Healthcare System in San Diego, CA. Multiple cardiac-related parameters were mea-sured including ST2, B-type natriuretic peptide (BNP), NT-proBNP, and blood urea nitrogen (BUN).
Plasma samples were collected at 6 time points between admission and discharge. Biomarker concentra-tions were correlated to survival at 90 days. Uni- and multivariate analyses were used to identify prognos-tic variables. From admission to discharge, percent change in ST2 was strongly predictive of 90-daymortality: those patients whose ST2 values decreased by 15.5% or more during the study period hada 7% chance of death, whereas patients whose ST2 levels failed to decrease by 15.5% in this time intervalhad a 33% chance of dying.
Conclusions: Percent change in ST2 concentrations during acute HF treatment is predictive of 90-daymortality and was independent of BNP or NT-proBNP levels. ST2 may provide clinicians with an addi-tional tool for guiding treatment in patients with acute destabilized HF. (J Cardiac Fail 2008;14:732e738)Key Words: ST2, B-type natriuretic peptide, Pro-B-type natriuretic peptide, heart failure.
Genomic and proteomic discovery programs have identi- receptor,'' because the natural ligand was unknown.
fied many clinically useful gene products, some of which Recently, Schmitz et al demonstrated that interleukin-33 may be measured for prognostic purposes.One gene char- is the natural ligand for the ST2 receptor acterized in this fashion is ST2, which was originally iden- A truncated version of the ST2 receptor is produced by tified in an in vitro model by examining mRNA transcripts alternative promoter splicing and 30 and is upregulated when neonatal rat cardiac myocytes were sub- subsequently secreted into circulation. This soluble form jected to mechanical stretch.Subsequently, several investi- is detected in the serum of patients early after acute myo- gators have shown that the protein product of the ST2 gene cardial infarction,and inversely correlates with ejection is increased in a number of pathologic conditions, including fraIt has been postulated that the volume overload asthma, myocardial infarction, and heart failure (HF).
engendered by poor cardiac pump function causes stretch- The ST2 gene codes for a transmembrane receptor, origi- ing of the myocardial fibers,which then serves as a signal nally identified as an interleukin 1erelated ‘‘orphan for nearby cells and the myocytes themselves to, amongother things, upregulate ST2 producSanada et al dem-onstrated that the soluble form of ST2 acts as a decoy From the 1University of California at San Diego, Veteran Affairs San Diego Healthcare System, San Diego, CA; 2Department of Pathology, Uni- receptor to block the interleukin 33emediated effects on versity of California at San Diego and 3Department of Cardiology, Massa- the transmembrane ST2 receptor.
chusetts General Hospital, Boston.
Shimpo et al demonstrated that baseline ST2 concentra- Manuscript received January 31, 2008; revised manuscript received May 21, 2008; revised manuscript accepted June 16, 2008.
tions predict 30-day mortality in patients after acute myo- Reprint requests: Saskia Boisot, MD, VASDHS Chemistry, 113, 3350 La cardial infarction.In the setting of chronic outpatient Jolla Village Dr, San Diego CA 92161 HF, Weinberg et al used multivariate analysis to show Drs. Januzzi and Maisel have a financial relationship with the study sponsor critical Diagnostics, as they are on the Scientific Advisory board that change in ST2 levels over a 2-week period predicted and are financially compensated.
outcome that was independent of B-type natriuretic peptide 1071-9164/$ - see front matter (BNP)In each of these studies, the subjects evaluated did Ó 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.cardfail.2008.06.415 not have acutely destabilized HF. Recent data suggest a role Serial Sampling of ST2 Predicts 90-Day Mortality after Destabilized Heart Failure  Boisot et al 733 for ST2 measurement in the prognosis of acutely destabi- abnormalities), LV systolic dysfunction defined by an ejection lized HF,but no studies have been conducted looking at fraction !50%, or diastolic dysfunction. Diastolic dysfunction serial ST2 measurements over the treatment course of acute was determined using Doppler measurements of mitral inflow HF. Accordingly, we collected serial blood samples for mea- and Doppler tissue imaging of the mitral annulus, as previously surement of ST2 levels in the setting of hospitalized HF described by Redfield et al.Subjects whose measurementswere borderline or suggestive of, but not definitive for, diastolic patients to determine if changes in concentrations of this dysfunction were classified as indeterminate.
biomarker would be useful in predicting 90-day outcomes.
Statistical Analyses ST2 values across freeze-thaw cycles were reported as geomet- ric means and evaluated with paired t-tests and Pearson's correla- This prospective study was designed to determine the prognos- tion. Continuous variables were represented with medians and tic utility of serial measurements of ST2 in patients admitted to the interquartile ranges, and graphed over time as geometric means San Diego Veterans Affairs Medical Center with a clinical diagno- and standard errors. Frequencies of categorical variables were ex- sis of acute decompensated heart failure between July 2002 and pressed as percentages and compared with Fisher exact tests, with December 2005 (the VA Effect of Therapy StudyPatients age (O65 years) and ejection fraction (!35%) treated as nominal younger than age 18 or unable to provide informed consent variables. Receiver operating characteristic (ROC) curve analysis were excluded from participation. After obtaining written in- for predicting 90-day mortality was performed for the remaining formed consent, data were obtained from verbal conversations continuous variables: baseline and percent change (first to last with the patients and a review of their electronic medical records.
sample) in ST2, BNP, NT-proBNP, and BUN. Potential prognostic Samples were obtained from venipuncture (potassium EDTA), cut points for the stronger of baseline or percent change in each of specimens were centrifuged, and plasma was immediately frozen these 4 parameters were identified as the point on the ROC curve (70C). Blood was collected at admission and discharge, as well nearest the top left corner, and their respective sensitivities and as up to 4 consecutive days between admission and discharge.
specificities were calculated. Comparisons between ROC curves After discharge from the hospital, patients were reevaluated were made using methods for correlated curves.Fisher exact through review of their electronic medical records or via telephone tests were used to identify significant univariate predictors of mor- conversations to evaluate the primary end point of 90-day mortal- tality. These predictors were then entered in a multivariate logistic ity. The diagnosis of HF was based on a number of parameters in- regression analysis. Patient characteristics were subjected to re- cluding, but not limited to, BNP, ejection fraction, and pulmonary verse stepwise removal, whereas biomarkers were forced into edema. Patients were subsequently stratified according to the New the final multivariate model. Group differences in log-transformed York Heart Association (NYHA) guidelines, with most patients ST2 concentrations at various time points were tested using t-tests (92%) falling into Class III or IV.
for independent samples.
Of the 200 patients enrolled in the study, 50 patients were ex- cluded from analysis because ST2 was measured at only 1 time point, leaving a final sample of 150 patients. The 50 patientswho were excluded were discharged before serial samples could be obtained and thus were not included in the data analysisbecause this study was designed to evaluate serial sampling of Variation in ST2 levels between fresh samples and each ST2 in a hospital setting.
freeze-thaw cycle was minimal. Geometric means after 0, Concentrations of ST2 were determined using an enzyme- 1, and 2 freeze thaw cycles were 0.57, 0.58, and 0.54 ng/ linked immunosorbent assay (Medical & Biological Laboratories mL, respectively. The difference between 0 and 1 cycle Co, Woburn, MA). This assay uses 2 monoclonal antibodies rec- or between 0 and 2 cycles was not statistically significant ognizing epitopes on human ST2 for both capture and detection, (P 5 .45 and P 5 .22, respectively). The correlations be- and has an interassay coefficient of variation of 14%. Samples tween 0 and 1 cycle (r 5 0.998, P ! .001) and between were stored at 70C before analysis. Because they were sub- 0 and 2 cycles (r 5 0.991, P ! .001) were high.
jected to a single freeze-thaw cycle, it was necessary to validatethat ST2 was stable through at least 1 freeze thaw cycle. ST2 sta- Patient Characteristics, ST2, BNP, and BUN Values bility was established by comparing the ST2 levels in fresh sam-ples with results from samples subjected to 2 freeze-thaw Detailed patient characteristics are shown in . By cycles. BNP and NT-proBNP concentrations were measured using the end of the 90-day study period, 126 patients were alive the Biosite-Triage assay (Biosite-Triage; San Diego, CA) and the and 24 had died, yielding a mortality rate of 16%. Of the ad- Roche ProBNP assay (Roche Diagnostics; Indianapolis, IN), re- ditional 50 subjects not included in the data analysis, 48 were spectively. Blood urea nitrogen (BUN) levels from admission alive and 2 died, yielding a mortality rate of 4% (P 5 .032).
and discharge samples were measured using the Beckman LX20 The overall 200 patient cohort mortality rate was 13%.
autoanalyzer (Beckman Coulter, La Brea, CA).
Almost all the study patients were male, with 61% being Echocardiograms were performed by certified sonographers and older than age 65. Seventy-three percent of the patients pre- interpreted by a cardiologist who was blinded to biomarker levels.
sented with worsening of previously diagnosed HF, and the Left ventricular (LV) systolic and diastolic volumes and ejectionfraction were derived from biplane apical views using the modi- spectrum of disease spanned the entire range of severity, fied Simpson's rule Patients classified with LV dys- with most patients falling in the NYHA Classes III-IV.
function had any or all of the following: LV segmental wall Using analysis of variance, the underlying etiology of con- motion abnormalities (with or without systolic or diastolic gestive HF had no statistically significant impact on either Journal of Cardiac Failure Vol. 14 No. 9 November 2008 Table 1. Patient Characteristics the strongest and most statistically significant capacityfor predicting 90-day mortality, namely changes in ST2, NT-proBNP, and BNP, as well as baseline BUN and NT-proBNP. Baseline BNP and NT-proBNP levels were (Interquartile Range) statistically significant, with AUCs of 0.629 (95% CI0.513e0.746; P 5 .05) and 0.738 (95% CI 0.635e0.841; P # .001), respectively, whereas baseline ST2 and change 0.177 (0.086e0.344) in BUN were not, with respective AUCs of 0.583 (95% 0.103 (0.043e0.219) CI 0.447e0.718; P 5 .201) and 0.590 (95% CI 0.467e0.714; P 5 .162). The percent change in ST2 had 5,878 (2297e11,918) an AUC of 0.783 (95% CI 0.690e0.877; P # .001), which 3,580 (1379e10,102) was identical to baseline BUN concentrations (AUC 0.783 [95% CI 0.690e0.877; P # .001]), and almost identical to percent change in NT-proBNP (AUC 0.781 [95% CI: 0.665e0.897; P # .001]), whereas the percent change in BNP had an AUC of 0.671 (95% CI 0.555e0.787; P 5 .008). ROC curves for the 4 strongest predictors are shown Number of patients in . Of the 4 strongest predictors, the only notable at each time point statistical difference was that NT-proBNP had a signifi- cantly higher AUC as compared with BNP (P 5 .011).
The optimal cut point for percent change in ST2 corre- sponded to a 15.5% decrease in ST2 levels (sensitivity 70%, specificity 73%). Those patients whose ST2 values decreased by 15.5% or more during the study period had a 7% chance of death, whereas those whose ST2 levelsfailed to decrease by at least 15.5% in this interval had a 33% chance of dying. Patients who died during the 90- at admission:Class 2 day follow-up period tended to have increasing concentra- tions of ST2 while they were in the hospital, whereas those who survived during follow-up tended to have decreasing ST2 concentrations Thus these groups did not dif- fer in ST2 concentrations at baseline (P 5 .22), but ST2 concentrations were elevated in subsequent observations BNP, B-type natriuretic peptide; BUN, blood urea nitrogen; CHF, con- gestive heart failure; CAD, coronary artery disease; MI, myocardial infarc-tion; CRI, chronic renal insufficiency; DM, diabetes mellitus; HTN,hypertension; EF, ejection fraction; NYHA, New York Heart Association.
*Interquartile ranges (25%e75%).
baseline or percent change values for ST2, nor any of the other measured biomarkers (data not shown).
The median length of stay was 4 days (interquartile range e7), with a mean of 5.77 days (SD 7.44). ST2 values could not be obtained for all patients at all time points,with the greatest deficits being at Days 4 and 5. The major-ity of patients had their first time point drawn at admission, AUC = 0.78 CI = 0.69 to 0.88 whereas in 16 patients the Day 2 ST2 level served as their AUC = 0.67 CI = 0.56 to 0.79 AUC = 0.78 CI = 0.69 to 0.88 first measured time point, because they did not have ST2 AUC = 0.78 CI = 0.67 to 0.90 levels measured on admission.
ST2, BNP, NT-proBNP, BUN, and Prognosis 1 - Specificity
ROC analysis was used to eliminate those continuous Fig. 1. Receiver operator characteristic curves for predicting 90- variables whose area under the curve (AUC) failed to reach day mortality. Blood urea nitrogen (BUN) is plotted from baseline statistical significance. ROC analysis was also used to iden- values, whereas ST2, B-type natriuretic peptide (BNP), and NT- tify prognostic cut points for those parameters that yielded proBNP are plotted as the percent change from first to last sample.
Serial Sampling of ST2 Predicts 90-Day Mortality after Destabilized Heart Failure  Boisot et al 735 Table 2. Univariate Analysis of Predictors of Death Variable Present Variable Absent Predictor Variable Decrease ST2 ! 15.5% Decrease BNP !10.0% ST2 (ng/ml)
Decrease NT-proBNP Fig. 2. Plot of ST2 concentrations (geometric means and standard errors) at serial time points in patients who survived as compared with those who died within 90 days. The groups did not differ in ST2 concentrations at baseline (P 5 .22), but were elevated in Atrial fibrillation subsequent observations for those who died within 90 days by Day 2 (P 5 .03) and thereafter (P ! .001 for each).
for those who died within 90 days by Day 2 (P 5 .03) and thereafter (P ! .001 for each). shows data on all patients including those without 5 observations. This figureis similar to what is observed if only patients with complete BNP, B-type natriuretic peptide; BUN, blood urea nitrogen; CHF, con- data through 3, 4, or 5 days are evaluated which all show gestive heart failure; CAD, coronary artery disease; MI, myocardial infarc-tion; HTN, hypertension; COPD, chronic obstruction pulmonary disease.
significant group by time interactions (P ! .001 for all,data not shown).
The optimal cut point for percent change in BNP was a de- a combination of these 3 variables had an additive effect crease of 10% (sensitivity 63%, specificity 67%). If the BNP on predicting mortality: of patients who had a combination concentration decreased by 10% or more, there was a 10% of BUN #40 mg/dL, an ST2 decrease of $15.5%, and an chance of death, whereas if it did not there was a 27% NT-proBNP decrease of #3%, 1.6% of them died, whereas chance of dying. The optimal cut point for percent change of patients who had a BUN O40 and did not attain either in NT-proBNP corresponded to a decrease of 3% (sensitivity a decrease in ST2 of at least 15.5% or a decrease in NT- 71%, specificity 23%). If the NT-proBNP concentration de- proBNP of at least 3%, 72.7% of them died. Using logistic creased by 3% or more, there was a 7% chance of death, regression, none of the interactions among the categories of whereas failing to do so incurred a 38% chance of dying.
ST2, NT-proBNP, and BUN was statistically significant (all Finally, 2 possible cut points can be noted for BUN in P O .7), indicating that these effects are additive and not . One represents a value of 26 mg/dL, the other a value of 40 mg/dL. Because a BUN of 26 mg/dL (sensitivity Nonparametric analysis of the relationship between these 88%, specificity 60%) is close to the high end of the normal 4 variables revealed a weak but significant correlation be- range, the higher cut point corresponding to a BUN of 40 tween % change ST2 and % change BNP (r 5 0.409, P mg/dL (sensitivity 63%, specificity 80%) was selected for ! .001), % change NTproBNP (r 5 0.491, P ! .001), further analysis. Patients who did not achieve a reduction and baseline BUN, although the strongest association was in their BUN levels #40 mg/dL had a 38% chance of dy- between % change BNP and % change NTproBNP (r 5 ing, whereas those who did had a 92% chance of survival.
0.724, P ! .001), and between baseline BNP and baseline Univariate analyses were performed on potential prog- NTproBNP (r 5 0.780, P ! .001).
nostic variables. As shown in percent change inST2 and BNP were both significant (P ! .001 and P 5 .01, respectively). Other significant predictors includedage O65 years, baseline BUN O40 mg/dL, decrease in This study was designed to evaluate a novel cardiac bio- NT-proBNP of !3%, wheezing, murmurs, history of coro- marker, ST2, as a potential predictor of mortality at 90 days nary artery disease and of myocardial infarction. In the in patients with acutely decompensated HF, requiring hos- multivariate analysis, the only parameters to reach statisti- pitalization. In 2002, Weinberg et al demonstrated that cal significance (were decrease in ST2 ! 15.5% soluble ST2 mRNA transcription could be induced in neo- (P 5 .007), BUN O40 mg/dL (P 5 .042), and decrease natal rat cardiac myocytes subjected to mechanical strain, in NT-proBNP of !3% (P 5 .030). As shown in or given Il-1b or phorbol ester.They also found increased

Journal of Cardiac Failure Vol. 14 No. 9 November 2008 Table 3. Multivariate Analysis of Predictors of Death a decrease in ST2 expression was even more predictive ofsurvival. Although the change in ST2 appeared to show a stronger relationship with outcome than either baseline or change in natriuretic peptides, we wanted to compareit with another known prognostic indicator of HF, renal Decrease ST2 ! 15.5% Decrease BNP !10% function in the form of BUN. Because of inadequate car- Decrease NT-proBNP !3% diac output, patients with HF often develop prerenal azote- mia, and may exhibit BUN to creatinine ratios in excess of CI, confidence interval; OR, odds ratio; BNP, B-type natriuretic peptide; 20:1.As such, elevations in BUN could plausibly be used BUN, blood urea nitrogen.
as a corollary of worsening HF, and might be predictive ofsubsequent mortality, whereas creatinine levels would not levels of soluble ST2 in the serum of mice after experimen- follow such a trend. Furthermore, because BUN was one tal myocardial infarction, and in the serum of human of only two other variables in our study aside from change patients 1 day after myocardial infarcBecause inter- in ST2 to reach statistical significance in both univariate leukin-1 is one of the cytokines activated during the stress and multivariate analyses, this made it an attractive choice response associated with acute myocardial infarc for such comparisons.Although baseline BUN predicted these findings strengthen the notion that ST2 serves in mod- 90-day mortality, it did not provide a means to monitor ulating the immune response and mediates subsequent therapy as changes in BUN were not predictive. Clearly repair mechanisms. Although the primary source for ST2 there is a need for biomarkers whose change during therapy production has not yet been uncovered, soluble ST2 protein predicts outcome.
has been shown to be secreted by cardiac myocytes when ROC analysis demonstrated that the % change in ST2 the cells are subjected to biomechanical ovIt has was equivalent to % change in NT-proBNP and that both also been postulated that the ST2 protein originates from of these biomarkers were better at predicting 90-day mor- multiple sources in response to the local inflammatory mi- tality than % change in BNP. In multivariate analysis, there croenvironment generated by cell injury.
was additive predictive power when % change in ST2, % Although the significance of these previous studies in change in NT-proBNP, and baseline BUN were analyzed relation to the pathophysiology of heart failure has yet to together. These findings agree with Weinberg et al's conclu- be established, it is plausible that the impaired regulation sions from the Prospective Randomized Amlodipine Sur- of these immune events might represent a precursor to the vival Evaluation (PRAISE)-2 trial, which showed that development of cardiac fibrosis, and thus subsequent pro- change in ST2 levels, but not the baseline ST2, was predic- gression of left ventricular dysfunction. In support of this, tive of 30-day mortality in patients with chronic NYHA pro-inflammatory cytokines have previously been shown to Class III-IV HFOur results confirmed that serial BNP participate in the development and progression of HF measurements did not yield any independently prognostic Our results confirmed the hypothesis that increased ST2 information in the context of ST2 and other independent expression was indeed predictive of mortality, whereas variables, although serial NT-proBNP measurements did.
Finally, the cut points we obtained for serial % changes in ST2 and NT-proBNP should be interpreted in the contextof both analytic and biologic variability of the respectiveassays.In the case of the ST2 assay, the interassay coef- ficient of variation in our study was 14%, so a decrease in ST2 of at least 15.5% could be regarded as clinically signif-icant, even in light of such high interassay variability. With improved analytical performance, the predictive value ofthe ST2 assay would likely increase. The clinical utilityof serial measurements of ST-2 would also depend on the biologic variability of this marker, which is currently un- known. For NT-proBNP, the interassay coefficient of varia- tion was approximately 3%, which is the same as the cut point we obtained from ROC analysis for a changing value predicting 90-day mortality. From our clinical experience, we know that the biologic variability of NT-proBNP mea- NT-proBNP De No
surements within an individual is much greater than 3%, crease ≥ 3%
and that reference change values, the amount of change ST2 Decrease ≥ 15.5%
needed to indicate a pathologic event, may be as high as23%.As such, it is difficult to ascertain whether a de- Fig. 3. Combined effect of blood urea nitrogen (BUN) $40, de-crease in NT-proBNP #3%, and decrease in ST2 # 15 on predict- crease in NT-proBNP of 3% is truly clinically significant ing mortality.
or if this relatively small change merely reflects adequate Serial Sampling of ST2 Predicts 90-Day Mortality after Destabilized Heart Failure  Boisot et al 737 therapy and a definitive downturn in NT-proBNP values; with destabilized HF, and may assist in clinical decision- indeed, prior data suggest that a more plausible in-hospital NT-proBNP change might be 30% or Current guidelines for the treatment of HF are largely based on subjective clinical criteria, such as cardiac stresstesting, echocardiography, and ultimately NYHA class.
The authors would like to thank James Snider from Crit- The need for more objective parameters is underscored by ical Diagnostics for providing the data on the freeze/thaw the vast body of research into cardiac biomarkers, and their stability of ST2.
potential use in this context.The use of both BNP andNT-proBNP concentrations in guiding the treatment ofHF in the outpatient setting has been investigated, and 2studies show significantly improved outcomes when com- pared with the use of standard treatment protHowever, no prospective studies have been conducted look- 1. Lee RT. Functional genomics and cardiovascular drug discovery. Cir- culation 2001;104:1441 ing at the use of biomarker-guided therapy in the inpatient 2. Weinberg EO, Shimpo M, De Keulenaer GW, MacGillivray C, setting. Our results suggest that serial ST2 measurements in Tominaga S, Solomon SD, et al. Expression and regulation of ST2, this context may help to stratify patients according to their an interleukin-1 receptor family member, in cardiomyocytes and myo- predicted mortality and further assist in guiding treatment.
cardial infarction. Circulation 2002;106:2961e6.
In addition, understanding the pathophysiology of eleva- 3. Oshikawa K, Yanagisawa K, Ohno S, Tominaga SI, Yukihiko S.
Expression of ST2 in helper T lymphocytes of malignant pleural effu- tions of ST2 in the context of acute decompensated HF sions. Am J Respir Crit Care Med 2002;165:1005e9.
may uncover novel approaches to treatment.
4. Weinberg EO, Shimpo M, Hurwitz S, Tominaga SI, Rouleau JL, Lee RT. Identification of serum soluble ST2 receptor as a novel heart failure biomarker. Circulation 2003;107:721e6.
The present study is hindered by a few unavoidable lim- McClanahan TK, et al. Il-33, and interleukin-1-like cytokine that sig- itations. The number of patients in the study population was nals via the Il-1 receptor-related protein ST2 and induces T helper type relatively small and consisted of subjects from a single VA 2-associated cytokines. Immunity 2005;23:479e90.
6. Walzl G, Matthews S, Kendall S, Gutierrez-Ramos JC, Coyle AJ, Medical Center with a predominantly male patient base.
Openshaw PJ, et al. Inhibition of T1/ST2 during respiratory syncytial This population also had a high incidence of cardiac disease virus infection prevents T helper cell type 2 (Th2)- but not Th1-driven with other significant comorbidities. Because of this, the re- immunopathology. J Exp Med 2001;193:785e92.
sults presented may not be applicable to other community 7. Iwahana H, Yanagisawa K, Ito-Kosaka A, Kuroiwa K, Tago K, settings. Because therapy was not guided by serial measure- Komatsu N, et al. Different promoter usage and multiple transcriptioninitiation sites of the interleukin-1 receptor-related human ST2 gene in ments of ST2 or NT-proBNP, we do not know if these bio- UT-7 and TM12 cells. Eur J Biochem 1999;264:397e406.
markers are useful for this purpose.
8. Bergers G, Reikerstorfer A, Braselman S, Graninger P, Busslinger M.
Assessment for 90-day outcome was largely performed Alternative promoter usage of the fos-responsive gene Fit-1 generates on the electronic medical record, making it possible that RNA isoforms coding for either secreted or membrane-bound proteins certain events might not have been included in our analysis.
related to the Il-1 receptor. EMBO J 1994;13:1176e88.
9. Shimpo M, Morrow DA, Weinberg EO, Sabatine MS, Murphy SA, However, the number of such missing events is likely to be Antman EM, et al. Serum levels of the interleukin-1 receptor family negligible, because all study participants had primary care member ST2 predict mortality and clinical outcome in acute myocar- physicians within the VA Medical Center system, and dial infarction. Circulation 2004;109:2186e90.
most study subjects obtained their care solely through a sin- 10. Feldman AM, Weinberg EO, Ray PE, Lorell BH. Selective changes in gle center. It is important to note that all of the events cardiac gene expression during compensated hypertrophy and the tran-sition to cardiac decompensation in rats with chronic aortic banding.
recorded from telephone follow-up were also contained in Circ Res 1993;73:184e92.
the electronic medical record.
11. Sanada S, Hakuno D, Higgins LJ, Schreiter ER, McKenzie ANJ, Lee RT. IL-33 and ST2 comprise a critical biomechanically inducedand cardioprotective signaling system. J Clin Invest 2007;117: 12. Januzzi JL, Peacock WF, Maisel AS, Chae CU, Jesse RL, Baggish AL, In conclusion, serial measurements of a novel heart failure et al. Measurement of the interleukin family member ST2 in patients biomarker, ST2, added independent prognostic information with acute dyspnea: results from the ProBNP Investigation of Dyspneain the Emergency Department (PRIDE) study. J Am Coll Cardiol among a population of patients undergoing therapy for desta- bilized HF. Whether used alone or together with NT-proBNP 13. Waldo SW, Beede J, Isakson S, Villard-Saussine S, Fareh J, Clopton P, and/or BUN, ST2 was strikingly prognostic, with those sub- et al. Pro-B-type natriuretic peptide levels in acute decompensated jects demonstrating a rise in ST2 during HF therapy suffering heart failure. J Am Coll Cardiol 2008;51:1874e82.
the worst short-term outcomes, independent of natriuretic 14. Schiller NB, Acquatella H, Ports TA, Drew D, Goerke J, Ringertz H, et al. Left ventricular volume from paired biplane two-dimensional peptide concentrations. The role of ST2, an emerging bio- marker bridging the inflammatory and neurohormonal sys- 15. Redfield MM, Jacobsen SJ, Burnett JC, Mahoney DW, Bailey KR, tems, appears promising in predicting outcomes in patients Rodeheffer RJ. Burden of systolic and diastolic ventricular dysfunction Journal of Cardiac Failure Vol. 14 No. 9 November 2008 in the community: appreciating the scope of the heart failure epidemic.
principles of internal medicine. 13th Edition. New York: McGraw- JAMA 289;194e202.
Hill; 1994. p. 1002.
16. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas 23. Bruins S, Fokkema MR, Romer JWP, Dejongste MJ, van der Dijs FP, under two or more correlated receiver operating characteristic curves: van den Ouweland JM, et al. High intraindividual variation of B-type a nonparametric approach. Biometrics 1988;44:837e45.
natriuretic peptide and amino-terminal proBNP in patients with stable 17. Mann DL. Stress activated cytokines and the heart. Cytokine Growth chronic heart failure. Clin Chem 2004;50(11):2052e8.
Factor Rev 1996;7:341e54.
24. Schou M, Gustafsson F, Nielsen PH, Madsen LH, Kjaer A, 18. St John Sutton MG, Sharpe N. Left ventricular remodeling after myo- Hildebrandt PR. Unexplained week-to-week variation in BNP and cardial infarction: pathophysiology and therapy. Circulation 2000;101: NT-proBNP is low in chronic heart failure patients during steady state.
Eur J Heart Fail 2007;68e74.
19. Yamaoka M, Yamaguchi S, Okuyama M, Tomoike H. Anti-inflammatory 25. Bettencourt P, Azevedo A, Pimenta J, Frioes F, Ferreira S, Ferreira A.
cytokine profile in human heart failure: behaviour of interleukin-10 in N-terminal-pro-brain natriuretic peptide predicts outcome after hospi- association with tumor necrosis factor-a. Jpn Circ J 1999;63:951e4.
tal discharge in heart failure patients. Circulation 2004;110:2168e74.
20. Ohtsuka T, Hamada M, Hiasa G, Sasaki O, Suzuki M, Hara Y, et al.
26. Troughton RW, Frampton CM, Yandle TG, Espiner EA, Nicholls MG, Effect of beta-blockers on circulating levels of inflammatory and Richards AM. Treatment of heart failure guided by plasma aminoter- anti-inflammatory cytokines in patients with dilated cardiomyopathy.
minal brain natriuretic peptide (N-BNP) concentrations. Lancet 2000; J Am Coll Cardiol 2001;37:412e7.
21. Brady HR, Brenner BM. Acute renal failure. In: Fauci AS, Isselbacher KJ, 27. Jourdain P, Gueffet P, Le Helloco J. Benefit of BNP for optimizing Braunwald E, Wilson JD, Martin JB, Kasper DL, editors. Harrison's princi- therapy in patients with heart failure due to systolic dysfunction: ples of internal medicine. 13th ed. New York: McGraw-Hill; 1994. p. 1270.
the Systolic Heart Failure Treatment Supported by BNP Trial failure. In: Fauci AS, Isselbacher (STARS-BNP) multicenter randomized study. Eur J Heart Fail Suppl Braunwald E, Wilson JD, Martin JB, Kasper DL, editors. Harrison's

Source: http://biochemmack.ru/upload/iblock/d17/d17d21ba3fbdff9c8b39e7806822267a.pdf