ST2 as a Cardiovascular Biomarker Lori B. Daniels, MD, MAS, FACC Professor of Medicine Director, Coronary Care Unit University of California, San Diego ST2 and IL-33: Cardioprotective ST2: member of the Interleukin-1 receptor family Exists in two main isoforms ST2L Circulating sst2 IL-33 binding to ST2L triggers cardioprotective effects. ** Cardioprotection ** Kakkar et al. Nat Rev Drug Discov 2008 Competitive Model of ST2/IL-33 Signaling sst2 acts as decoy receptor IL-33 can bind to sst2, reducing [IL-33] available to ST2L ST2 in Heart Failure Kakkar et al. Nat Rev Drug Discov 2008 1
ST2 is higher in patients with acute heart failure ST2 is Associated with Symptom Severity 3.00 2.50 P<0.001 200.0 NYHA Class Acute HF (PRIDE) 200.0 NYHA Class Chronic HF (PHFS) ST2 (ng/ml) 2.00 1.50 1.00 0.50 ST2 Concentration (ng/ml) 180.0 160.0 140.0 120.0 100.0 80.0 60.0 40.0 P<0.0001 ST2 Concentration (ng/ml) 180.0 160.0 140.0 120.0 100.0 80.0 60.0 40.0 P<0.0001 20.0 20.0 0.00 Not acute HF (n=385) Acute HF (n=208) - NYHA Class 1 NYHA Class 2 NYHA Class 3 NYHA Class 4 - NYHA Class I NYHA Class II NYHA Class III NYHA Class IV Januzzi et al, J Am Coll Cardiol, 2007 ST2 Predicts 30-Day Mortality after Acute Decompensation in HF (PRIDE Study) After acute event: ST2 Concentrations and 1-Yr Mortality in Acute HF: As ST2 levels increase, so does risk ST2 identifies patients at highest risk during first 30- days Prognostic value persists to 1 year and beyond. Kaplan-Meier Curve based on supplemental data provided by investigators from PRIDE study. Original study results published: Januzzi J et al. JACC 2007 One-Year mortality >50% for highest decile Rehman et al. Correlates of ST2 in Acute HF JACC Vol. 52, No. 18, 2008.1458 65 2
Mortality Associated with Elevated ST2: Multivariate PRIDE: 1 Yr Mortality Risk Stratification Dyspnea: ST2 Has Best Prognostic AUC In multivariable analysis (adjusted for age, vital signs, treatment at presentation, CRP and either NT-proBNP or BNP), an ST2 >median was independently predictive of death at one year * * All p-values <0.001 except where asterisked Rehman SU et al. Clinica Chimica Acta 2008 ST2 in Acute Decompensated HF Cohorts HR for risk of death at 1 year, with ST2 >35 ng/ml ST2 in Chronic, Ambulatory HF Cohorts HR for risk of death at 1 year, with ST2 >35 ng/ml Univariable Risk-Adjusted Univariable Risk-Adjusted Adjusted for age, sex, NYHA class, EF, GFR, diabetes, HTN, and smoking Adjusted for age, sex, NYHA class, EF, GFR, diabetes, HTN, and smoking Daniels LB, Future Cardiol 2014 Daniels LB, Future Cardiol 2014 3
Predicting Short-Term Adverse Events: ST2 vs Natriuretic Peptides Acute HF pooled studies HFpEF TIME-CHF ST2 similarly predictive in HFpEF and HFrEF PARAMOUNT ST2 gives an earlier and stronger signal for short-term adverse events than NPs. HFpEF vs HFrEF ST2 and 18m survival in HFpEF vs HFrEF 622 symptomatic HF patients PARAMOUNT: HFpEF Entresto vs Valsartan and short-term Δ in LA volume (6 to 9 months) ST2 equally predictive Sanders-van Wijk et al. EJHF 2015. TIME-CHF Study Zile et al. Circ HF 2016 ST2 <median predicts improved LA volume with LCZ696 4
Additive value of ST2 to NTproBNP in long term prognosis Cumulative hazard 0.8 0.6 0.4 0.2 Both sst2 and NT-proBNP elevated (n=276) Only sst2 elevated (n=95) Only NT-proBNP elevated (n=54) Neither elevated (n=168) P <.001 ST2 Levels: Monitoring and Response to Treatment 0.0 0 300 600 900 Days from enrollment 1200 1500 Short-Term Changes in ST2 are Associated With Long-Term Events in HF ST2 as a Function of Events in Chronic HF ST2 Ratio: Week 2/Baseline 1-year cardiovascular death, hospital admission for HF, or urgent cardiac transplantation Bayes-Genis A, et al., Rev Esp Cardiol. 2010 Baseline to Week 2 ST2 change strongly predictive of risk and additive to NT-proBNP Bayes-Genis A, et al., Rev Esp Cardiol. 2010 5
ST2 in Val-HeFT Val-HeFT (background): 5010 patients with NYHA class II-IV HF Randomized to valsartan or placebo Valsartan associated with risk of mortality or HF hospitalization. ST2 Substudy: Do changes in ST2 levels over time provide additional prognostic information? ST2 measured at baseline and at 4 months 4 groups: Low-low, low-high, high-low, highhigh Based on cut-point of 35 ng/ml Changes in ST2 in Val-HeFT: placebo vs valsartan Patients randomized to valsartan had less in ST2 levels over time Results from Val-HeFT Serial ST2 Measurements Predict Outcome: Basel ADHF Cohort 207 AHF patients in ED Samples at: presentation, Day 1, Day 2 ST2 Non-Responders Fully adjusted hazard ratios after correction for 20 clinical variables. Conclusion: Change in ST2 over time was independently and significantly related to subsequent risks of mortality, 1 st morbid event or HF hospitalization 72% had persistently low ST2 8% moved from high risk to low risk 20% had persistently high or transitioned to high ST2 ST2 Responders 30 6
Serial ST2 Measurements for In-Patient Monitoring (SDVA): Mortality Serial ST2 Measurements Categorize Responder Status N=35 ST2 Non-Responders N=115 Decrease 50% Decrease 25 49% Increase or decrease <25% Adjusted for ADHERE Risk Factors and BNP change. N= 150 ST2 Values Shown Using 1 st Generation Research Assay. Boisot et al. J Card Failure 2008 Basel ADHF cohort 32 ST2 Levels Predict Response to Treatment: Beta-blockers ST2 Predicts Response to Treatment: Aldosterone Blockade in STEMI Cumulative survival 1 2 3 4 BB treated ST2 responder BB treated ST2 non-responder untreated ST2 responder untreated ST2 non-responder Days 1 2 3 4 p=0.62 p=0.1 Risk can be attenuated! Patients identified by ST2 are at increased risk and should receive therapy! Eplerenone prevents adverse ventricular remodeling ST2 predicts which pts are most at risk AND which pts will benefit most from aldosterone blockade High and low ST2 separated at median. Eplerenone attenuates remodeling more in pts with higher baseline ST2. Basel ADHF cohort Weir AP, et al. J. Am. Coll. Cardiol. 2010;55;243-250. 7
Predicting Reverse Remodeling: ST2-R2 score Developed in 304 HF pts with EF <40% Validated in 569 pts from 3 international HF cohorts, EF <40%: Barcelona, TIME-CHF, PROTECT Followed 4 yrs for Δ EF, Δ LV size, mortality Dichotomized Variables Used for Score Construction OR Assigned Points Non-ischemic Etiology 6.33 5 No LBBB 5.56 4 ST2<48 ng/ml 2.96 3 Duration of HF <12 Months 2.13 2 Beta Blocker treatment 1.89 2 Baseline LVEF < 24% 1.52 1 Reverse Remodeling Percentage ST2-R2 Score: Probability of Reverse Remodeling using ST2 2x likelihood of reverse remodeling if score is 12-14 vs 9-11 4x : score 15-17 vs 9-11 Improved NRI by 14.8% Lupon et al. Int J Card, 2015 and 2016 Lupon et al. Int J Card, 2015; and 2016 Admission vs Discharge ST2 in Acute HF N=182 pts admitted with AHF; real life population Admission and d/c ST2 and NT-proBNP Admission vs Discharge ST2 Admission ST2 Level Discharge ST2 Level Discharge ST2 more predictive of outcomes Discharge ST2 more predictive of outcomes Llibre et al, Biomarker 2016. Llibre et al, Biomarker 2016. 8
Discharge ST2 and NTproBNP Add ST2 Add BNP vs +ST2 +BNP ST2 in Monitoring and Treatment Response Early ST2 changes identify which acute HF patients respond to inpatient treatment Discharge levels most predictive of outcomes Change in ST2 over time predicts mortality and morbidity in chronic HF Only 1 marker (+): Big gain in prognosis with +ST2 alone Only 1 vs Both Markers (+): Addition of +NTpro complementary primarily for composite outcome Llibre et al, Biomarker 2016. Predicts response to (and varies with) aldosterone blockade, beta-blockers, and diuretics Algorithm for Using ST2 in Chronic HF Summary ST2 is a powerful predictor of short-term mortality in acute & chronic HF and MI ST2 is useful in both HFpEF and HFrEF ST2 and natriuretic peptides are synergistic Changes in ST2 levels have important clinical implications Admission and D/C levels give different info Non-responders are at higher risk, and optimizing therapy can attenuate this 9