The STICH Trial: Evidence-Based Conclusions

This editorial refers to ‘The STICH trial unravelled’ by G.D. Buckberg et al., published in this issue on page 1024–1027. Adverse left ventricular (LV) remodelling is one of the most important determinants of prognosis post-myocardial infarction (post-MI).1 The process begins with infarct expansion nearly immediately post-MI and is followed by progressive eccentric hypertrophy, cardiac fibrosis, and reduced contractility of the remaining cardiomyocytes and can result in significant LV distortion, impaired ventricular filling, reduced ventricular function, and mitral regurgitation.2 As LV dysfunction progresses and heart failure develops, global or focal myocardial ischaemia can occur, regardless of the severity of coronary artery disease (CAD), which in turn further worsens LV dysfunction.3 Indeed, in patients with severe heart failure, chronic global cardiac lactate production can occur, regardless of whether the patient has significant residual coronary artery obstruction or not.3 Perhaps, it is this enhanced susceptibility of patients with LV dysfunction to ongoing myocardial ischaemia that explains the remarkable benefits of coronary artery bypass grafting (CABG) in patients with LV dysfunction and advanced CAD.4 Several interventions have been shown to slow or reverse adverse LV remodelling and to improve patient outcome in patients with congestive heart failure. These include medications, such as beta-blockers and blockers of the renin–angiotensin–aldosterone system, and the implantation of cardiac resynchronization devices.5 Surgical ventricular reconstruction (SVR) has also been used as a way to rapidly and dramatically reverse adverse ventricular remodelling.6 The SVR procedure consists of surgically creating a smaller ventricle with a more normal shape via an endoventricular circuloplasty technique used to treat ventricular aneurysms and adapted by Dor et al.6 Initially, the SVR procedure was applied to patients with frank LV aneurysms, but, more recently, its use has been extended to patients with akinetic and/or dyskinetic anterior LV walls but without a frank aneurysm. Several series have documented that this procedure can be performed with reasonably low peri-operative mortality and that it results in good patient outcomes.7,8 Experience from these large series led to the development of empirical standards for patient selection and measures of surgical success for the SVR procedure.7,9 The optimal patient for SVR is thought to be a patient with an LV end-systolic volume index (LVESVI) ≥60 mL/m2, which is approximately greater than twice normal, akinesia/dyskinesia ≥35% of the anterior wall, and a reduced LV ejection fraction (≤35%). Although encouraging, the results from these series were never compared with CABG alone, and there exists no systematic evaluation of the long-term effects of adding SVR to CABG on LV remodelling and clinical outcome. Clinical follow-up from these registries is often quite incomplete, and the timing of the follow-up evaluation of the effects of SVR on ventricular remodelling is variable, often as early as 1 week after SVR surgery. The temporal relation of LV assessment post-SVR is critical to understanding the true impact of SVR on LV remodelling, as the early impact of SVR on LV size and shape may be attenuated by the progressive LV dilatation and distortion thought to occur post-SVR operation.10 The long-term effects of SVR on mitral regurgitation whether mitral valve repair is performed or not also remain to be determined. Finally, the long-term effects of SVR on LV filling and diastolic function appear to be of some concern, at least in some patients, and may contribute to worsening heart failure in a significant subgroup of patients.11 Diastolic dysfunction post-SVR has been shown to adversely affect prognosis.8 The Surgical Treatment for Ischemic Heart Failure (STICH) trial was designed to address two specific primary hypotheses in patients with clinical heart failure, LV dysfunction, and CAD amenable to surgical revascularization and was conducted as two parallel trials. Buckberg et al. address the portion of the STICH trial that has already reported and that randomized 1000 patients with ischaemic cardiomyopathy and dominant anterior LV dysfunction amenable to SVR, that underwent CABG alone or had CABG + SVR.12,13 The primary outcome was a composite of death from any cause or cardiac hospitalization with a median follow-up of 48 months. The addition of SVR to CABG had no effect on the primary outcome or either of its components. Cardiac symptoms and exercise tolerance improved from baseline to a similar degree in both study groups. Surgical ventricular reconstruction resulted in lower LV volumes 4-month post-operation. STICH is the first large randomized trial evaluating the long-term effects of SVR. In this issue of the journal,14 a group of surgeons, some of whom participated in the STICH study as investigators, raise a number of concerns regarding several aspects of the STICH trial. Some of their concerns have been addressed at recent major meetings in presentations that have focused on the subgroups of patients these authors believe are the most likely to benefit with SVR. The complete papers of these presentations have been submitted or are in preparation for submission for publication. Other concerns expressed by these authors appear not to be warranted. The authors begin by expressing concerns that the patients selected in STICH did not meet all of the optimal characteristics for the SVR procedure. They correctly point out that in the initial protocol, patients were to have an LVESVI of ≥60 mL/m2, akinesia/dyskinesia ≥35% of the anterior wall, and reduced LV ejection fraction (≤35%). However, due to challenges in recruiting, the STICH study needed to expand the number of centres participating in the study and, due to the empirical nature of the entry criteria, chose to liberalize the inclusion criteria to include patients amenable to SVR surgery in the opinion of the investigators, but that may not have strictly met all of the initial entry criteria. This and all other protocol changes, as summarized in the electronic appendix of the rationale and design paper,12 were approved by the STICH steering committee and conducted under the oversight of an independent Data and Safety Monitoring Board (DSMB) reporting to the NHLBI and by the Ethics Committee or institutional review board at each site, including the institutions of several of the authors of the Buckberg editorial. Nevertheless, it should be noted that the majority of STICH patients met all of these initial entry criteria, and analyses presented by Oh et al. at the late-breaking sessions of the American Heart Association, 2009, and by Michler et al. at the late-breaking sessions of the American College of Cardiology, 2010, provide clear evidence that patients meeting all criteria fared no better than those amenable to SVR surgery but not meeting all of the initial empirical LV characteristics. The authors of the editorial go on to suggest ‘that only half of patients had akinesia or dyskinesia and 13% had no prior history of infarction’. In fact, the authors misquote the paper. In STICH, patients randomized to SVR had a median of 50% (inter-quartile range, 40–60%) of the anterior wall that was akinetic or dyskinetic (Table 1 of the NEJM STICH paper).13 In other words, half of the patients had >50% of the anterior wall affected by akinesia/dyskinesia, and 75% had ≥40% of the anterior wall affected. As for 13% of patients not having a history of MI, we fail to see the problem, as in most series, close to 20% of MIs are silent.15 Indeed, to have had 100% of patients with a documented MI would have been surprising. The authors correctly point out that viability studies were not performed in each patient and express concerns as to whether some of the patients required only CABG to recover LV function. Some, but not all, STICH patients had viability studies performed, whereas in others, based on the characteristics of the LV and its scar, a clinical decision not to perform a viability study prior to SVR was made by the treating team, a decision permitted by the protocol. The assessment of potential differences in patient outcomes, according to whether they had a viability study or not, is presently being performed. As for the concerns that echocardiography is inadequate to assess LV characteristics, in STICH, the imaging modality used to assess LV characteristics depended on the ease of acquisition and the quality of the study performed, as it does in clinical practice. The result was the use of cardiac magnetic resonance in some patients, gated single photon (SPECT) in others, and echocardiography in still others. In patients who had more than one study performed, the correlation between techniques was good. There is ample evidence in the literature to document and confirm the utility of echo to assess LV geometry and function in patients before and after SVR surgery.16 The authors go on to point out that only a small proportion of patients in the original STICH paper were reported to have had a measurement of the change in LV volume assessed at 4 months, as originally proposed. Indeed, in the original paper, paired LV volume studies were reported for only 373 patients in which these data were available at the time. However, as presented by Michler et al., we now have 917 patients with baseline studies and nearly 600 patients (∼66% of the patients still alive) with paired baseline and 4-month follow-up studies of good to excellent quality, a database that is unique and confirmatory of the original conclusions. The authors go on to indicate that patients in STICH had only a 19% reduction in LVESVI and that a large number of patients did not meet the arbitrary 30% reduction in LVESVI. They suggest that patients not meeting this empirical reduction in LV volume should be excluded from analyses. As no systematic evaluation of LV volume changes over time with SVR exists, this appears to be an arbitrary suggestion. This is particularly true as time-dependent LV dilatation is thought to occur post-operatively12 and it may be that early 30% reductions in LVESVI gradually drift towards less impressive reductions by 4 months. Michler, in his presentation at the ACC, went one step further and analysed the influence of changes in LVESVI on outcome and found no clear advantage over CABG alone in patients with a >30% reduction in LVESVI with SVR. The authors go on to point out that the number of centres originally chosen to perform the STICH study was significantly expanded and suggest that the expanded STICH surgical team may not have consistently performed true SVR procedures. As, apart from the STICH study, no systematic information exists regarding the 4-month reduction in LVESVI obtained with SVR, any response to this suggestion is speculative. However, a number of lines of evidence suggest that the concerns expressed by the authors are unwarranted. First, the analyses by Michler et al. suggest that changes in LVESVI alone do not predict outcome. Secondly, because mortality in the STICH patients with only CABG was at least as good as that of the RESTORE registry patients undergoing SVR (28% at a median 4 years of follow-up vs. 37% at 5 years), despite similar characteristics (except for a slightly higher NYHA class in RESTORE), one would question the added benefits of adding SVR to CABG in RESTORE patients that resembled those in STICH. Thirdly, no interaction was found between geographical area and the influence of SVR, suggesting that the globalization of STICH did not influence its findings.13 Finally, the authors suggest that adequate LV dilatation (LVESVI ≥ 60 mL/m2) is required in order for SVR to attain its full benefit and that this pre-requisite was not present in all STICH patients. This assertion has some logic, but it should be noted that the median LVESVI in STICH was similar to that in the RESTORE registry (79.7 vs. 80.4 mL/m2) and this cut-off is not based on any objective information. Indeed, as the LV dilates, it frequently becomes progressively more distorted, such that one could imagine that restoration of a very dilated LV to a more normal volume and shape becomes more difficult and may favour suboptimal results. Results presented by Oh et al., and by Michler et al. suggest that the relationship between baseline LVESVI and outcomes with SVR may be more complex than previously thought. We thus conclude that the editorial by Buckberg et al. is misleading. The results of STICH are clear, and further analyses of the STICH trial database have addressed many of the issues raised. Specifically, in these analyses, we have focused on patients meeting the entrance criteria initially empirically chosen based on registry data and expert opinion and have found no benefit of SVR. We are not suggesting that some patients may not benefit from SVR; however, we are stating that in the STICH study, we were unable to identify a patient population that did. These results may not meet the expectations of those that believed or still believe that this procedure should gain more widespread use, however, their position is clearly not evidence based. Finally, it is a matter of record that the STICH study was performed according to the best clinical practices under the watchful eye of the NHLBI and an independent expert DSMB, and to suggest otherwise is inaccurate. Conflict of interest: The authors are the members of the STICH trial executive committee.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.