Revascularization for Patients with HFrEF: CABG and PCI and the Concept of Myocardial Viability

Revascularization for Patients with HFrEF: CABG and PCI and the Concept of Myocardial Viability 22nd Annual Heart Failure 2018: an Update on Therapy April 2018 Eric J. Velazquez, MD, FACP, FACC, FASE, FAHA Professor of Medicine Duke University Health System Duke Clinical Research Institute

Revascularization for HFrEF Representative Case What we knew then What we know now STICH Main Hypotheses Patient Selection What we still need to learn Role for PCI in HFrEF? Personal reflections All Rights Reserved, Duke Medicine 2007

Mr. S 12/2002-1/2003 A 57 yo M initially seen by vascular surgery due to non-healing foot ulcer requiring BKA Admitted with progressive dyspnea and LE edema to OSH and transferred to DUMC for further evaluation PMH PVD s/p R BKA, DM, HL, HTN PEX JVP ~12 cm H 2 0, b/l decreased breath sounds/rales, 1-2 pitting edema Labs Na 127, scr 1.2, Hct 40, Alb 3.0 EKG NSR, no Q waves or acute ST segment deviation/t wave abnormalities All Rights Reserved, Duke Medicine 2007

Mr. S Inpatient 1/2003 Aggressively diuresed and discharged on beta-blocker, ACEI, digoxin, furosemide, baby ASA and simvastatin Clinic 4/2003 Continued to complain of dyspnea and orthopnea and was grossly volume overloaded by PEX despite OMT Randomized to CABG + MED as part of the STICH protocol and underwent 3vCABG (i.e., LIMA-LAD, SVG->OMT1, SVG->L PDA) on 5/2003 with an uneventful post-operative course All Rights Reserved, Duke Medicine 2007

Mr. S 2004-Present Continues to follow in clinic every 6-12 months and remains NYHA functional class I on OMT Has not been admitted for HF since undergoing CABG in 2004, has not undergone repeat LHC/PCI, and has not had a TTE ordered since 2009 All Rights Reserved, Duke Medicine 2007

Revascularization for HFrEF Representative Case What we knew then What we know now STICH Main Hypotheses Patient Selection What we still need to learn Personal reflections All Rights Reserved, Duke Medicine 2007

Ischemic Cardiomyopathy Ischemic cardiomyopathy first described by Burch in 1970 myocardial dysfunction that results from occlusive coronary disease apparently, once damage has occurred to the extent of the ischemic cardiomyopathy described the disease is irreversible even with absolute bed rest Burch et al Am Heart J 1972 All Rights Reserved, Duke Medicine 2007

CASS RCT: Reduced LVEF Subset (EF 35-50%) Proportion Surviving 1-Vessel Disease 2-Vessel Disease 3-Vessel Disease 1.0.9.8.7.6.5.4 CABG CABG CABG No.3 Statins, Medicine ACEi, ARBs, Medicine MRAs, ICDs, Medicine CRTs.2.1 0 p=0.45 1.0 p=0.40 p=0.0094.93.86.79.86 Medical.82 Therapy in CASS:.72.75 8S 11M Limited ASA and Beta-Blockers 8S 9M 1S 4M 28S 26S 13S 35M 31M 15M 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 Year.90.81.88.65 42S 37S 23S 36M 29M 16M Passamani et al. NEJM 1985 All Rights Reserved, Duke Medicine 2007

CASS Registry: LVEF < 35% 631 patients excluded from CASS RCT After 3 years, limiting symptom was: Medical Surgical Angina 26% 0% Dyspnea 36% 30% Fatigue 18% 40%* After 3 years, survival was: In those presenting with angina 68% 84%* In those presenting with CHF 55% 55% Alderman et al. Circulation 1983 All Rights Reserved, Duke Medicine 2007

2002 Surgical Revascularization? HF is increasingly prevalent; CAD is the most common associated etiology Substantial improvements in GDMT Risks and benefits of cardiac surgery among patients with HFrEF and CAD compared to medical therapy never studied The role of symptoms and/or myocardial imaging to select revascularization candidates unclear All Rights Reserved, Duke Medicine 2007

Surgical Treatment of IsChemic Heart Failure Trial Surgical Revascularization Hypothesis In patients with HF, LVD and CAD amenable to surgical revascularization, CABG added to intensive MED will decrease all-cause mortality compared to MED alone. Surgical Ventricular Reconstruction Hypothesis In patients with HF, LVD and CAD amenable to CABG and SVR, SVR added to CABG with MED will decrease all cause mortality and cardiac hospitalization compared to CABG alone. Velazquez EJ et al. JTCVS 2007 All Rights Reserved, Duke Medicine 2007

Inclusion Criteria LVEF 0.35 within 3 months of trial entry CAD suitable for CABG MED eligible Absence of left main CAD as defined by an intraluminal stenosis of 50% Absence of CCS III angina or greater (angina markedly limiting ordinary activity) SVR eligible Dominant LV anterior akinesia/dyskinesia Experienced surgeon Velazquez EJ et al. JTCVS 2007

Major Exclusion Criteria Recent acute MI (within 30 days) judged to be an important cause of LVD Cardiogenic shock (within 72 hours of randomization) Plan for percutaneous intervention of CAD Aortic valve disease clearly indicating the need for aortic valve repair or replacement Velazquez EJ et al. JTCVS 2007

Revascularization Hypothesis Conduct 1,212 Subjects Randomized CABG + MED n = 610 Randomization MED n = 602 Age (median) 60 years; 12% women Prior MI 77%; Diabetes 39% Baseline NYHA II-IV 89% LVEF 28% and ESVI 78 ml/m 2 (median) Multi-vessel disease 74%; Proximal LAD 68% Velazquez EJ et al. New Engl J Med 2016

CABG Conduct Variable Randomized to CABG (N=610) CABG received no (%) 555 (91) Time to CABG, days Median (IQR) 10 (5, 16) Performed electively % 95 Arterial conduits 1, % 91 Total conduits 3, % 56 Velazquez EJ et al. New Engl J Med 2016

Medication Use Medication All Patients Randomized (N=1212) Baseline Last Follow-Up Aspirin or clopidogrel, % 86 85 ACE inhibitor or ARB, % 90 84 Beta-blocker, % 85 88 Statin, % 81 85 Loop Diuretics, % 65 72 K+ Sparing Diuretics, % 46 53 Velazquez EJ et al. New Engl J Med 2016

Study Conduct 1,212 Subjects Randomized CABG + MED n = 610 Randomization MED n = 602 n = 13 Withdrew or lost n = 12 Withdrew or lost Analyzed n = 610 (100%) Final Analysis Median Follow-up 9.8 yrs. Max. Follow-up 13.4 yrs. Analyzed n = 602 (100%) Velazquez EJ et al. New Engl J Med 2016

All-cause Mortality NNT = 14 MED CABG Velazquez EJ et al. New Engl J Med 2016

Cardiovascular Mortality NNT = 11 Velazquez EJ et al. New Engl J Med 2016

All-cause Mortality or Cardiovascular Hospitalization NNT = 10 Velazquez EJ et al. New Engl J Med 2016

Other Outcomes Outcomes CABG (N=610) MED (N=602) Hazard Ratio (95% CI) (CABG vs. MED P- value Death or heart failure hospitalization 404 (66.2%) 450 (74.8%) 0.81 (0.71, 0.93) 0.002 Death or all-cause hospitalization 506 (83.0%) 538 (89.4%) 0.81 (0.71, 0.91) 0.001 Death or revascularization (PCI or CABG) 388 (63.6%) 478 (79.4%) 0.63 (0.55, 0.73) <0.001 Death or non-fatal myocardial infarction 376 (61.6%) 409 (67.9%) 0.86(0.74, 0.98) 0.032 Death or non-fatal stroke 367 (60.2%) 406 (67.4%) 0.85 (0.74, 0.98) 0.032 Velazquez EJ et al. New Engl J Med 2016

KCCQ Overall Summary Score: % of Pts with Significant Change (>5) 100 CABG + MED MED Alone 80 P = 0.007 0.0004 0.002 0.03 60 63 55 69 67 66 58 57 58 40 20 0 4 months 12 months 24 months 36 months Mark D et al. Ann Intern Med 2015

All-cause Mortality as Treated NNT = 9 Velazquez EJ et al. N Engl J Med 2016

Outline: Revascularization in HFrEF What we knew then What we know now STICH Main Hypotheses Patient Selection What we still need to learn Role for PCI in HFrEF? Personal reflections All Rights Reserved, Duke Medicine 2007

Diabetes and Treatment: All-Cause Mortality and CV Hospitalization MacDonald M et al. Eur J HF 2015

Mortality According to Angina and Treatment Arm Jolicour M J et al. JACC 2015

Ischemia and All-Cause Mortality Panza et al. JACC 2013

Interaction Between Ischemia and Treatment All-Cause Mortality Panza et al. JACC 2013

Myocardial Viability and Mortality Mortality Rate 1.0 0.8 0.6 0.4 0.2 0.0 Without viability With viability Without Viability With Viability HR 95% CI P 0.64 0.48,0.86 0.003 0 1 2 3 4 5 6 Years from Randomization Variables Associated with Mortality Chi- Square Risk score 33.26 < 0.001 LV ejection fraction 114 99 85 80 63 36 Multivariable 16 0.91 0.339 487 432 409 371 294 188 102 p 24.80 < 0.001 LV EDVI 35.36 < 0.001 LV ESVI 33.90 < 0.001 Myocardial Viability 8.54 0.003 Univariate 8.81 0.003

Baseline Characteristics: Patients With and Without Myocardial Viability 100 Previous MI p<0.001 50 LVEF p<0.001 200 180 LVEDVI p<0.001 LVESVI p<0.001 Percent 80 60 40 20 Ejection Fraction (%) 40 30 20 10 LV Volume Index (ml / m 2 ) 160 140 120 100 80 60 40 20 0 0 0 With myocardial viability Without myocardial viability

Interaction Between Viability and Treatment: All-Cause Mortality Without Viability With Viability Subgroup N Deaths HR 95% CI Without Viability 114 58 0.70 0.41, 1.18 Interaction P value 0.528 With Viability 487 178 0.86 0.64, 1.15 0.25 0.5 1 2 CABG Better MED Better Bonow RO et al. N Engl J Med 2011

Interaction of Viability and Treatment Endpoint Events Treatment p value Mortality 236 As randomized 0.528 As treated 0.962 Mortality or CV hospitalization 422 As randomized 0.390 As treated 0.975 CV mortality 187 As randomized 0.697 As treated 0.261

Effect of CABG vs MED on LVEF at 4 months Velazquez EJ et al. AHA Scientific Sessions 2014

LVEF Recovery Not Related to CABG Survival LVEF Improved No Improvement Failure to improve LVEF is not related to post -CABG HF symptoms Angina Survival Samady H et al. Circulation 1999 All Rights Reserved, Duke Medicine 2007

CABG leads to earlier and greater benefit in those at higher risk A) 3v CAD B) EF <27% C) LVESVI >79 ml/m2 Panza J et al., JACC 2014

Higher risk leads to earlier and increasing benefit Panza J et al., JACC 2014

Cause of Death by Age Age Quartiles Q1 Q2 Q3 Q4 (Age 54 years) (n=330) (54<Age 60 years) (n=295) (60<Age 67 years) (n=279) (Age>67 years) (n=308) All cause, % 53.9 56.3 67.4 73.1 CV, % 42.2 45.1 47.0 45.5 Non-CV, % 5.8 5.8 14.7 21.1 Unknown, % 5.8 5.4 5.7 6.5 Petrie M Velazquez EJ. Circulation 2016

Effect of CABG vs. GDMT on Cardiovascular Mortality by Age Petrie M Velazquez EJ. Circulation 2016

What do we know about PCI in HFrEF? No RCTS of PCI in chronic HFrEF Very few patients with HFrEF in CABG vs PCI RCTs SYNTAX included ~5% with any HF and ~ 2% with an LVEF < 30% FREEDOM included 2.6% with LVEF < 40% Very few patients with HFrEF in RCTs of PCI vs. MED COURAGE excluded patients with LVEF < 30% All Rights Reserved, Duke Medicine 2007

CABG Compared to PCI with Stenting 3 VD All patients Patients with EF < 40% Patients with EF 40% Disease of nonproximal LAD artery Stenting group 2166 342 1824 CABG group 4946 1196 3750 Unadj. hazard ratio (95% CI) 0.89 (0.74-1.06) 0.61 (0.46-0.81) 0.94 (0.75-1.17) Adj. hazard ratio (95% CI) 0.74 (0.62-0.90) 0.64 (0.48-0.87) 0.76 (0.60-0.96) Disease of proximal LAD artery Stenting group 2165 399 1766 CABG group 20857 5597 15260 Unadjusted HR (95% CI) 0.67 (0.59-0.77) 0.55 (0.44-0.69) 0.64 90.53-0.76) Adjusted HR (95% CI) 0.64 (0.56-0.74) 0.68 (0.54-0.85) 0.60 (0.50-0.72) Hannan EL et al. N Engl J Med. 2005 All Rights Reserved, Duke Medicine 2007

PCI vs MED in HF N = 700 Inclusion Criteria Chronic heart failure LVEF<35% Extensive CAD (BCIS Jeopardy Score>6) At least 4 segments of viable myocardium Heart team does not recommend CABG Exclusion Criteria Acute HF CCS 3 HB<9 egfr<25 Primary Endpoint Death, MI and HF hospitalization All Rights Reserved, Duke Medicine 2007

Revascularization in Ischemic Cardiomyopathy 2018: Implications (so far) Extent of CAD should be assessed and GDMT optimized for all patients presenting with HFrEF CABG improves survival with less morbidity Viability, angina and ischemia status should not define candidacy for CABG in HFrEF Patients at higher risk due to the extent of CAD, LVSD and remodeling have a greater (earlier) reward with CABG CABG mechanisms of benefit are multifactorial PCI not well studied in HFrEF; RCT data needed All Rights Reserved, Duke Medicine 2007

Putting it Together Velazquez EJ, Bonow RO JACC 2015

STICH Investigators Zabrze, Poland