AU2013204539B2 - Predicting mortality and detecting severe disease - Google Patents

Abstract

Measurement of circulating ST2 and/or IL-33 concentrations is useful for the prognostic evaluation of subjects, in particular for the prediction of adverse clinical outcomes, e.g., mortality, and the detection of severe disease.

Description

PREDICTING MORTALITY AND DETECTING SEVERE DISEASE This is a divisional of Australian patent application No. 2012202069, which in turn is a divisional of Australian patent application No. 2007244927, the entire content of which is incorporated herein by reference. FIELD OF THE INVENTION The invention relates to methods for predicting mortality and detecting the presence of severe disease by measuring circulating levels of ST2 and/or IL-33, alone or in combination with other biomarkers. BACKGROUND Clinical evaluation of subjects, particularly those with non-specific symptoms such as chest pain or discomfort, shortness of breath, nausea, vomiting, eructation, sweating, palpitations, lightheadedness, fatigue, or fainting, can present significant challenges, as the cause and severity of the subject's condition may not always be apparent. The decision whether to treat a subject aggressively or conservatively, or to admit the subject as an inpatient or to send them home, may sometimes be made solely on a physician's clinical assessment or "gut feeling" as to the individual's actual condition. Biomarkers that indicate a subject's likelihood of an adverse outcome, e.g., mortality, and/or the presence of severe disease, would significantly enhance the physician's ability to make informed treatment decisions. SUMMARY The present invention is based, at least in part, on the discovery that serum levels of the biomarker ST2 (Growth Stimulation-Expressed Gene 2, also known as Interleukin 1 Receptor Like 1 (IL lRL-1)) can be used to predict clinical outcome, e.g., death, within a specific time period, e.g., 30 days, 3 or 6 months, or a year or more, or to detect the presence of severe disease, regardless of the underlying causes of symptoms or ultimate diagnosis. Changes in the level of ST2 over time, e.g., over a few days or more, can also be used to predict clinical outcome, e.g., in patients hospitalized after an acute event. The methods described herein include measuring ST2 levels as well as monitoring changes in ST2 levels over time (e.g., ratios) to provide diagnostic and prognostic evaluation of patients, e.g., patients with non-specific symptoms, e.g, acutely dyspneic patients and those with chest pain. 1 WO 2007/127749 PCT/US2007/067333 IL-33 has been identified as the ligand for ST2. Thus, the invention includes methods for evaluating patients by monitoring biomarker levels of ST2 and/or IL-33 levels, as well as ST2IL-33 complexes, and ratios of ST2:IL-33 complexes to free ST2 and/or IL-33. 5 In addition, the methods can include using additional diagnostic methods, including evaluating organ function and/or levels of adjunct biomarkers such as troponin (Tn, e.g., TnI or TnT), brain natriuretic peptide (BNP), proBNP, NT proBNP, atrial natriuretic peptide (ANP), NT-proANP, proANP, C-reactive peptide (CRP), Blood Urea Nitrogen (BUN), D-dimers (degradation products of cross-linked 1o fibrin, whose level becomes elevated following clot formation), albumin, liver function enzymes, measures of renal function (e.g., creatinine, creatinine clearance rate, or glomerular filtration rate) and/or bacterial endotoxin. In some embodiments, the methods include measuring ST2 and/or a change in ST2 levels over time in addition to BUN, NT-proBNP or BNP, and/or TnI. 15 Thus, in one aspect, the invention features methods for evaluating the risk of death or readmission within a specific time period, e.g., 30, 60, 90, or 180 days (e.g., one, two, three, or six months), or one, two, or five years, for a subject. The methods include obtaining a sample, e.g., blood, serum, plasma, urine, or body tissue from the subject; determining a biomarker level of ST2 and/or IL-33 in the sample; and 20 comparing the biomarker level of ST2 and/or IL-33 in the sample to a reference level of ST2 and/or IL-33. A comparison of the biomarker level of ST2 and/or IL-33 in the sample versus the reference indicates the subject's risk of death or readmission within the specific time period. In some embodiments, the specific time period is one year. In some embodiments, the reference level represents a level in a subject who 25 has a low risk of death within one year. In some embodiments, e.g., wherein the biomarker level of ST2 is measured using an immunoassay, e.g., an enzyme-linked immunosorbent assay (ELISA), e.g., as described in Example 1, the reference level of ST2 is between about 0.2 and 0.3 ng/ml of serum, e.g., the level can be 0.20, 0.23, 0.25, 0.27, or 0.29 ng/ml of serum, and a level in the sample that is greater than or 3o equal to the reference level indicates that the subject has an elevated, i.e., statistically significantly elevated, risk of death within one year. If an analytical technique other than the ELISA described in Example 1 is employed, the reference ST2 level may be different than described herein; the specific numbers recited herein should be 2 Hisvnnlerwo~vn\N~RPoriDCC\SVh65T44_ Idoe,-IMI/W20 15 construed to be equivalent to corresponding numbers generated using other analytical techniques. In some embodiments, the elevated risk of death is at least 20% higher, e.g., 30%, 40%, or 50% higher. In another aspect, the invention features methods for determining the severity of one or more diseases, e.g., the present severity of diseases, in a subject. The methods include obtaining a sample from the subject; determining a biomarker level of ST2 and/or IL-33 in the sample; and comparing the biomarker level of ST2 and/or IL-33 in the sample to a reference level of ST2 and/or IL-33. The biomarker level of ST2 and/or IL-33 in the sample as compared to the reference indicates whether the one or more diseases the subject has are severe, e.g., life-threatening. In a further aspect, the invention includes methods for monitoring a subject's condition, e.g., for deciding whether a subject has improved, e.g., improved sufficiently to be discharged from the hospital. The methods include determining a first biomarker level of ST2 and/or IL-33 in the subject, e.g., a baseline level; and determining at least one subsequent biomarker level of ST2 and/or IL-33 in the subject, e.g., a treatment level. Then, the first level and the subsequent levels are compared. If the biomarker level of ST2 and/or IL-33 decreases sufficiently, e.g., statistically significantly, or by at least 5%, 10%, 15%, 20%, or more, from the first to the subsequent levels, then the subject's condition is likely to be improving and, if either one or both levels are low enough, e.g., below a selected threshold, then the subject can be discharged, e.g., for outpatient treatment. In a further aspect, the invention includes an in vitro method for selecting a treatment for a subject having undiagnosed dyspnea, undiagnosed chest pain, undiagnosed chest pain, or undiagnosed chest discomfort, the method comprising: (a) determining a first level of soluble ST2 in a biological sample from a subject having undiagnosed dyspnea, undiagnosed chest pain, or undiagnosed chest discomfort at a first time point; 3 H :\sym\Intemoven\ NR~oibIDCCS\S34_ .docs4I5/O/20I15 (b) determining a second level of soluble ST2 in a biological sample from the subject at a second time point; (c) determining a ratio of the second level of soluble ST2 to the first level of soluble ST2; and (d) identifying a subject that has a ratio of the second level of soluble ST2 to the first level of soluble ST2 that is elevated as compared to a reference ratio as having an increased risk of death within one year, and selecting aggressive treatment for the subject; or (e) identifying a subject that has a ratio of the second level of soluble ST2 to the first level of soluble ST2 that is decreased as compared to a reference ratio as having a decreased risk of death within one year, and selecting a-conservative treatment for the subject. In some embodiments, the methods include determining a level of ST2 that indicates a subject's risk, and optionally selecting or modifying a treatment for the subject, based on a ratio of a first ST2 level, e.g., a baseline level, to a second ST2 level, e.g,. a level taken some time later, e.g., one, two, three, four, or more days later. For example, if the second level of ST2 is more than a selected percentage of the first level, then the subject has a high risk and should be treated more aggressively; if the subject is already being treated, then the subject is not responding favorably to the current treatment and a new treatment should be selected, i.e., an alternate treatment to which the patient may respond more favorably. As one example, if the second level is about 85% or more of the first level (i.e., has decreased by about 15% or less), then the subject is not improving and still has a high risk of death. 3A WO 2007/127749 PCT/US2007/067333 In some embodiments, the level of ST2 in a subject is compared to a reference level that represents a level in a subject who does not have severe disease, e.g., has no disease or has no acute, severe disease, e.g., when measured using an ELISA, e.g., as described herein. The reference level of ST2 can be between about 0.2 and 0.3 ng/ml, 5 e.g., the level can be about 0.20, 0.23, 0.25, 0.27, or 0.29 ng/ml of serum or plasma (as noted above, the thresholds recited herein apply when using an ELISA method as described herein; other threshold numbers can be considered as equivalent to these numbers when determined using a different method). A level in the sample that is greater than or equal to the reference level indicates that the subject has one or more 10 severe diseases, e.g., present diseases. In some embodiments, the reference level represents a subject with a certain prognosis. For example, when the level of ST2 is measured using an ELISA, e.g., as described herein in Example 1, the reference level can be used to determine prognosis as follows: an ST2< about 0.2 or 0.23 ng/ml indicates that the subject has a good 15 prognosis, e.g., is likely to recover; an ST2 of from about 0.2 or 0.23 ng/ml to 0.7 ng/ml (or an equivalent thereof) indicates that the subject has a poor prognosis, e.g., is less likely to recover. Finally, an ST2 of greater than 0.7 ng/ml indicates a very poor prognosis, e.g., the subject is not likely to recover. In this embodiment poor prognosis would indicate that the patient is at risk of death or developing more severe 20 disease within one year possibly requiring hospital admission. Very poor prognosis indicates that the patient has a high probability of death or developing more severe disease within 90 days possibly requiring hospital admission. In one study patients with an ST2 level higher than 0.7 ng/ml had a mortality rate of over 30%. In some embodiments, the subject exhibits one or more non-specific 25 symptoms, e.g., chest pain or discomfort, shortness of breath (dyspnea), nausea, vomiting, eructation, sweating, palpitations, lightheadedness, fatigue, and fainting. In some embodiments, the symptom is dyspnea or chest pain. In some embodiments, the subject does not have a cardiovascular disorder. In various embodiments, the subject has a pulmonary disorder, e.g., acute infection (e.g., 30 pneumonia), chronic obstructive pulmonary disease (COPD), and pulmonary embolism. 4 WO 2007/127749 PCT/US2007/067333 In certain embodiments, the subject has a liver disorder, e.g., a liver disorder associated with chemotherapy, alcohol toxicity, or drug toxicity as determined by standard liver function laboratory tests. In some embodiments, the methods further include determining the level of an 5 adjunct (non-ST2, non-IL-33) biomarker, e.g., Troponin, NT-proBNP, BNP, proBNP, NT-proANP, proANP, ANP, CRP, D-dimers, BUN, albumin, liver function enzymes, measures of renal function, e.g., creatinine, creatinine clearance rate, or glomerular filtration rate, and/or bacterial endotoxin, in the sample; and comparing the level of the adjunct biomarker in the sample to a reference level of the adjunct biomarker. The 10 level of the adjunct biomarker in the sample as compared to the reference, in combination with the level of ST2 in the sample as compared to an ST2 reference level, indicates whether the subject has an elevated risk of death within a specific time period, and/or has a present severe disease. In some embodiments, the methods include determining a change in levels over time (e.g., a ratio) for the adjunct 15 biomarker, by comparing a first level, e.g., a baseline level, to a second level, e.g., a level taken some time later, e.g., one, two, three, four, or more days later. In embodiments where a ratio of ST2 is calculated, a ratio of an adjunct biomarker can also be calculated, e.g., based on the same time period as the ratio of ST2. In some embodiments, the subject has a BMI of 25-29, a BMI of 30, or 20 renal insufficiency, e.g., the subject is selected on the basis that they have a BMI of 25-29, a BMI of > 30, or renal insufficiency. In some embodiments, the methods include determining a level of ST2 and a level of IL-33 in the sample; determining a ratio of ST2:IL-33 in the sample; and comparing the ratio of ST2:IL-33 to a reference ratio. The ratio of ST2:IL-33 in the 25 sample as compared to the reference ratio indicates whether the subject has an elevated risk of death within a specific time period, and/or has present severe disease. In another aspect, the invention provides methods for evaluating the risk of death within a specific time period, e.g., 30, 60,90, or 180 days (6 months), or one, two, or five years, e.g., for a subject who exhibits one or more non-specific 30 symptoms. The methods include obtaining a sample from the subject; determining a biomarker level of ST2 and/or IL-33 in the sample, and optionally a level of NT proBNP, proBNP, or BNP in the sample; and comparing the level of ST2 and/or IL-33 in the sample, and the level of NT-proBNP in the sample (if determined), to 5 WO 2007/127749 PCT/US2007/067333 corresponding reference levels. The level of ST2 and/or IL-33 in the sample, and the level of NT-proBNP, proBNP, or BNP in the sample, as compared to the respective reference levels indicate the subject's risk of death within the specific time period. In some embodiments, the methods include determining levels of (i) NT 5 proBNP, proBNP, or BNP and (ii) ST2 in the sample. In some embodiments, the subject's risk of death within one year is as follows: HIMI .......... 0.20 <986pgm Lowest Risk Medium Risk T-proBNP 986 pg/m Medium Risk Highest Risk In certain embodiments, if the subject has the highest level of risk of death or hospital readmission, the subject is treated aggressively. 10 In a further aspect, the invention includes methods for monitoring a subject's condition, e.g., for deciding whether a subject has improved, e.g., improved sufficiently to discharge the subject from the hospital. The methods include determining a first level of (i) a non-ST2 biomarker, e.g., NT-proBNP, proBNP, or BNP and (ii) ST2 and/or IL-33 in the subject, e.g., a baseline level; and determining at 15 least one subsequent level of (i) the non-ST2 biomarker, e.g., NT-proBNP, proBNP, or BNP and (ii) ST2 and/or IL-33 in the subject, e.g., a treatment level. Then, the first level and the subsequent levels are compared. If the level of ST2 and/or IL-33 decreases from the first to the subsequent levels, then the subject is likely to be improving and, if the level is low enough, then the subject has improved, e.g., 20 improved sufficiently to be discharged, e.g., for outpatient treatment. In some embodiments, the methods include determining at least a first, second, and third level of (i) a non-ST2 biomarker, e.g., NT-proBNP, proBNP, or BNP and (ii) ST2 and/or IL 33 in the subject, and comparing the levels. A difference between the levels indicates whether the subject has improved sufficiently to be discharged. Thus, for example, a 25 decision to discharge or continue to treat on an inpatient basis can be made as follows: 6 WO 2007/127749 PCT/US2007/067333 ST2> threshold ST2 < threshold Prognosis very poor, Patient may still need treatment, BNP > threshold requires intensive favorable prognosis, possibly treatment, do not discharge as outpatient _______________discharge______________ Prognosis poor, patient BNP <threshold may have complicating Prognosis good illness, if discharge monitor closely In some embodiments, the threshold for ST2 is 0.2 ng/ml, and the threshold for BNP is 986 pg/ml. Also provided herein are kits including one or more antibodies that bind 5 specifically to ST2 and/or one or more antibodies that bind specifically to IL-33, and instructions for performing one or more of the methods described herein. In a further aspect, the invention provides methods for evaluating a subject's condition. The methods include obtaining a sample from the subject; determining a biomarker level of ST2 and/or 1L-33 in the sample, and determining presence or a 10 level of one or more, e.g., all, of the following other biomarkers: (i) NT-proBNP, proBNP, or BNP; (ii) NT-proANP, proANP, orANP; (iii) cardiac troponin (cTn), e.g., cTnI; (iv) D-dimers; 15 (v) C-reactive protein (CRP); (vi) creatinine, creatinine clearance rate, or glomerular filtration rate; (vii) Blood Urea Nitrogen (BUN); (viii) bacterial endotoxin; (ix) one or more liver function enzymes; and 20 comparing the level of ST2 and/or IL-33 in the sample, and the level of the one or more other biomarkers in the sample, to corresponding reference levels. The level of ST2 and/or IL-33 in the sample, and the level of the other biomarker in the sample, as compared to the reference levels indicate the severity of the subject's condition. 25 In some embodiments, the methods described herein include measuring levels or ratios of ST2 and/or IL-33 in combination with BNP or NT-proBNP; with troponin, e.g., TnI or TnT; or with a measure of renal function, e.g., creatinine, creatinine clearance rate, or glomerular filtration rate. 7 WO 2007/127749 PCT/US2007/067333 In another aspect, the invention includes methods for evaluating the efficacy of a treatment in a subject. The methods include determining a first (e.g., baseline) level of circulating ST2 and/or IL-33 in a subject; comparing the first level of 5 circulating ST2 and/or IL-33 in the subject to a predetermined reference level; selecting the subject if the first level of ST2 is above the predetermined reference level; administering a treatment to the subject; determining a second level of circulating ST2 and/or IL-33 in a subject; and comparing the first and second levels of circulating ST2 and/or IL-33. A difference 10 between the first and second levels of circulating ST2 and/or M-33 indicates the efficacy of the treatment in the subject. For example, a second level of circulating ST2 and/or IL-33 that is lower than the first level indicates that the treatment is effective. As used herein, a "sample" includes any bodily fluid or tissue, e.g., one or 15 more of blood, serum, plasma, urine, and body tissue. In certain embodiments, a sample is a serum, plasma, or blood sample. An antibody that "binds specifically to" an antigen, binds preferentially to the antigen in a sample containing other proteins. The methods and kits described herein have a number of advantages. For 20 example, the methods can be used to determine whether a patient should be admitted or held as an inpatient for further assessment, regardless of whether a definitive diagnosis has been made. For example, the methods can be used for risk stratification of a given subject, e.g., to make decisions regarding the level of aggressiveness of treatment that is appropriate for the subject, based on their ST2 levels. Better 25 treatment decisions can lead to reduced morbidity and mortality, and better allocation of scarce health care resources. The methods described herein can be used to make general assessments as to whether a patient should be further tested to determine a specific diagnosis. The methods described herein can also be used for patient population risk stratification, e.g., to provide information about clinical performance 30 or expected response to a therapeutic intervention. The methods described herein can be used regardless of the underlying cause or ultimate diagnosis, and therefore are not limited to specific indications. 8 WO 2007/127749 PCT/US2007/067333 Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Methods and materials are described herein for use in the present invention; other, suitable methods and materials known in the art can also 5 be used. The materials, methods, and examples are illustrative only and not intended to be limiting. All publications, patent applications, patents, sequences, database entries, and other references mentioned herein are incorporated by reference in their entirety. In case of conflict, the present specification, including definitions, will control. 10 Other features and advantages of the invention will be apparent from the following detailed description and figures, and from the claims. DESCRIPTION OF DRAWINGS FIG 1 is a graph illustrating receiver operating characteristic analysis for ST2 and death within one year. ST2 was useful for this purpose, as indicated by the high 15 area under the curve (AUC). FIG. 2 is a bar graph illustrating the crude rates of death across deciles of ST2 in the ProBNP Investigation of Dyspnea in the Emergency Department (PRIDE) study cohort. A clear threshold effect is noted at decile 5, corresponding to an ST2 concentration of 0.23 ng/ml for the particular assay. 20 FIGs. 3A and 3B are a pair of Kaplan-Meier hazard curves depicting the rates of death from presentation to one year of follow up in patients with dyspnea, stratified as a function of ST2 concentrations. Among dyspneic patients with ST2 concentrations of> 0.20 ng/ml, a high rate of mortality was noted within days of presentation, and extending to a full year from presentation. The rates of death were 25 similar among those with (3A) and without (3B) acute heart failure (all Log-rank p values < 0.001). FIG 4 is a bar graph illustrating mortality rates as a function of marker concentrations for NT-proBNP and ST-2. FIG. 5 is receiver operating curve (ROC) of specificity versus sensitivity, for 30 change in ST2 (light grey line) and change in BNP (dark line). FIG 6 is a combination bar and line graph. The bars illustrate the percent mortality in populations with the indicated levels of BUN and ST2 ratios. The line indicates the number of patients that are in each category. 9 WO 2007/127749 PCT/US2007/067333 FIG 7A is a line graph of average ST2 values for survivors (light grey squares) and non-survivors (dark diamonds) on each day of hospitalization. FIGs. 7B-7D are whisker box plots bracketing the 25th and 75th percentiles of ST2 levels (7B), BNP concentration (7C), and NT-proBNP (7D), plotted against time 5 (over 6 days of hospitalization); S = survivors, D = decedents. FIG 8 is a line graph of ratios of ST2 values for survivors (light grey squares) and non-survivors (dark diamonds) as compared to baseline (around admission) on each day of hospitalization. FIG 9 is a ROC for ST2 ability to predict death by one year in PRIDE subjects 10 with a pulmonary diagnosis. Area under ROC = 0.73; 95% CI = 0.62-0.83; P < 0.0001; Optimal cut: 0.20 ng/ml; 88% sensitive, 52% specific; PPV = 22%; NPV= 96%. FIG 10 is a box graph illustrating the correlation between ST2 concentrations and risk of death within one year in PRIDE subjects with a pulmonary diagnosis. 15 FIG 11 is a line graph illustrating the mortality rate as a function of ST2 concentration in PRIDE subjects with a pulmonary diagnosis. P <0.001. FIGs. 12A-B are box graphs illustrating mean Glomerular Filtration Rate (GFR, 12A) and ST2 levels (12B) in a population of 133 subjects with moderate to severe renal insufficiency. 20 FIG 13 is a bar graph illustrating the distribution of ST2 levels in the population described in Example 8. DETAILED DESCRIPTION Clinical evaluation of patients, particularly patients with non-specific symptoms such as dyspnea or chest pain, is often challenging. The results described 25 herein provide evidence that ST2 is useful in the prognostic evaluation of patients, regardless of the underlying cause of their disease. ST2 is a powerful indicator of severe disease and imminent death, as demonstrated herein in several completely different populations with completely different symptoms (see Examples 1-6). Predicting Mortality 30 Elevated concentrations of ST2 are markedly prognostic for death within one year, with a dramatic divergence in survival curves for those with elevated ST2 soon after presentation, regardless of the underlying diagnosis. As one example, there is a 10 WO 2007/127749 PCT/US2007/067333 dramatic relationship between elevations of ST2 and the risk for mortality within one year following presentation with dyspnea. The relationship between ST2 and death in dyspneic patients was independent of diagnosis, and superseded all other biomarker predictors of mortality in this setting, including other markers of inflammation, 5 myonecrosis, renal dysfunction, and most notably NT-proBNP, a marker recently described as having value for predicting death in this population (Januzzi et al., Arch. Intern. Med. 2006; 166(3):315-20). Indeed, most of the mortality in the study was concentrated among subjects with elevated ST2 levels at presentation; however, the combination of an elevated ST2 and NT-proBNP was associated with the highest rates 10 of death within one year. Such a multi-marker approach for risk stratification has been proposed for patients with acute coronary syndromes (Sabatine et al., Circulation 2002;105(15):1760-3), but no such strategy has yet been proposed for the evaluation for the patient with non- specific symptoms such as undifferentiated dyspnea or 15 general complaint of chest pain. Determining Severity of Disease Elevated concentrations of ST2 are correlated with the presence of severe disease in a subject, regardless of the underlying cause of the disease. As one 20 example, in a population of patients presenting with chest pain, the highest levels of disease were associated with severe disease including chronic obstructive pulmonary disease (COPD), lymphoma, sepsis, alcohol abuse, and pulmonary embolism (see Example 5). Therefore, for undiagnosed subjects, the methods described herein can be used 25 to determine how aggressively a diagnosis should be sought; a high ST2 level would indicate the presence of severe disease, and suggest that the subject should be treated as a high-risk case. For subjects with a known diagnosis, the methods described herein can be used to help determine the severity of the underlying pathology; again, a higher ST2 level is associated with more severe disease. 30 General Methodology In general, the methods described herein include evaluating circulating levels (e.g., levels in blood, serum, plasma, urine, or body tissue) of ST2 and/or IL-33 in a 11 WO 2007/127749 PCT/US2007/067333 subject, e.g., a mammal, e.g., a human. These levels provide information regarding the subject's likelihood of experiencing an adverse outcome, e.g., mortality, e.g., within a specific time period, e.g., 30 days, 60 days, 90 days, 6 months, one year, two years, three years, or five years. These levels also provide information regarding the 5 severity of disease in the subject. In some embodiments, the level of ST2 and/or IL 33 is determined once, e.g., at presentation. In some embodiments, the level of ST2 and/or IL-33 is determined 2, 4, 6, 8, 12, 18, and/or 24 hours, and/or 1-7 days after the onset of symptoms. Where more than one level is determined, a ratio of ST2 can be calculated that quantifies whether and how much the level of ST2 in the subject has 10 increased or decreased. In some embodiments, the level of ST2 and/or IL-33 is determined more than once; in that case, the higher measurement, or the most recent measurement, can be used. In embodiments where the level of ST2 and/or IL-33 is determined more that once, the highest level can be used, or the difference between the levels (i.e., the 15 magnitude and direction of the difference) can be determined and used. Thus, a ratio of ST2 can be determined that represents the change (e.g., the magnitude and direction, e.g., increase or decrease) in ST2 levels over time, e.g., over the course of a few days, e.g., 3 days or more, or over the course of weeks or months; the ratio is indicative of the subject's risk level and the presence of severe disease. Levels of ST2 20 and/or IL-33 can also be determined multiple times to evaluate a subject's response to a treatment. For example, a biomarker level of ST2 and/or IL-33 taken after administration of a treatment, e.g., one or more doses or rounds of a treatment, can be compared to levels of ST2 and/or IL-33 before the treatment was initiated. The difference between the ST2 levels would indicate whether the treatment was effective; 25 e.g., a reduction in ST2 levels would indicate that the treatment was effective. The difference between the ST2 levels can also be used to monitor a subject's condition, e.g., to determine if the subject is improving, e.g., improving enough to be discharged from a hospital, to be treated less aggressively, or to be followed up at greater time intervals. 30 Evaluating circulating levels of ST2 and/or IL-33 in a subject typically includes obtaining a biological sample, e.g., serum, plasma or blood, from the subject. Levels of ST2 and/or IL-33 in the sample can be determined by measuring levels of polypeptide in the sample, using methods known in the art and/or described herein, 12 WO 2007/127749 PCT/US2007/067333 e.g., immunoassays such as enzyme-linked immunosorbent assays (ELISA). Alternatively, levels of ST2 and/or IL-33 mRNA can be measured, again using methods known in the art and/or described herein, e.g., by quantitative PCR or Northern blotting analysis. 5 Once a level or ratio of ST2 and/or IL-33 has been determined, the level or ratio can be compared to a reference level or ratio. In some embodiments, e.g., where the level of ST2 is determined using an ELISA, e.g., as described in Example 1, the reference level will represent a threshold level, above which the subject has an increased risk of death, and/or has a severe disease. The reference level chosen may 10 depend on the methodology used to measure the levels of ST2. For example, in some embodiments, where circulating levels of soluble ST2 are determined using an immunoassay, e.g., as described herein, the reference level is about 0.20, 0.23, or 0.29 ng/ml of serum, and a level of ST2 above that reference level indicates that the subject has an increased risk of death, and/or has a severe disease. 15 Where a ratio has been determined, e.g., using a first and second measurement of ST2, the reference ratio will represent an amount and direction of change that indicates whether the subject has an increased risk of death and/or has a severe disease. As one example, an ST2 ratio can be calculated based on a first measurement, e.g., a baseline measurement taken when a subject presents for 20 treatment, e.g., to an ED, and a second measurement, e.g., a measurement taken about three to four days later. If the ratio of first and second ST2 levels over time is about 0.85 or higher, i.e., the ST2 levels have decreased less than about 15% (or have stayed the same or increased), then the subject has a very high risk of imminent death. Ratios below about 0.85 (where the ST2 levels have decreased more than about 15%) 25 indicate that the subject has a lower risk of imminent death. This information allows a treating physician to make more accurate treatment decisions; for example, when the results of the determination indicate that the subject has a level equal to or above a reference level, e.g., above about 0.20 ng/ml, 0.23 ng/ml, or 0.29 ng/ml of serum, or a ratio above a reference ratio, the subject may be 30 admitted to the hospital as an inpatient, e.g., in an acute or critical care department. In some embodiments, comparison of ST2 to a reference level or ratio can be used to determine a subject's prognosis. For example, when the level of ST2 is measured using an ELISA, e.g., as described herein in Example 1, the reference level can be 13 WO 2007/127749 PCT/US2007/067333 used to determine prognosis as follows: an ST2 <about 0.20 ng/mIl or 0.23 ng/ml indicates that the subject has a good prognosis, e.g., is likely to recover; an ST2 of from about 0.20 ng/ml or 0.23 ng/ml to about 0.7 ng/ml indicates that the subject has a poor prognosis, e.g., is less likely to recover. Finally, an ST2 of greater than about 5 0.7 ng/ml indicates a very poor prognosis, e.g., the subject is not likely to recover. As another example, a ratio of first and second ST2 levels over time above about 0.85 indicates a poor prognosis, while a ratio of about 0.85 of below indicates a good prognosis. Additional testing may be performed, to determine the subject's actual 10 condition. More aggressive treatment may be administered either before or after additional testing. For example, in the case of a suspected MI the subject may be sent for more extensive imaging studies and/or cardiac catheterization. In some embodiments, both levels of ST2 and IL-33 are determined, and the information from the comparison of both biomarkers with their respective reference 15 levels provides cumulative information regarding an increased risk of death, and/or presence of a severe disease in the subject. In some embodiments, the ratio of ST2 to IL-33 may be determined, and the ratio compared to a reference ratio that represents a threshold ratio above which the subject has an increased risk of death, and/or has a severe disease. In some embodiments, the presence of IL-33/ST2 complexes is 20 detected, and the level of such complexes is indicative of risk of death and/or the presence of severe disease. In some embodiments, the methods include the use of additional diagnostic methods to identify underlying pathology. Any diagnostic methods known in the art can be used, and one of skill in the art will be able to select diagnostic methods that 25 are appropriate for the subject's symptoms. In some embodiments, the methods described herein include other diagnostic methods in addition to or as an alternative to the measurement of other biomarkers, e.g., physical measurements of lung function or cardiac function as are known in the art. For example, the methods described herein include measuring levels of ST2 30 and/or IL-33 and one or more additional biomarkers that aid in the subject's diagnosis. As one example, for a subject who has chest pain or dyspnea, biomarkers indicative of cardiac disease can be measured, e.g., cardiac troponin (cTn), e.g., cTnI, BNP, and/or ANP; alternatively or in addition, biomarkers of pulmonary disease can be measured, 14 WO 2007/127749 PCT/US2007/067333 e.g., D-dimers for pulmonary embolism. Thus, in subjects presenting with symptoms that include MI in their differential diagnoses, the methods can include measuring levels of, cTnI, BNP or NTproBNP or proBNP in addition to ST2 and/or IL-33, to determine whether the subject is having an MI. In subjects presenting with symptoms 5 that include heart failure (HF) in their differential diagnoses, the methods can include measuring levels of BNP or NTproBNP or proBNP in addition to ST2 and/or IL-33, to determine whether the subject is having HF. In subjects presenting with symptoms that include COPD in their differential diagnoses, the methods can include measuring lung function in addition to levels of ST2 and/or IL-33, to determine whether the 10 subject has COPD. One of skill in the art will appreciate that there are a number of additional diagnostic methods that can be applied, depending on the situation and the subject's condition. In some embodiments, the methods include measuring levels of BUN, and the presence of elevated BUN and elevated ST2 places the subject in the highest risk category. 15 ST2 The ST2 gene is a member of the interleukin-1 receptor family, whose protein product exists both as a trans-membrane form, as well as a soluble receptor that is detectable in serum (Kieser et al., FEBS Lett. 372(2-3):189-93 (1995); Kumar et al., J. 20 Biol. Chem. 270(46):27905-13 (1995); Yanagisawa et al., FEBS Lett. 302(l):51-3 (1992); Kuroiwa et al., Hybridoma 19(2):151-9 (2000)). ST2 was recently described to be markedly up-regulated in an experimental model of heart failure (Weinberg et al., Circulation 106(23):2961-6 (2002)), and preliminary results suggest that ST2 concentrations may be elevated in those with chronic severe HF (Weinberg et al., 25 Circulation 107(5):721-6 (2003)) as well as in those with acute myocardial infarction (MI) (Shimpo et al., Circulation 109(18):2186-90 (2004)). The trans-membrane form of ST2 is thought to play a role in modulating responses of T helper type 2 cells (Lohning et al., Proc. Natl. Acad. Sci. U. S. A. 95(12):6930-5 (1998); Schmitz et al., Immunity 23(5):479-90 (2005)), and may play a 30 role in development of tolerance in states of severe or chronic inflammation (Brint et al., Nat. Immunol. 5(4):373-9 (2004)), while the soluble form of ST2 is up-regulated in growth stimulated fibroblasts (Yanagisawa et al., 1992, supra). Experimental data suggest that the ST2 gene is markedly up-regulated in states of myocyte stretch 15 WO 2007/127749 PCT/US2007/067333 (Weinberg et al., 2002, supra) in a manner analogous to the induction of the BNP gene (Bruneau et al., Cardiovasc. Res. 28(10):1519-25 (1994)). Tominaga, FEBS Lett. 258:301-304 (1989), isolated murine genes that were specifically expressed by growth stimulation in BALB/c-3T3 cells; they termed one 5 of these genes St2 (for Growth Stimulation-Expressed Gene 2). The St2 gene encodes two protein products: ST2, which is a soluble secreted form; and ST2L, a transmembrane receptor form that is very similar to the interleukin- 1 receptors. The HUGO Nomenclature Committee designated the human homolog, the cloning of which was described in Tominaga et al., Biochim. Biophys. Acta. 1171:215-218 10 (1992), as Interleukin I Receptor-Like I (ILIRLI). The two terms are used interchangeably herein. The mRNA sequence of the shorter, soluble isoform of human ST2 can be found at GenBankAcc. No. NM__003856.2, and the polypeptide sequence is at GenBankAcc. No. NP_003847.2; the mRNA sequence for the longer form of human 15 ST2 is at GenBankAcc. No. NM_016232.4; the polypeptide sequence is at GenBank Acc. No. NP_057316.3. Additional information is available in the public databases at GeneID: 9173, MIM ID # 601203, and UniGene No. Hs.66. In general, in the methods described herein, the soluble form of ST2 polypeptide is measured. Methods for detecting and measuring ST2 are known in the art, e.g., as 20 described in U.S. Pat. Pub. Nos. 2003/0124624, 2004/0048286 and 2005/0130136, the entire contents of which are incorporated herein by reference. Kits for measuring ST2 polypeptide are also commercially available, e.g., the ST2 ELISA Kit manufactured by Medical & Biological Laboratories Co., Ltd. (MBL International Corp., Woburn, MA), no. 7638. In addition, devices for measuring ST2 and other biomarkers are 25 described in U.S. Pat. Pub. No. 2005/0250156. In some embodiments, the methods include determining the identity of the nucleotide sequence at RefSNP ID: rs1041973. IL-33 30 IL-33 was recently identified as the ligand for ST2, and the presence of increased levels of IL-33 in various inflammatory disorders has been described (see Schmitz et al., Immunity 23(5):479-90 (2005); U.S. Pat. Pub. No. 2005/0203046). In 16 WO 2007/127749 PCTIUS2007/067333 the methods described herein, IL-33 can be measured instead of or in addition to ST2. The ratio of ST2 to IL-33 can also be determined. The nucleic acid sequence of IL-33 can be found at GenBank Acc. No. NM_033439.2, and the polypeptide sequence is at GenBankAcc. No. NP_254274.1. 5 Additional information is available in the public databases at GeneID: 90865, MIM ID # *608678, and UniGene No. Hs.348390. IL-33 is also known as Chromosome 9 Open Reading Frame 26 (C90RF26); Nuclear Factor from High Endothelial Venules (NFHEV); and Interleukin 33. See also Baekkevold et al., Am. I. Path. 163: 69-79 (2003). 10 Methods for measuring levels of IL-33 are known in the art, see, e.g., Schmitz et al., Immunity. 23(5):479-90 (2005), and U.S. Pat. Pub. No. 2005/0203046. Other Biomarkers The methods described herein can also include measuring levels of other 15 biomarkers in addition to ST2 and/or IL-33. Suitable biomarkers include proBNP, NT-proBNP, BNP, NT-proANP, proANP, ANP, troponin, CRP, IL-6, D-dimers, BUN, liver function enzymes, albumin, measures of renal function, e.g., creatinine, creatinine clearance rate, or glomerular filtration rate, and/or bacterial endotoxin. Methods for measuring these biomarkers are known in the art, see, e.g., U.S. Pat. Pub. 2o Nos. 2004/0048286 and 2005/0130136 to Lee et al.; Dhalla et al., Mol. Cell. Biochem. 87:85-92 (1989); Moe et al., Am. Heart. J. 139:587-95 (2000); Januzzi et al., Eur. Heart J. 27(3):330-7 (2006); Maisel et al., J. Am. Coll. Cardiol. 44(6):1328 33 (2004); and Maisel et al., N. Engl. J. Med. 347(3):161-7 (2002), the entire contents of which are incorporated herein by reference. Liver function enzymes 25 include alanine transaminase (ALT); aspartate transaminase (AST); alkaline phosphatase (ALP); and total bilirubin (TBIL). In these embodiments, levels of ST2 and/or IL-33 and one or more additional biomarkers are determined, and the information from the comparison of the biomarkers with their respective reference levels provides additional information 3o regarding the subject's risk of death and/or the presence of a severe disease in the subject, which may provide more accurate and specific information regarding the subject's risk. The levels can then be compared to a reference ratio that represents a 17 WO 2007/127749 PCT/US2007/067333 threshold ratio above which the subject has an increased risk of death, and/or has a severe disease. As one example, the methods can include determining levels of NT-proBNP and ST2. The levels indicate the subject's risk of death, e.g., as shown in Table IA. 5 Table 1A: Risk of Death Based on Circulating Levels of NT-proBNP and ST2 ST2 <0.20i nI ST2 .2o/ml, NT-proBNP <986'pg/i Lowest Risk Medium Risk NT-proBNP>986 pgi Medium Risk Highest Risk As shown in Table IA, the lowest risk of death, e.g., no greater risk of death 10 than in normal patients or healthy individuals, occurs when both ST2 and NT-proBNP levels are low, and the highest risk of death, i.e., a statistically significantly increased risk, e.g., greater than 20% increased risk of death, e.g., a greater than 30, 40, or 50% higher risk than a normal patient or healthy individual, occurs when both ST2 and NT proBNP levels are high. 15 As another example, the methods can include determining levels of ST2 and BUN. The levels indicate the subject's risk of death, e.g., as shown in Table 1B. Table IB: Risk of Death Based on Circulating Levels of BUN and ST2 ST2<.~20 ng/ml ST22 0.20 ng/mil BUN < 40 mg/dL Lowest Risk Medium Risk BUN40 mg/dL Medium Risk Highest Risk As shown in Table IB, the lowest risk of death, e.g., no greater risk of death than in normal patients or healthy individuals, occurs when both ST2 and BUN levels 20 are low, and the highest risk of death, e.g., a statistically significantly increased risk of death, e.g., a risk greater than 30, 40, or 50% higher risk than a normal patient or healthy individual, occurs when both ST2 and BUN levels are high. Selecting a Treatment - Aggressive vs. Conservative 25 Once it has been determined that a subject has a circulating level of ST2 and/or IL-33 above a predetermined reference level, the information can be used in a 18 WO 2007/127749 PCT/US2007/067333 variety of ways. For example, if the subject has elevated ST2 levels, e.g., as compared to a reference level, a decision to treat aggressively can be made, and the subject can be, e.g., admitted to a hospital for treatment as an inpatient, e.g., in an acute or critical care department. Portable test kits could allow emergency medical 5 personnel to evaluate a subject in the field, to determine whether they should be transported to the ED. Triage decisions, e.g., in an ED or other clinical setting, can also be made based on information provided by a method described herein. Those patients with high ST2 and/or IL-33 levels can be prioritized over those with lower ST2 or IL-33 levels. 10 The methods described herein also provide information regarding whether a subject is improving, e.g., responding to a treatment, e.g., whether a hospitalized subject has improved sufficiently to be discharged and followed on an outpatient basis. In general, these methods will include determining the levels of ST2 and/or IL 33 in the subject multiple times. A decrease in ST2 and/or IL-33 levels over time 15 indicates that the subject is likely to be improving. The most recent levels of ST2 and/or IL-33 can also be compared to a threshold, as described herein, to determine whether the subject has improved sufficiently to be discharged. The subject may also be considered for inclusion in a clinical trial, e.g., of a treatment that carries a relatively high risk. The subject can be treated with a regimen 20 that carries a relatively higher risk than would be considered appropriate for someone who had a lower risk of imminent mortality, e.g., mortality within 30 days or within 1 year of presentation. Beyond the clinical setting, information regarding a subject's ST2 and/or IL 33 can be used in other ways, e.g., for payment decisions by third party payors, or for 25 setting medical or life insurance premiums by insurance providers. For example, a high level of ST2 and/or IL-33, e.g., a level above a predetermined threshold level, may be used to decide to increase insurance premiums for the subject. Patient Populations 30 The methods described herein are useful in a wide variety of clinical contexts. For example, the methods can be used for general population screening, including screening by doctors, e.g., in hospitals and outpatient clinics, as well as the ED. As 19 WO 2007/127749 PCTUS2007/067333 one example, levels of ST2 and/or IL-33 can be determined at any time, and if ST2 and/or IL-33 is elevated, the physician can act appropriately. Although the methods described herein can be used for any subject, at any time, they are particularly useful for those subjects for whom a diagnosis, or the 5 severity of a condition, is difficult to determine. For example, such subjects may present with non-specific symptoms, e.g., symptoms that do not indicate a specific diagnosis. Non-specific symptoms include, but are not limited to, chest pain or discomfort, shortness of breath, nausea, vomiting, eructation, sweating, palpitations, lightheadedness, fatigue, and fainting. Each symptom can have varied etiology. 10 Chest Pain Chest pain is the chief complaint in about I to 2 percent of outpatient visits, and although the cause is often noncardiac, heart disease remains the leading cause of death in the United States. Therefore, distinguishing between serious and benign causes of chest pain is crucial. The methods described herein are useful in making 15 this determination. A subject presenting to the ED with chest pain may have esophageal pain, an ulcer, acute lung problems such as pulmonary embolus (PE) (potentially fatal), rupturing or dissecting aneurysm (highly lethal), gall bladder attack, pericarditis (inflammation of the sack around the heart), angina pectoris (cardiac pain without 20 damage), or an MI (potentially fatal). A precise diagnosis may be difficult to make immediately, but the decision whether to admit the subject or to treat them conservatively should generally be made immediately. If the methods described herein indicate that the subject has an increased risk of an adverse clinical outcome, e.g., imminent mortality or severe disease, then the decision can be made to treat the 25 subject aggressively, to potentially prevent the adverse outcome. Additional information about treatment and diagnosis of chest pain may be found, e.g., in Cayley, Am. Fam. Phys. 72(10):2012-2028 (2005). Dyspnea Dyspnea, or shortness of breath (also defined as abnormal or uncomfortable 30 breathing), is a common symptom of subjects on presentation to the ED. The differential diagnosis for dyspnea includes four general categories: (1) cardiac, (2) pulmonary, (3) mixed cardiac or pulmonary, and (4) noncardiac or nonpulmonary. 20 WO 2007/127749 PCT/US2007/067333 Cardiac causes of dyspnea include right, left, or biventricular congestive heart failure with resultant systolic dysfunction, coronary artery disease, recent or remote myocardial infarction, cardiomyopathy, valvular dysfunction, left ventricular hypertrophy with resultant diastolic dysfunction, asymmetric septal hypertrophy, 5 pericarditis, and arrhythmias. Pulmonary causes include obstructive (e.g., chronic obstructive pulmonary disease (COPD) and asthma) and restrictive processes (e.g., extrapulmonary causes. such as obesity, spine or chest wall deformities, and intrinsic pulmonary pathology such as interstitial fibrosis, pneumoconiosis, granulomatous disease or collagen 10 vascular disease). Mixed cardiac and pulmonary disorders include COPD with pulmonary hypertension and cor pulmonale, deconditioning, pulmonary emboli, and trauma. Noncardiac or nonpulmonary disorders include metabolic conditions such as anemia, diabetic ketoacidosis and other, less common causes of metabolic acidosis, 15 pain in the chest wall or elsewhere in the body, and neuromuscular disorders such as multiple sclerosis and muscular dystrophy. Obstmctive rhinolaryngeal problems include nasal obstruction due to polyps or septal deviation, enlarged tonsils, and supraglottic or subglottic airway stricture. Dyspnea can also present as a somatic manifestation of psychiatric disorders, 20 e.g., an anxiety disorder, with resultant hyperventilation. Additional information regarding the evaluation and treatment of dyspnea can be found, e.g., in Morgan and Hodge, Am. Fam. Phys. 57(4):711-718 (1998). Special Populations Certain populations of subjects may benefit particularly from the methods 25 described herein. These subjects include people for whom BNP or NT-proBNP is less useful, such as in those with impaired renal function (Anwaruddin et al., J. Am. Coll. Cardiol. 47(1):91-7 (2006); McCullough et al., Am. J. Kidney Dis. 41(3):571-9 (2003)), or in those who are overweight (Body Mass Index (BMI) of 25-29) or obese (BMI> 30) (Krauser et al., Am. Heart J. 149(4):744-50 (2005); McCord et al., Arch. 30 Intern. Med. 164(20):2247-52 (2004)). It is known and accepted in the field that patients with a high BMI usually have levels of natriuretic peptide that are lower than expected relative to a normal body mass patient for the same level of disease; the exact mechanism for this phenomenon is not known. It has been shown that 21 WO 2007/127749 PCT/US2007/067333 circulating levels of ST2 are not influenced by BMI, therefore, the determination of ST2 levels is more useful than natriuretic peptide levels in subjects with high BMI. Thus, the methods described herein can include determining a subject's BMI, and if the subject is overweight or obese, selecting the patient for determination of ST2 5 and/or IL-33 levels, as described herein. EXAMPLES The invention is further described in the following examples, which do not limit the scope of the invention described in the claims. Example 1: Sandwich ELISA Assay 10 This example uses the ST2 ELISA Kit manufactured by Medical & Biological Laboratories Co., Ltd. (MBL International Corp., Woburn, MA), no. 7638. This kit is a sandwich ELISA assay utilizing monoclonal antibodies for both capture and detection. This procedure is intended to analyze a full plate of samples assayed in replicates at a 1:3 dilution factor and closely follows the manufacturers' protocol. 15 Kits should be stored at 4"C until use. The procedure described in this example is optimized for human serum or plasma collected in citrate or EDTA anticoagulant tubes. Plasma collected in heparin anticoagulant tubes should not be used in this assay as heparin binds ST2 and inhibits measurement by this ELISA protocol. Plasma or serum samples may be used fresh or stored frozen. This assay is not adversely 20 effected by up to 3 freeze and thaw cycles of plasma samples. Reagents should be prepared fresh from a new kit immediately before performing the assays. Allow the kit to equilibrate to room temperature prior to use. Reagents not explicitly discussed below are provided by the manufacturer ready to use. 25 1. Wash solution - wash solution is provided by the manufacturer as a lOX concentrate solution. To make 1 liter of wash solution dilute 100 ml of the loX concentrate provided with 900 ml of distilled water. 2. Detector solution - the detector solution is prepared by diluting the detector concentrate 1:101 with the detector diluent. For a full 96 well plate of samples 30 10 ml of detector solution is required. To prepare 10 ml of detector solution use a pipette to transfer 10 ml of the blue colored detector diluent to a 15 ml 22 WO 2007/127749 PCTIUS2007/067333 orange top polypropylene tube. Ad 100 pl of the detector concentrate to this volume of detector diluent. a. NOTE: this reagent should be prepared during the first assay incubation step. 5 3. Calibrator stock - reconstitute the calibrator protein by dissolving the lyophilized protein in the amount of distilled water defined by the manufacturer for this manufacturing lot to yield a stock solution of 8 ng/ml. This volume specification is included in the product insert. Preparation of standards and samples: 10 . All of the following should be prepared in labeled 1.5 ml polypropylene tubes to be transferred to the assay plate with the P200 pipetter. Standards: The standard curve is prepared by making 2 fold serial dilutions of the 8 ng/ml stock solution. 15 1. Using a P1000 pipette transfer 250 g1 of Assay Diluent to 8 1.5 ml polypropylene tubes labeled S I-S8. 2. Using the same P1000 pipette transfer 250 pl of the 8 ng/ml Calibrator stock solution to tube S 1. This tube is now 4 ng/ml calibrator protein. a. Mix thoroughly by gently pipetting 3 times being careful not to create 20 bubbles. 3. Using the same P1000 pipette and a fresh tip for each of the following transfer 250 ptl of the reagent in tube S1 to tube S2, repeat the mixing. 4. Repeat step 3 for S2 to S3, S3 to S4, S4 to S5, S5 to S6 and S6 to S7. S8 will be the reagent blank so do not transfer the calibrant protein to this well. 25 a. Tubes Sl-S6 and S8 will now have 250 pl of reagent and tube S7 will have 450 pl Samples: The plate is set up so that each sample is analyzed as a 1:3 dilution in duplicate. 30 1. Label a 1.5 ml polypropylene tube for each sample. 2. Using the P200 pipette transfer 160 pl of Assay Diluent to each tube. 3. Using a P200 pipette transfer 80 pl of serum or plasma from sample I to tube 1. Mix carefully by pipetting 3 times without making bubbles. 23 WO 2007/127749 PCT/US2007/067333 4. Continue transferring samples to the sample tubes by repeating step 2 for each sample. Procedure: 1. Use the P200 pipette transfer the standards and diluted serum samples quickly 5 to the 96 well assay plate, as shown in Table 2. a. Set the P200 pipette for 100 pl b. Transfer 100 pl of the standard curve dilutions to each of columns 1 & 2 in the assay plate c. Transfer 100 pl of each of the serum samples to the assay plate in 10 exactly the same positions as shown in the plate map below. 2. Cover the assay plate with the provided shield and incubate at room temperature for 60 minutes. 3. Using the plate autowasher wash the plate 4 times. 4. Detector: using the 8 channel multichannel pipette transfer 100 p 1 of the 15 detector solution to each well and incubate at room temperature for 60 minutes. a. NOTE: this reagent was to be prepared during the first incubation step. b. NOTE: use a disposable reagent vessel for this reagent addition. ALWAYS use a fresh disposable reagent vessel for each reagent. It is 20 not necessary to change pipette tips during this step. 5. Wash the plate as in step 3 6. Substrate: using the 8 channel multichannel pipette transfer 100 pl of the Substrate to each well and incubate at room temperature for 30 minutes. a. The Substrate reagent is provided ready to use by the manufacturer. 25 7. Stop: at the completion of the Substrate incubation using the 8 channel multichannel pipette transfer 100 pl of the Stop solution to each well. a. The Stop Solution reagent is provided ready to use by the manufacturer. 8. Read the plate at 450 mn with background correction at 620 nm. 30 a. The plate should be read within 30 minutes after stopping the reaction. 9. Enter the absorbance readings in the provided spreadsheet for analysis. 24 WO 2007/127749 PCT/US2007/067333 Table 2: Map of Exemplary 96 Well Assay Plate 1 2 3 4 5 6 7 8 9 10 11 12 A 4.0 1 1 9 9 17 17 25 25 33 33 B 2.0 2 2 10 10 18 18 26 26 34 34 C 1.0 3 3 11 11 19 19 27 27 35 35 D 0.5 4 4 12 12 20 20 28 28 36 36 E 0.25 5 5 13 13 21 21 29 29 37 37 F 0.125 6 6 14 14 22 22 30 30 38 38 G 0.0625 7 7 15 15 23 23 31 31 39 39 H 0.0 8 8 16 16 24 24 32 32 40 40 Table 2 is a map of an exemplary 96 well assay plate, with control reactions in column 1, and each sample 1-40 analyzed in duplicate in columns 3-12. 5 Example 2. Measurement of Soluble ST2 Concentrations for the Evaluation of Patients with Acute Dvspnea In this example, the utility of ST2 measurement for evaluation of dyspneic patients was assessed. The subjects used in this Example took part in the ProBNP Investigation of 10 Dyspnea in the Emergency Department (PRIDE) Study, a prospective, blinded study of 599 dyspneic subjects presenting to the ED of the Massachusetts General Hospital, and was performed for the purpose of validation of the diagnostic and prognostic use of NT-proBNP testing. The results of the PRIDE study were recently reported (Januzzi et al., Am. J. Cardiol. 95(8):948-54 (2005)). 15 The gold standard for the diagnosis of acute HF was based on the impression of reviewing physicians, blinded to NT-proBNP values, who had all available information from presentation through the 60-days of follow-up; for the few patients in whom a diagnosis was uncertain, the reviewers were instructed to utilize the guidelines as reported by the Framingham Heart Study (McKee et al., N. Engl. J. 20 Med. 285(26):1441-6 (1971)). As reported, 209 subjects (35%) in the PRIDE study were adjudicated to have dyspnea due to acute destabilized HF, of whom 17 had mild (Class II) symptoms by the New York Heart Association (NYHA) classification, 80 had moderate (Class III) symptoms, and 112 had severe (Class IV symptoms). Of those without acute HF, the 25 most common diagnoses were exacerbation of obstructive airways disease (n = 150; 25 WO 2007/127749 PCT/US2007/067333 includes exacerbation of chronic obstructive pulmonary disease (n = 120) and asthma (n = 30) as well as acute pneumonia (n = 64). At the end of one year, the managing physician for each patient was contacted to ascertain the vital status of the patient. As reported, follow-up at one year was 5 complete in 597 subjects overall (Januzzi et al., Arch Intern Med 2006;166(3):315 20). NT-proBNP was measured in the PRIDE study using a commercially available immunoassay (ELECSYS* ProBNP assay, Roche Diagnostics, Indianapolis, IN), using established methodology. In the PRIDE study, the assay had inter-run 10 coefficients of variation of 0.9%. Blood collected at the time of presentation was later analyzed for concentrations of ST2, using an enzyme-linked immunosorbent assay (Medical & Biological Laboratories Co., Ltd.), as described in Example I herein. This assay utilizes monoclonal antibodies to human ST2 for both capture and detection, and had a relative percent difference of 17.5% in the present analysis. The 15 plasma used for the present study had been previously subjected to a single freeze thaw cycle. Comparisons Between Groups Comparisons of clinical characteristics between patients were performed utilizing chi-square tests for categorical data and the Wilcoxon rank-sum test for 20 continuous data. Comparisons of ST2 concentrations between diagnostic, New York Heart Association (NYHA) symptom classes and outcome categories were performed using non-parametric testing. Correlations ST2 and NT-proBNP results were log-transformed to establish normal 25 distribution. Correlations between these log-transformed variables were evaluated with the Spearman correlation coefficient. There was a modest correlation between concentrations of log-transformed ST2 and log-NT-proBNP in all subjects (r = 0.58, p <0.001), those without acute HF (r = 0.47, p <0.001) and those with acute HF (r = 0.40, p <0.00 1). 30 Cut-point analyses Patient characteristics as a function of an ST2 concentration of above or below 0.20 ng/ml are detailed in Table 3, which demonstrates the expected prevalence of factors consistent with a diagnosis of incident HF. 26 WO 2007/127749 PCT/US2007/067333 Table 3: Characteristics of Study Subjects as a Function of ST2 Concentrations Characteristie ST2 : 0.2 ng/mI ST2<0O.2 P valute Age (mean SD), years 67.7± 15.0 56.5 ± 17.5 <0.001 Past medical history Prior cardiomyopathy 13% 7% 0.02 Prior congestive heart failure 37% 12% <0.001 Arrhythmia 20% 13% 0.02 Hypertension 54% 43% 0.009 Diabetes mellitus 35% 16% <0.001 Coronary artery disease 33% 22% 0.005 Myocardial infarction 15% 11% NS Obstructive airway disease 15% 8% 0.005 Symptoms/signs Paroxysmal nocturnal dyspnea 16% 8% 0.002 Orthopnea 22% 12% 0.001 Lower extremity edema 26% 7% <0.001 Chest pain 35% 52% <0.001 Dyspnea at rest 50% 24% <0.001 Medications at presentation Beta blocker 44% 32% 0.002 Loop diuretic 41% 17% <0.001 Digoxin 13% 8% 0.04 Angiotensin converting enzyme 25% 16% 0.009 inhibitor Physical examination Body-mass index (Kg/m 2 , mean±SD) 28.0 ± 7.0 28.5 ± 6.5 NS Pulse, beats per minute (mean±SD) 91.7 ± 23.8 82.9 ± 20.6 <0.001 Jugular venous distension 13% 4% <0.001 Murmur 14% 8% 0.009 Hepatojugular reflux 3% 0% 0.002 Lower extremity edema 34% 14% <0.001 Rales 35% 16% <0.001 Electrocardiographic findings Atrial fibrillation 16% 9% 0.009 Chest radiographic findings Interstitial edema 25% 8% <0.001 Pleural effusion 28% 5% <0.001 Cephalization of vessels 2% 0% 0.04 Laboratory findings Serum creatinine, mg/dl, mean SD 1.2 ± 0.5 0.98 0.3 <0.001 Creatinine clearance, ml/min/1.73m 2 , 67.3 ± 32.0 85.2 61.0 <0.001 mean ± SD Blood urea nitrogen, mg/dl, mean ± SD 25.5 ± 17.0 17.2 10.0 <0.001 Troponin T, ng/dl, mean ± SD 0.063 ±0.32 0.022 0.16 0.04 NT-proBNP, pg/ml, median (IQR) 1800 (365-6745) 97 (40477) <0.001 27 WO 2007/127749 PCT/US2007/067333 These results indicate that ST2 is not generally correlated with BMI, but is associated with a number of other indices including serum creatinine levels and clearance, prior congestive heart failure, and diabetes mellitus. Example 3. Measurement of Soluble ST2 Concentrations for the Determination 5 of Risk of Death in Patients with Acute Dyspnea Factors predictive of mortality within one year following presentation with dyspnea were evaluated in the population described in Example 2. Candidate ST2 diagnostic cut points were evaluated with the use of bootstrapping techniques using the STATA SWBOOT program; this was followed by multivariable logistic 10 regression analyses. Each of the estimation procedures was coded in programming language then subjected-to the STATA bootstrap prefix command for 10 bootstrap repeated random samples, followed by 100 replications for those variables selected in the initial analyses. The bootstrap sample size was 593 (the size of the entire data set). Factors entered into the analysis included elements from past and present 15 medical history, symptoms and signs, medication use, as well as results of diagnostic studies including radiographic studies, electrocardiography, hematology, and blood chemistries. Measures of renal function included serum creatinine results as well as estimated glomerular filtration rate (Levey et al., Ann. Intern. Med. 130(6):461-70 (1999)). Following, candidate predictors of mortality (i.e., those with greater than 70 20 selections in bootstrap replications) were entered into multivariable logistic regression analyses, with acute HF as the dependent variable. For each logistic regression, results were entered in a single forward step with tail-probability to enter set at p = 0.01 and to remove the effect from the regression at p = 0.02, and goodness-of-fit was evaluated using the Hosmer-Lemeshow test. 25 In mortality analyses including NT-proBNP, this variable was modeled dichotomously, with a threshold value of 986 pg/ml for predicting death within one year, as identified previously (Januzzi et al., 2006, supra). After the SWBOOT run, the resulting validated candidate independent variables were entered stepwise into a Cox Proportional Hazards model; the proportions for this model were checked and 30 were found to be appropriate. Hazard ratios (HR) with 95% CI (confidence interval) were generated for each independent predictor of death by one year. Kaplan-Meier survival curves were constructed to compare mortality rates within one year in groups divided as a function of ST2 concentrations (as well as 28 WO 2007/127749 PCT/US2007/067333 diagnosis), using the log-rank test to compare the significance of the rates of mortality. For all statistical analyses, either SPSS (Chicago, IL, USA) or STATA (College Station, TX) software was used; all p values are two-sided, with composite 5 results <0.05 considered significant. Within one year, 93 subjects (15.7%) had died. The median concentrations of ST2 were significantly higher among decedents (1.03 ng/ml, IQR (interquartile range) =0.38 -2.43) than survivors (0.18 ng/ml, IQR =0.08-0.51; p <0.001). This pattern of higher ST2 concentrations in decedents remained when subjects were considered as 10 a function of the absence of acute HF (1.14 vs 0.13 ng/ml; p <0.001), as well as in those with acute HF (0.90 vs 0.45 ng/ml; p <0.001). Receiver operating characteristic (ROC) curve analysis using Analyse-It software (Analyse-It, Ltd, Leeds, UK) was performed for ST2, using the gold standard diagnosis of HF or survival within one year as the reference standard, and 15 area under the curve (AUC) was estimated; following, potential cut points for both diagnosis and prognosis were identified, and their sensitivity, specificity, as well as positive and negative predictive values (PPV, NPV) were estimated. To further evaluate the utility of optimal ROC-optimal cut-points, patients were also divided into deciles based on their ST2 concentrations, and evaluated for ST2 threshold effects on 20 the frequency of mortality. ROC analyses demonstrated an AUC of 0.80 (95% CI= 0.75 - 0.84, p<0.001) for ST2 and one year mortality (Figure 1); the optimal cut point as identified by ROC was 0.29 ng/ml, which was 87% sensitive (95% CI = 79 - 93%), 63% specific (95% CI = 58 - 67%), and had a PPV of 30% and NPV of 96%. In docile analyses, a 25 threshold effect for mortality was noted at the ST2 median of 0.20 ng/ml (Figure 2). A graded relationship between ST2 concentrations and the likelihood for death was also found, such that those subjects below the median concentration of ST2 (n = 236) enjoyed the lowest rates of death (2%) compared to those in the highest two deciles (n = 117) who demonstrated a 42% rate of death within one year, a HR of 43.0 (95% CI 30 = 15.0 - 123.0, p<0.001). In a final bootstrap model for prediction of death within one year (Table 4), an ST2 concentration > 0.20 ng/ml was selected in 96 of 100 replications and represented the strongest predictor of death within one year in breathless patients (HR = 5.6, 95% 29 WO 2007/127749 PCTUS2007/067333 CI = 2.2-14.2; p <0.001). Notably, even with the inclusion of NT-proBNP into the models, an ST2 >0.20 ng/ml remained the most selected variable in bootstrap replications (86 of 100 selections), and remained the strongest predictor of death within one year (HR=4.6, 95% CI= 1.8-11.8; p = 0.002). 30 WO 2007/127749 PCT/US2007/067333 Table 4: Identification of Independent Predictors of Death Within One Year in Dyspneic Subjects Bootstrap Analysis Multivaiiable analysis Variable Number selected out of Hazard 95% confidence P value 100 replications ratio intervals Log-transformed 99 2.7 1.6-4.4 <0.001 ST2 ST2 2 0.20 ng/ml 96 5.6 2.2-14.2 <0.001 Hemoglobin 73 0.91 0.85-0.98 0.008 Pleural effusion on 95 1.8 1.2-2.8 0.005 chest X-ray With NT-proBNP i. th noe Bootstrap Analysis Mlidiiabhb1 al6iin" Variable Number selected out of Hazard 95% confidence P value 100 replications ratio intervals ST2 2 0.20ng/ml 86 4.6 1.8-11.8 0.002 NT-proBNP 986 85 2.3 1.3-4.0 0.002 pg/ml Cough 80 0.60 0.39-0.95 0.03 Pleural effusion on 80 1.6 1.0-2.4 0.05 chest X-ray Log-transformed 77 2.6 1.6-4.4 <0.001 ST2 __ 5 The results of Table 4 are based on 100 bootstrap replications, followed by multivariable Cox Proportional Hazards analysis are shown in the absence and presence of NT-proBNP results in the model. Only those variables selected > 70 of 100 times in bootstrap replications are shown. 10 Kaplan-Meier hazard curves demonstrate that among subjects with ST2 concentrations 2 0.20 ng/ml, rates of death rose rapidly from enrollment, and continued to rise within one year (Figure 3A; Log rank p value <0.001). Similar relationships between ST2 values 0.20 ng/ml in those without and with the diagnosis of acute HF at presentation (Figure 3B; Log-rank p value for both <0.001). 15 Therefore, ST2 levels are an excellent predictor of mortality. Example 4. Measurement of Soluble ST2 and NT-proBNP Concentrations for the Determination of Risk of Death in Patients with Acute Dyspnea As ST2 and NT-proBNP were both independent predictors of death within one year, crude rates of death in subjects as a function of ST2 and NT-proBNP 31 WO 2007/127749 PCT/US2007/067333 concentrations were examined using the methods described above. The percentage of subjects who died in each category are shown in Table 5. Table 5: Mortality Rates as a Function of NT-proBNP 5 and ST2 Concentrations in the PRIDE Study NT-prBNP <986 pgml 1.7% 13.1% NT-prBNP> 986 pml 4.5% 37.0% The majority of mortality among subjects in the PRIDE study occurred in those with elevated levels of ST2 (equal to or above 0.20 ng/ml) and NT-proBNP 10 (equal to or above 986 pg/ml) (Table 5 and Figure 4). Thus, the combination of NT-proBNP and ST2 levels is useful in predicting risk of mortality. Example 5. Elevated ST2 Concentrations in Patients in the Absence of MI ST2 concentrations were determined as described in Example 1, above, in a 15 population of 350 patients who presented to the ED with chest pain. Serum samples were obtained and ST2 measurements made at baseline, and 90 and 180 minutes later for most patients. Also, for most patients, the baseline sample was collected within 2 hours of onset of symptoms. 17 patients had final diagnosis of MI, and 5 of these had ST2 ' 0.23 (0.25 20 0.65). Two of these patients were troponin negative. 11 patients had very high ST2 levels (0.97 - 9.22), but none of these patients had confirmed final diagnosis of MI, though all had severe diseases, including COPD, lymphoma, sepsis, alcohol abuse, and pulmonary embolism. The diagnoses for these 11 patients are shown in Table 6; ST2 I is the baseline level, ST2 2 is 90 minutes later, and ST2 3 is at 180 minutes. 25 Table 6: Non-MI Patients with High ST2 Levels Patient ST2 1 ST2 2 ST2 3 Final Diagnosis (gm)(ng/m) (ng/ml) 811 1.43 1.62 1.63 COPD with heart failure following coronary artery bypass graft surgery and pulmonary hypertension 847 2.37 4.44 3.53 Pulmonary embolism 873 2.36 2.42 2.74 RAD 898 1.32 1.24 1.66 History of heart failure following coronary artery bypass surgery 920 6.03 9.22 Bacteremia sepsis 32 WO 2007/127749 PCT/US2007/067333 928 3.80 4.69 13.99 H ypetension and alcohol abuse 952 6.76 1 Alcohol abuse, gastritis and pulmonary hypertension 953 3.77 History of heart failure following coronary artery bypass surgery 1055 1.42 1.28 1.13 URI 1213 0.97 1.19 1.07 Pulmonary embolism and pericarditis 1245 4.11 6.46 Lymphoma and hypertension 1280 1.30 1.33 COPD These results demonstrate that elevated ST2 is associated with severe disease, regardless of the underlying pathology. Example 6. Serial Analysis of ST2 Concentrations in Patients Hospitalized With 5 Acute Decomnensated Heart Failure (ADHF) ST2 concentrations were determined as described in Example 1, above, as were BUN, NT-Pro-BNP, and BNP, in a population of 150 subjects who were diagnosed with and admitted for acute decompensated heart failure (ADHF) to the San Diego Veterans' Administration Hospital. Some patients had a new diagnosis of 10 ADHF and other patients had an acute exacerbation of existing heart failure. Samples were taken on successive days in a number of patients, though not every patient gave a sample every day. Length of stay (LOS) in the hospital for these patients ranged from 1 to 24 days with an average of 5 days. Table 7 shows the characteristics of the population; as expected, given that the population was drawn from the San Diego 15 Veteran's Hospital, the population was overwhelmingly male and Caucasian. 33 WO 2007/127749 PCTUS2007/067333 Table 7: Frequency of Patient Characteristics Frequency Percent Gender Male 148 98.7 Female 2 1.3 Total 150 100 Ethnicity White 115 76.6 Black 22 14.7 Hispanic 9 6 Asian 3 2 other 1 0.7 Total 150 100 Diagnosis CHF new 14 9.3 CHF worse 110 73.3 Other 26 17.3 NYHA* Class II 9 6 Class III 75 50 Class IV 66 44 Etiology Unknown 61 40.7 Ischemic 43 28.7 Hypertensive 23 15.3 Other 9 6.1 Alcohol 5 3.3 Idiopathic 3 2 Valve 1I. Drug 1 0.7 *NYHA = New York Heart Association Stages of Heart Failure: Class I: No limitation of physical activity. Ordinary physical activity does not cause undue fatigue, palpitation, or dyspnea (shortness of 5 breath). Class II: Slight limitation of physical activity. Comfortable at rest, but ordinary physical activity results in fatigue, palpitation, or dyspnea. Class III: Marked limitation of physical activity. Comfortable at rest, but less than ordinary activity causes fatigue, palpitation, or dyspnea. 10 Class IV: Unable to carry out any physical activity without discomfort. Symptoms of cardiac insufficiency at rest. If any physical activity is undertaken, discomfort is increased. The population of patients was followed for at least 90 days, and adverse events were tabulated. A summary of events is shown in Table 8. 34 WO 2007/127749 PCT/US2007/067333 Table 8: Summary of Events Frequency Percent In hospital mortality 8 5.3 30 day readmission 13 8.7 30 day mortality CHF 10 6.7 30 day mortality other 7 4.7 90 day readmission 26 17.3 90 day mortality CHF 26 17.3 90 day mortality other 9 6.0 *90 day outcome frequency is cumulative of all events that occurred at earlier time points. Receiver operating characteristics and area under the curve (AUC) were 5 determined as described herein for the correlation of levels of BNP, NT-ProBNP, and ST2 at admission and at discharge with the adverse events listed in Table 8. The results are shown in Table 9. AUC for ST2 change from first to last = 0.820. Table 9: ROC SUMMARY (AUC) In Hospital 30 day 30 day 90 day 90 day mortality event mortality event mortality BNP Admit 0.735 0.536 0.630 0.569 0.653 Discarge NA 0.545 0.648 0.575 0.692 NT-proBNP 0.790 0.638 0.785 0.624 0.763 Admit NT-proBNP NA 0.637 0.831 0.631 0.815 Discharge______ ST2 admit 0.638 0.549 0.574 0.558 0.603 Discharge 0.752 0.642 0.760 0.657 0.772 Change in NA 0.573 0.793 0.604 0.809 In this small and extremely homogeneous population, the biomarkers that ia most accurately predicted the reported events included change in ST2 for 30 and 90 day mortality, and NT-proBNP levels at discharge for both 30 and 90-day mortality. As expected, the homogeneity of this cohort resulted in the AUC using the admission measurement for each biomarker being lower than what was observed in the PRIDE cohort (described in examples 2-3). The high number of patients without diagnosed 15 heart failure reported in the PRIDE cohort resulted in a high discriminatory power of 35 WO 2007/127749 PCT/UJS2007/067333 ST2 at the admission measurement. In this specialized cohort, where every patient already had a diagnosis of heart failure, the ST2 level at admission is expected to be somewhat elevated; consequently, the predictive power for subsequent events is somewhat diminished as compared to a heterogenous cohort that is more 5 representative of the general population. A univariate model of predictors of 90 day mortality was analyzed including the same markers, as well as BUN and creatinine clearance, which are markers of renal function. The results are shown in Table 10. Table 10: Univariate Predictors of 90-day Mortality Predictor Odds Ratio P-Value Change in ST 2: 2 -0.02 vs. < -0.02 11.55 0.002 Admission ST 2: 2 0.2 vs. < 0.2 1.14 0.83 Discharge ST 2:> 0.2 vs. < 0.2 7.00 0.003 Admission BNP: 2 400 vs. <400 1.91 0.35 Discharge BNP: 400 vs.< 400 2.59 0.14 Admission NT-ProBNP: 2 5000 vs. <5000 3.96 0.09 Discharge NT-ProBNP: 5000 vs. <5000 6.72 0.007 Admission BUN continuous 1.033 0.002 Admission creatinine continuous 1.50 0.07 10 In this analysis, for the dichotomous variables (all but BUN and creatinine) the odds ratio represents the increased odds of death within 90 days for a positive test vs. a negative test. For the continuous variables (BUN and creatinine) the odds ratio represents the increased odds of death within 90 days for each 1 mg/dL increase. As shown in Table 10, change in ST2 provided the best indication of 90 day mortality, 15 with an OR of 11.55 (p = 0.002). A multivariate model was also constructed that included BNP, NT-ProBNP, BUN and creatinine clearance levels with change in ST2. The results are shown in Table 11. 36 WO 2007/127749 PCTIUS2007/067333 Table 11: Predictors of 90-day Mortality With Change in ST2 Additional Predictor in Change in ST 2 Additional Predictor Model Model Change inST_2_R_2 Odds Ratio P-Value Odds P-Value ________Ratio Admission 12.18 0.002 2.04 0.34 0.246 BNP____________ _ Discharge 13.98 0.001 3.56 0.07 0.285 BNP I____ Admison 14.49 0.001 5.48 0.044 0.312 Discharge NT- 10.04 0.005 5.49 0.021 0.325 ProBNP__ __ Admission 7.87 0.014 1.023 0.034 0.300 BUN ___ Admission 10.08 0.005 1.260 0.30 0.241 creatinine I______ I___________ When only change in ST2 (the difference between the ST2 levels) was considered, the R 2 for the model was 0.223. As can be seen in Table 11, in each case change in ST2 was the stronger predictor due to its having the smaller p-value. 5 Moreover, only admission NT-ProBNP, discharge NT-ProBNP, and admission BUN provided additional prediction beyond change in ST2 that achieved a statistical significance (p < 0.05) (see bold R 2 values), while discharge BNP achieved marginal statistical significance (p = 0.07). Next, the ROC was compared for change in BNP and change in ST2. As 1o shown in Figure 5, the AUC for change in BNP during hospitalization was 0.67, while the AUC for change in ST2 was 0.78, significantly higher, indicating that change in ST2 is a much more accurate predictor of 90-day mortality than BNP. The change in ST2 was analyzed in conjunction with BUN levels to determine whether the two markers, when used together provided additional predictive 15 information. The results, shown in Figure 6, demonstrate that BUN levels and changes in ST2 levels can be used to categorize patients by risk: patients with low levels of BUN (less than 40 mg/dL) and levels of ST2 that decreased by 15% or more had the lowest risk levels; patients with higher BUN and levels of ST2 that decreased by less than 15% (or increased or held steady) had the highest risk levels, while 20 subjects with either high BUN and ST2 decreased by at least 15%, or low BUN and 37 WO 2007/127749 PCTIUS2007/067333 ST2 decreased by 15% or more, had a medium level of risk (which was still much increased over that of the lowest risk category, see Figure 6). Changes in ST2 levels over time were evaluated on a lapsed-time basis to determine an optimal length of time between first and second ST2 measurements. 5 The results, shown in Tables 11A and 11B, indicate that a lapsed time of at least two days and within four days between the first and second measurement of ST2 provides the best information regarding likelihood of death within 90 days. Table 11A: Summary of ST2 Values over Time N average median 1 day 130 0.97 0.78 2 days 99 1.00 0.60 3 days 66 1.42 0.42 4 days 44 1.87 0.57 5 days 34 1.86 0.74 10 Table 11B: Correlation of ST2 Values with Mortality within 90 Days N> 0.85 # dead % death N < 0.85 # dead % death I day 54 7 13% 76 7 9% 2 days 27 7 26% 72 4 6% 3 days 19 4 21% 47 2 4% 4 days 18 6 33% 26 1 4% 5 days 14 6 43% 20 1 5% In Table 11 B, "N > 0.85" indicates the number of people who had a reduction of less than 15% in ST2 over the time period indicated (so that the second level is 85% or more of the first level). After at least two days, the percentage of people with 15 a ratio of 0.85 or higher who died was over 20% while the percentage of patients who died with a ratio below 0.85 initially decreased and remained low. When average ST2 values over time were tracked in this population, a clear increase in correlation of risk of mortality within 90 days was seen at three to four days after the initial measurement was made, as shown in Table 12A and Figures 7A 20 and 7B. Figure 7B, a whisker box plot bracketing the 2 5 h and 75*h percentiles, 38 WO 2007/127749 PCT/US2007/067333 illustrates that distinction between survivors and decedents is clearly resolved with the ST2 value. The ST2 results contrast with results generated by comparing either BNP or NT-proBNP over these same days of treatment. Although the mean values for each of these markers is distinct between survivors and decedents the groups are not 5 statistically distinct from each other, as shown in Figures 7C and 7D. Unlike ST2 the survivor and decedent groups for BNP and NT-proBNP overlap significantly between the 2 5 " and 75* percentiles. In this analysis ST2 is a more accurate marker than BNP or NT-proBNP alone for predicting mortality as a function of change over time during hospitalization. 10 When average ratios of ST2 were calculated, again, a significant difference was seen; the results are shown in Table 12B and Figure 8. Table 12A: Average ST2 Values Over Time Average ST2 value ST2-1 ST2-2 ST2-3 ST2-4 ST2-5 ST2-6 non-survivor 0.33 0.38 0.44 1.14 0.89 0.71 survivor 0.27 0.24 0.15 0.17 0.15 0.17 Table 12B: Change in ST2 Over Time Average Ratio of ST2 1 day 2 days 3 days 4 days 5 days non-survivor 1.04 2.44 9.57 8.43 6.46 survivor 0.96 0.82 0.60 0.63 0.67 15 It is also possible to use the ratio of ST2 along with the ratio of NT-proBNP in a 2 x 2 matrix analysis to further refine the risk stratification of these patients, as shown in Table 12C. Table 12C: Change in ST2 and Change in NT-proBNP for Risk 20 Stratification NT-proBNP and ST2 ratios NT < 0.7, NT <0.7, NT >0.7, NT > 0.7, ST2<0.85 ST2>0.85 ST2 <0.85 ST2>0.85 N 45 7 34 28 39 WO 2007/127749 PCT/US2007/067333 Mort 10 12 13 16 % Mort 0.0% 28.6% 8.8% 21.4% These results demonstrate that the change in ST2 over time is a powerful predictor of mortality, e.g., within 90 days, in both univariate or multivariate predictive analysis; change in ST2 was shown to be the strongest univariate predictor of patient outcome in this population, with optimal predictive value seen after 3 days 5 of hospitalization. Combining change in ST2 data with a measure of renal function (e.g., BUN) provided additional predictive value, as did including an NT-proBNP measurement with change in ST2. Example 7. ST2 Concentrations for Risk Stratification in Patients with Pulmonary Disease 10 Of the 599 subjects in the PRIDE population (described in Examples 2 and 3), 209 had acute decompensated heart failure (ADHF). Of the remaining 390 subjects, 236 had "pulmonary disease"; of those with available data, 5 had uncomplicated bronchitis; 18 had pulmonary embolism (PE); 64 had pneumonia; and 149 had COPD/asthma. Of the 149 with COPD/asthma, 69 had asthma; 67 had emphysema; 15 and 13 had chronic bronchitis. To evaluate the usefulness of ST2 as a marker of risk in these patients, ST2 levels were assessed in each diagnostic category. Concentrations for each diagnostic category are shown in Table 13. Table 13: ST2 Concentrations by Diagnosis Category ST2 median 25th-75th Percentile Whole group 0.23 0.09-0.69 PE 0.16 0.08-1.4 Pneumonia 0.69 0.14-2.29 Bronchitis 0.19 0.11-0.41 Asthma/COPD 0.20 0.08-0.54 Asthma 0.14 0.05-0.47 Emphysema 0.24 0.12-0.55 Chronic bronchitis 0.33 0.06-1.2 40 WO 2007/127749 PCT/US2007/067333 ST2 concentrations were evaluated using ROC analysis for death at one year, and concentrations in those dead and alive at one year were compared. Mortality rates as a function of ST2 concentrations to one year were analyzed. The results are shown in Table 14 and Figures 9-11. 5 Table 14: Mortality by Diagnosis Acute PE PNA Asthma/COPD bronchitis N 18 64 5 149 N survivors 15 52 5 139 0.14 (0.07- 0.53 (0.13- 0.17 (0.07 ST2 survivors See above 0.36) 1.81) 0.50) ST2 1.98 (1.2-3.83) 1.17 (.4-3.17) n/a 0.31 (.2-1.1) decedents These results demonstrate that, as shown in Figure 9, ST2 concentrations showed excellent specificity and sensitivity (Area under ROC = 0.73; 95% CI = 0.62 0.83; P < 0.0001). The optimal cut point was 0.20 ng/ml, with 88% sensitivity and 52% specificity (PPV = 22%; NPV = 96%). Figure 10 further illustrates this point. 10 The median ST2 concentrations in subjects who were still alive at one year was 0.19 ng/ml (IQR 0.08-0.59, n = 211); while median concentrations in subjects who had died within one year were 1.19 ng/ml (IQR 0.28-2.2, n = 25). Finally, as shown in Figure 11, mortality rate increased dramatically as a function of ST2 concentration, using a threshold of 0.2 ng/ml. 15 Multivariate Cox Proportional Hazards analyses were used to identify independent predictors of death at one year; the results are shown in table 15. 41 WO 2007/127749 PCT/US2007/067333 Table 15: Independent Predictors of Death in Pulmonary Disease Characteristic Hazard ratio 95% CI P value Age 1.01 0.98-1.04 0.71 Gender 1.36 0.60-3.07 0.74 ST2 20.20 ng/mI 6.14 1.80-21.0 0.004 Pleural effusion on 2.99 1.30-6.83 0.009 CXR Emphysema 0.29 0.13-0.65 0.003 Spironolactone 13.7 3.60-52.0 <0.001 These data demonstrate that ST2 levels are an excellent predictor of death within one year, regardless of the underlying pathology, in subjects presenting with dyspnea. 5 Example 8. ST2 Concentrations are Not Affected by Renal Insufficiency The effect of renal impairment on ST2 concentrations was evaluated in a population of 135 patients with moderate to severe renal insufficiency. None of the patients were on dialysis, and none were previously diagnosed with CVD. All of the patients were evaluated using glomerular filtration rate (GFR in mls/min) as 10 determined by the Modification of Diet in Renal Disease (MDRD) method as a measure of renal function. Echocardiography and coronary artery calcium (CAC) measurements were also performed on each subject to detect latent CVD. Multiple biomarkers were also evaluated. The descriptive statistics for this cohort are shown in Table 16; the mean GFR 15 and ST2 are illustrated graphically in Figures 12A-B. Table 16: Glomerular Filtration Rate (GFR) and ST2 Levels GFR ST2 levels (ng/ml) Mean 34.5 0.122 Median 34 0.107 Std Error 0.989 0.005 Std Dev. 11.4 0.059 Coeff. Var. 33.3 48.346 Lower 95%CL 32.5 0.112 Upper 95%CL 36.4 0.132 25th Percentile 27 0.090 75th Percentile 43 0.127 Minimum 9 0.068 Maximum 59 0.476 Count 135 135 42 WO 2007/127749 PCT/US2007/067333 In this cohort of patients with stable, chronic disease, only ten (8%) had ST2 levels above 0.2, the highest of which was 0.476 ng/ml. The distribution of ST2 values is shown in Figure 13. This was as expected in this population of subjects with chronic, managed renal insufficiency; one would not expect to see very high ST2 5 levels. Pearson Correlation analysis was performed in this population to determine whether there was a correlation between ST2 levels and GFR. The results are shown in Tables 17 and 18. Table 17: Pearson Correlation Results - GFR and ST2 Descriptive Statistics Variable Mean Std Dev. Std Err N GFR 34.5 11.5 0.989 135 ST2 (ng/mL) 0.122 0.059 0.005 135 Correlation Matrix ) GFR ST2 (ng/mL) GFR 1.000 0.028 ST2 (ng/mL) 0.028 1.000 Correlation Significance (P) GFR ST2 (ng/mL) GFR - 0.748 ST2 (ng/mL) 0.748 10 Table 18: Pearson Correlation Results - Creatinine Clearance and ST2 Descriptive Statistics Variable Mean Std Dev. Std Err N Screening Cr 2.175 0.859 0.081 113 ST2 n/mL) 0.122 0.058 0.006 113 Correlation Matrix Screening Cr ST2 (ng/mL) Screening Cr 1.000 1-0.018 ST2 (ng/mL) -0.018 11.000 Correlation Significance (P) Screening Cr ST2 (ng/mL) Screening Cr - 10.851 ST2 (n /mL) 0.851 43 WO 2007/127749 PCT/US2007/067333 These results demonstrate that, as was expected in this population of subjects with chronic, managed renal insufficiency, there is no correlation between ST2 levels and either GFR (p = 0.75) or creatinine clearance (p =0.851) in this population. This indicates that renal insufficiency, by itself does not cause an elevation of ST2 levels. 5 The same analyses were carried out in a population of 139 subjects at the San Diego Veteran's Administration Hospital. All of the subjects had previously been diagnosed with acute decompensated heart failure (ADHF), and the mean ST2 level was about twice that seen in the population of patients with chronic renal insufficiency but no HF (see tables 17-18). There is an almost ubiquitous correlation 10 between renal insufficiency and heart failure, with an almost 80% confluence of patients with stage III/W HF also having impaired renal function (Fonarow and Heywood, Am. J. Med. (2006) 119(12A):S17-S25. Thus, because ADHF is correlated with ST2 levels, one would expect to see a correlation between renal insufficiency (as measured by GFR) and ST2 levels. This was exactly what was seen, 15 as shown in Tables 19 and 20. Table 19: Pearson Correlation Results - GFR and ST2 in ADHF Descriptive Statistics Variable Mean Std Dev. Std Err N GFR 59.1 25.3 2.143 139 ST2 (ng/mL) 0.283 0.332 0.028 139 Correlation Matrix (R) GFR ST2 (ng/mL) GFR 1.000 -0.062 ST2 (ng/mL) -0.062 1.000 Correlation Significance (P) GFR ST2 (ng/mL) GFR - 10.470 ST2 (ng/mL) 0.470 44 WO 2007/127749 PCT/US2007/067333 Table 20: Pearson Correlation Results - GFR and ST2 Ratios in ADHF Descrip tive Statistics Variable Mean Std Dev. Std Err N GFR 59.1 25.3 2.143 139 ST2 ratio 1.038 3.038 0.258 139 Correlation Matrix @ GFR ST2 ratio GFR 1.000 -0.161 ST2 ratio -0.161 1.000 Correlation Significance (P) GFR ST2 ratio GFR - 0.058 , ST2 ratio 0.058 These results demonstrate that, in subjects with ADHF, ST2 values, whether represented as a single level or a ratio, are correlated with measures of renal insufficiency, but are independent of the renal insufficiency. 5 Example 9. Risk of Mortality Correlates with ST2 Concentrations The amount of increased risk of mortality was evaluated in the subjects who participated in the PRIDE study and the veterans in the population described in Example 8. Both 0.2 and 0.7ng/ml ST2 thresholds were evaluated across the entire 10 population (Table 21) and the just the ADHF subset (Table 22). The biggest difference between the two thresholds is that the PPV increases and the NPV decreases for the 0.7 ng/ml threshold in both patient sets. Table 21 - % Mortality by ST2 level in PRIDE Cohort ST2 0.7 ST2<0.7 ST22 0.2 ST2<0.2 Total N 146 447 317 275 Mortality N 52 41 88 5 % Mortality 35.6% 9.2% 27.8% 1.8% The NPV for 0.2 ng/ml is 98.2% and for 0.7 ng/nl is 90.8% in this population. 15 Table 22 - % Mortality by ST2 level in Subjects with ADHF ST2 2 0.7 ST2 <0.7 ST2 2 0.2 ST2 <0.2 Total N 81 127 167 41 Mortality N 30 26 55 1 % Mortality 37.0% 20.5% 32.9% 2.4% 45 The NPV for 0.2 ng/ml is 97.6% and for 0.7 ng/ml is 79.5% in this cohort. For the ST2 ratio from the VET study the results are shown in Table 23. 5 Table 23 -- % Mortality in Subjects with Chronic Renal Failure ST2 ratio > 0.85 <0.85 N .35 79 Mortality 8 3 % Mort 22.9% 3.8% The NPV for this measurement is 96.2%. OTHER EMBODIMENTS 10 It is to be understood that while the invention has been described in conjunction with the detailed description thereof, the foregoing description is intended to illustrate and not limit the scope of the invention, which is defined by the scope of the appended claims. Other aspects, advantages, and modifications are within the scope of the following claims. The reference in this specification to any prior publication (or information derived 15 from it), or to any matter which is known, is not, and should not be taken as an acknowledgment or admission or any form of suggestion that that prior publication (or information derived from it) or known matter forms part of the common general knowledge in the field of endeavour to which this specification relates. Throughout this specification and the claims which follow, unless the context 20 requires otherwise, the word "comprise", and variations such as "comprises" and "comprising", will be understood to imply the inclusion of a stated integer or step or group of integers or steps but not the exclusion of any other integer or step or group of integers or steps. 46

Claims (17)

1. 2. The method of claim 1, wherein the reference ratio is a threshold ratio.
2. 3. The method of claim I or claim 2, wherein the method comprises identifying a subject that has a ratio of the second level of soluble ST2 to the first level of soluble ST2 that is elevated as compared to a reference ratio as having an increased risk of death within one year, and selecting an aggressive treatment for the subject.
3. 4. The method of claim 1 or claim 2, wherein the method comprises identifying a subject that has a ratio of the second level of soluble ST2 to the first level of soluble ST2 that is decreased as compared to a reference ratio as having a decreased risk of death within one year, and selecting a conservative treatment for the subject. 47 H:s nm\]ncevn\NRPcrtb1DCC\SVf65D44_I.doex-151102015
4. 5. The method of any one of claims 1 to 3, wherein the aggressive treatment includes admission of the subject to a hospital, cardiac catheterization, and/or extensive imaging.
5. 6. The method of any one of claims 1, 2, and 4, wherein the conservative treatment is outpatient treatment.
6. 7. The method of any one of claims 1 to 6, wherein the subject is already being treated and the aggressive or conservative treatment is an alternative treatment.
7. 8. The method of any one of claims 1 to 7, wherein at least two days pass between the first time point and the second time point.
8. 9. The method of claim 8, wherein at least three days pass between the first time point and the second time point.
9. 10. The method of claim 9, wherein at least four days pass between the first time point and the second time point.
10. 11. The method of any one of claims 1 to 10, the method further comprising: determining a level of one or more additional biomarkers selected from the group consisting of: cardiac troponin, brain natriuretic peptide (BNP), proBNP, amino-terminal (NT) BNP, atrial natriuretic peptide (ANP), NT-proANP, proANP, C-reactive peptide, blood Urea Nitrogen (BUN), and D-dimers.
11. 12. The method of any one of claims 1 to 11, wherein the subject has a body mass index of 25-29.
12. 13. The method of any one of claims I to 11, wherein the subject has a body mass index of> 30. 48 H:\si\intem-oven\NRPertbl\DCOSVM\65734- I docx--I/100025
13. 14. The method of any one of claims I to 13, wherein the subject is hypertensive.
14. 15. The method of any one of claims I to 14, wherein the subject does not have a cardiovascular disorder.
15. 16. The method of any one of claims 1 to 15, wherein the biological samples in (a) and (b) are serum, plasma, or blood.
16. 17. The method of any one of claims I to 16, wherein the first level of soluble ST2 at the first time point and the second level of soluble ST2 at the second time point are determined using an immunoassay.
17. 18. The method of claim 17, wherein the immunoassay utilizes one or more monoclonal antibodies that binds specifically to human soluble ST2. 49