Author + information
- Clyde W. Yancy, MD, MSc∗ ()
- ↵∗Reprint requests and correspondence:
Dr. Clyde W. Yancy, Division of Cardiology, Northwestern University, Feinberg School of Medicine, 676 North St. Clair, Suite 600, Chicago, Illinois 60611.
The landmark STICH (Surgical Treatment of Ischemic Heart Failure) trial was a disappointment—a worthy hypothesis but difficult execution with unrewarding results (1). STICH posed 2 fundamental questions: is surgical revascularization for ischemic heart failure (HF) with reduced ejection fraction (EF) preferable to medical therapy; and does surgical reverse remodeling therapy change the natural history of ischemic heart failure compared to medical therapy? Although largely accepted as dictum, surgical revascularization for multivessel coronary artery disease and reduced EF emanated from databases addressing, at best, mildly reduced left ventricular ejection fraction (LVEF) or mild to moderate clinical HF. Our current threshold for reduced EF HF (i.e., LVEF < 0.40) was not a consideration in the earlier CASS (Coronary Artery Surgical Study) and European surgical studies (2). Thus, both of these questions were appropriate. To be clear, the predicate data, consistently used now to guide everyday practice, addressing the surgical treatment of ischemic heart disease did not address significantly reduced EF phenotypes and, beyond anecdotal and uncontrolled experiential data, did not provide an answer to address surgical geometric modification of the left ventricle. This evidence gap persisted for decades. The results that have since emerged from STICH, including the data addressing viability, provided direction for surgical reverse remodeling, (generally not indicated) but left us in a quandary regarding surgical treatment of reduced EF HF due to ischemic heart disease. The primary results of STICH revealed no differences in mortality between surgical and medical therapy for reduced EF HF due to ischemic heart disease (p = 0.12) and additionally provided no insight regarding the utility of viability testing to discriminate between surgical and medical therapy.
The recent American College of Cardiology Foundation/American Heart Association 2013 Heart Failure Clinical Practice Guidelines (3) highlight the uncertainty which remains: “CABG [coronary artery bypass graft] or medical therapy is reasonable to improve morbidity and cardiovascular mortality for patients with severe LV dysfunction (EF <35%), HF, and significant CAD [coronary artery disease] (COR IIa, LOE B). CABG may be considered with the intent of improving survival in patients with ischemic heart disease with severe LV systolic dysfunction (EF <35%), and operable coronary anatomy whether or not viable myocardium is present (COR IIb, LOE B).”
In short, clinical judgment remains the modus operandi when a physician is confronted with the patient who has reduced EF HF and ischemic heart disease.
The new data brought forward in this issue of the Journal by Panza et al. (4) and the STICH investigators are therefore illuminating and worthy of our most careful consideration. A solution may have now emerged from STICH.
Important provisos must be acknowledged before delving into these new data. First, this was a retrospective, nonprespecified analysis; second, these were subgroup analyses that, although thoughtfully derived, remain empirical in construct and importantly are subject to the power limitations of small sample size and confounding variables due to lack of randomization; and third, these findings have not been replicated or validated in any other dataset and may represent both the play of chance and the consequence of extensive data dredging.
However, we now discover that within the STICH trial, for those patients with 2 of the 3 generally agreed upon poor prognostic indicators, namely, multivessel coronary artery disease, reduced EF, and increased end-systolic volume, there is a 2-year survival advantage that is statistically (and likely clinically) important that eclipses the early postoperative mortality risk seen in this higher risk cohort (hazard ratio [HR]: 0.71, 95% confidence interval [CI]: 0.56 to 0.89; p = 0.004). Given the totality of evidence in contemporary literature, these data cannot easily be dismissed. The FREEDOM (Future Revascularization Evaluation in Patients with Diabetes Mellitus: Optimal Management of Multi-vessel Disease) trial would argue that those patients with diabetes, as a marker of a less-than-good prognosis, and coronary artery disease benefit from surgical revascularization (5). Additionally, SYNTAX (Synergy between PCI with TAXUS and Cardiac Surgery) study data demonstrated the superiority of surgical revascularization relative to percutaneous coronary intervention for multivessel or left main coronary artery disease (6). The resultant SYNTAX score now predicts outcomes related to anatomical characteristics. In the context then of these seminal clinical trial results supporting surgery for higher risk cohorts, data reported by Panza et al. (4) reach a higher tier of veracity. The data are even more provocative when one observes the dismal 2-year outcomes for the medically treated group with 2 or 3 poor prognostic factors (HR: 1.88; 95% CI: 1.45 to 2.43, p < 0.001). As well, the sensitivity analyses that included per protocol treatments and “as treated” analyses are concordant. However, when only 1 poor prognostic factor was present, the early operative mortality associated with bypass surgery was not overcome at 2 years (HR: 0.88; 95% CI: 0.59 to 1.31; p = 0.535). Thus, higher risk cohorts did better with surgery, and lower risk cohorts did better with medical therapy. An answer from STICH appears to have materialized; however, as intriguing as these data are, we must return to our reasons for pause.
These data must be carefully interpreted. The group of surgeons in STICH had a noteworthy CABG mortality rate of <5%. We also should recall that >85% of patients enrolled in STICH were outside the United States, with important questions recently raised regarding event rates and standards of care in non-U.S. populations. Additionally, although the analysis based on 3-vessel coronary artery disease is irrefutable, the cut points of an LVEF of 0.27 and an left ventricular end-systolic volume index (LVESVI) of 76.8 ml/m2 are debatable and represent retrospectively derived median values only. Why not an LVEF of 0.25 or <0.20 or an LVESVI of >70 ml/m2 or >90 ml/m2? The data presented indicate the higher representation of advanced HF (New York Heart Association functional classification III or IV) and atrial fibrillation in the higher risk group, but we do not know which of those important clinical characteristics predicted the observed response to surgical therapy. Many of us remain befuddled that objective evidence of ischemia did not predict a surgical advantage. These data are reported at follow-up of ≥2 years out to 56 months. We will need to await the results of the STICH ES (STICH Extension Study) to better understand longer term outcomes. Finally, in a contemporary reduced EF HF population, with a penetration of up to 50% receiving implantable cardioverter-defibrillator therapy and approximately 20% receiving cardiac resynchronization therapy/defibrillator therapy, would this survival advantage remain (7)? Indeed, there are many caveats to consider.
However, another randomized trial of reduced EF HF due to ischemic heart disease is not likely. The imperfection of these data may prevail for some time, so how do we move beyond these findings as “hypothesis generating” worthy of a prospective trial versus provocative new data sufficient to change clinical practice and accept this solution from STICH? We should not allow the cloak of uncertainty to define these data; we should be more definitive, if possible, with these less-than-definitive data. The authors make a brilliant point, “our findings should not be considered as a dogmatic postulate of specific values to be used when making decisions on individual patients…” (4). I agree. The solution now emerging from STICH is not dogma but reason. The severity of ischemic heart disease is probably important and surgical treatment for more advanced disease in an experienced center with careful deployment of evidence-based medical and device therapy likely reflect optimal therapy for reduced EF HF due to ischemic heart disease, but the primary results of STICH remain valid. For all comers, the decision to operate depends upon the patient’s anatomy, the comfort of the surgeon, and the context of best-possible guideline-directed medical therapy (3). However, for those patients with more advanced disease as typified by reasonable descriptors, a more deliberate revascularization approach, or at least a more purposeful discussion among all parties, especially including the patient, might be the pragmatic solution now provided by STICH.
The guidelines will not change based on these data, nor should they, but perhaps our practice will. Finally, a solution emerges from STICH.
↵∗ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
Dr. Yancy has reported that he has no relationships relevant to the contents of this paper to disclose.
- American College of Cardiology Foundation
- ↵(1983) Coronary artery surgery study (CASS): a randomized trial of coronary artery bypass surgery. Survival data. Circulation 68:939–950.
- Yancy C.W.,
- Jessup M.,
- Bozkurt B.,
- et al.
- Panza J.A.,
- Velazquez E.J.,
- She L.,
- et al.