Cell therapy for acute myocardial infarction: Requiescat in Pace

This editorial refers to ‘The effect of intracoronary infusion of bone marrow-derived mononuclear cells on all-cause mortality in acute myocardial infarction: the BAMI trial’†, by A. Mathur et al., on page 3702. Will cell therapy be approved for the treatment of acute myocardial infarction (MI)? The answer is almost certainly no. The publication of the long-awaited results of the BAMI trial1 in this issue of the European Heart Journal provides an opportunity to reflect on this matter and on the status of cell therapy for heart disease in general. After the exciting results of REPAIR-AMI (a Phase II trial of 204 patients) were published in 2006 suggesting that intracoronary infusion of bone marrow (BM)-derived mononuclear cells (MNCs) improves left ventricular (LV) function and reduces major adverse cardiac events (MACE) in patients with ST-segment elevation MI (STEMI),2 a long series of Phase I–II trials consistently failed to demonstrate benefit3–12 (Figure 1). These studies, however, were relatively small, and their power to show improved clinical outcome was questionable; thus, for a definitive answer, the cardiovascular community awaited the results of BAMI, the logical follow-up to REPAIR-AMI and the largest study of cell therapy in acute MI to date.13 BAMI was designed in 2011 as a pivotal Phase III trial that would conclusively determine whether intracoronary infusion of BM-MNCs improves survival in patients with STEMI, successful primary angioplasty, and LV ejection fraction (EF) ≤45%.13 If BAMI met this hard endpoint, it would probably lead to regulatory approval of BM-MNCs for patients with STEMI and LV dysfunction. Based on an estimated all-cause mortality of 12% at 2 years in the control cohort, a sample size of 3000 patients was calculated to be necessary to demonstrate a 25% reduction in this endpoint.13 Clinical translation of cell therapy for heart disease. The progress toward clinical application of cell therapy has been different in different clinical settings. In STEMI, the likelihood of clinical application (represented by the green-blue-black triangle) has decreased steadily over the past 10 years because of the negative outcomes of clinical trials (red arrows), and because the decrease in the incidence of STEMI and in mortality and MACE after STEMI has made investigations of clinical outcomes challenging. As a result, it is unlikely that cell therapy will be used in STEMI. In contrast, clinical application of cell therapy remains an open question in chronic heart failure and refractory angina, two settings where Phase I or II trials (blue arrows) have been encouraging. Unfortunately, major problems surfaced soon after the trial started recruiting in September 2013. Enrolment was much lower than expected; instead of the target number of 3000 patients, BAMI ended up with 375 patients. In addition, mortality was much lower than anticipated; instead of the estimated 12%, the 2-year all-cause mortality in the control group was 3.82%.1 The combined effect of low enrolment and low mortality was a dramatic loss of statistical power, which plummeted from 90% to 7%. That is, with 375 patients and a mortality of 3.82% in the control group, the probability that BAMI would demonstrate a 25% decrease in mortality, even if the treatment did produce such an effect, was only 7%, making it impossible to test the original hypothesis that BM-MNC therapy decreases mortality. Testing that hypothesis would have required prohibitively large sample sizes; for example, if cell therapy reduces mortality by 25% (from 3.82% to 2.87%), a trial would have to enrol 15 052 patients to have a 90% probability (or power) of demonstrating this reduction with a two-sided type I error of 0.05. In view of these sobering facts, in 2016 the BAMI Steering Committee decided to modify the trial from a hypothesis-testing (interventional) study to an observational study. This meant that BAMI would not be able to achieve its intended objective—to determine whether BM-MNC therapy reduces mortality. Instead, it became a registry that could be used to estimate the incidence of death and other MACE in STEMI patients, in the presence and absence of BM-MNC therapy. The report by Mathur et al. 1 in the current issue of the journal accurately reflects this reality. The authors should be applauded for carefully avoiding unwarranted or misleading statements. They describe outcomes in the control and treated cohorts without making statistical comparisons. Throughout the text, they clearly state that BAMI does not provide a statistical test of the effect of BM-MNCs on clinical outcome and that no conclusions regarding the efficacy of cell therapy are possible.1 Not only was all-cause mortality low, but MACE [e.g. the combination of cardiovascular death and hospitalization for heart failure (HF)] were also sufficiently rare (9.7% in the control group) that showing improvement with cell therapy (or any therapy) would be challenging. Thus, despite widespread expectations, the results of BAMI do not tell us whether or not BM-MNC therapy improves survival or reduces MACE in STEMI. What they do tell us is that in contemporary medicine it is not realistically possible to conduct a trial in STEMI patients using clinical outcome as the primary endpoint (unless one is willing and able to enrol tens of thousands of patients). Before performing a postmortem on this study, it is important to stress how daunting the task of the BAMI investigators was and what a gargantuan effort they delivered. They should be lauded for this enormous undertaking, which involved coordinating the regulatory and logistical aspects of a Consortium of 37 clinical sites in 10 European countries. They did enrol high-risk patients (86% had anterior STEMI, and LVEF averaged 39%1). That the trial did not achieve its intended objectives is not related to execution but to the shifting landscape of medicine, as elaborated below. What went wrong with BAMI? In a nutshell, it succumbed to the seismic changes in the epidemiology of STEMI over the past three decades, driven by improvements in prevention and treatment.14 , 15 Not only has the incidence of STEMI plummeted by >70%, but MACE after STEMI have become much less common, even among patients with LV dysfunction.14 , 15 These dramatic changes are documented by the experience of all major cell therapy trials conducted in STEMI patients in the past 10 years, which have reported low mortality rates, averaging 1.2% at 1 year (range 0–3%), and low MACE rates (e.g. average rate of HF admissions at 1 year, ∼3%; range 0–7%) even in ostensibly high-risk patients with moderate to severe LV dysfunction.3–12 Given this reality, vividly illustrated by the vicissitudes of BAMI, it seems very unlikely that there will be another study of cell therapy in acute MI that uses clinical outcome as the primary endpoint. BAMI is almost certainly the last one. Will there be large-scale trials in acute MI using surrogate endpoints, such as LV function or scar size? This is also unlikely, for two reasons. First, the inability to assess clinical outcome undermines the significance of surrogate endpoints. Secondly, the experience with surrogate endpoints has not been encouraging. Nine major cell therapy trials published since 2010 have failed to demonstrate an improvement in LV function or a reduction in scar size, not only with BM-MNCs4–8 , 10 , 12 but also with CD34+ cells9 and allogeneic cardiac stem cells11 (Figure 1). Recent meta-analyses do not support beneficial effects on scar size or LV function (when the latter is measured by magnetic resonance imaging).3 It is difficult to imagine that there will be much enthusiasm for other similar studies. Given the challenges of showing improved outcome in a population with low death and MACE rates, and given the disappointing results obtained with surrogate endpoints, it seems likely that investigations of cell therapy in acute MI will dwindle or cease altogether (Figure 1). So, does BAMI mean the end of cell therapy for heart disease? Not at all. Perhaps the biggest mistake that one could make in the post-BAMI era would be to extrapolate the experience of acute MI trials to other clinical settings, as if all heart disease were the same. The danger is real. The opponents of cell therapy may pounce on BAMI to support their opinion that all clinical research on cell therapy for heart disease is futile and should stop.16 Such an extreme position is not justified by medical science. In fact, as pointed out before,17 the available evidence indicates that cell therapy may be beneficial in chronic HF and refractory angina, two setting where well-designed Phase II trials of mesenchymal stromal cells (MSCs) and CD34+ cells, respectively, have yielded encouraging results17 (Figure 1). In particular, ixCELL-DCM18 and MSC-HF19 have shown that MSCs improve LV function and MACE in chronic HF, and a patient-level meta-analysis has suggested improved exercise time and angina frequency in no-option refractory angina after CD34+ cell administration.20 The publication of BAMI is an opportunity to redirect efforts towards these clinical settings, where cell therapy is more promising. Importantly, there is evidence that selected cell types (such as MSCs and CD34+ cells) may be more effective than unfractionated BM-MNCs.17 Four major trials in HF [CONCERT-HF,21 DREAM-HF,22 SCIENCE,23 and CSCC ASCII (NCT03092284)] will be completed soon; together, they will probably answer the question of whether cell therapy has a place in the treatment of HF. While awaiting rigorous clinical data, we must keep an open mind. Was BAMI a ‘negative’ study? This would be a gross misinterpretation. It is worth repeating that, as executed, BAMI did not test whether cell therapy is beneficial in STEMI, nor did it show lack of efficacy. BAMI was not a negative study, but a description of the clinical outcomes of two groups of STEMI patients with LV dysfunction (control and treated). As the authors point out, no comparisons between the two groups are possible or valid. It may well be that BM-MNCs are ineffective in these patients, but the results of BAMI do not enable such a conclusion. Does BAMI raise concerns regarding the safety of cell therapy? Two decades of clinical investigation in thousands of patients have shown cell therapy to be safe.24 Although in BAMI there were four hospitalizations for stroke in the treated group and none in the controls, case review at local centres did not suggest a relationship of the strokes to the administration of BM-MNCs, which occurred 1–19 months earlier.1 In addition, when considering the possible increase in stroke, one should also consider the possible decrease in rehospitalization for HF (2.7% of cell-treated patients vs. 8.1% of controls).1 It is, however, unknown whether BM-MNC therapy alleviated HF in this trial. Additional data [e.g. symptoms, New York Heart Association (NYHA) class, exercise tolerance, LV function, brain natriuretic peptide (BNP) levels, etc.] would be necessary to answer this question. An inevitable consequence of a long lead time is that, by the time a trial is published, new ideas or techniques have emerged that were not tested in the trial. The field of cell therapy in 2020 is very different from that in 2011, when BAMI was conceived. Enthusiasm for treating STEMI has receded, due to numerous negative studies in this population3–12 (Figure 1). The focus has shifted toward HF and refractory angina, where cell therapy appears more promising (see above). BM-MNCs have been virtually abandoned, due to the disappointing results of studies that used these cells in the past decade. Instead, almost all contemporary studies use selected cell products, such as MSCs and CD34+ cells.17 Autologous products have been largely replaced by allogeneic cells, which are less heterogeneous, easier and cheaper to mass-produce, and not affected by age and comorbidities.24 Intracoronary cell infusion has been mostly supplanted by transendocardial injection, which enables greater retention of cells in the heart. Thus, the BAMI results are less relevant today than they would have been a decade ago. BAMI was an outstanding effort. It began as an ambitious, definitive Phase III trial; because of low power, it ended up as an observational study. It did not answer the question as to whether BM-MNCs improve clinical outcome in STEMI. (Most probably, we will never know the answer to that question.) It did, however, document very low event rates in STEMI and demonstrate the challenges of conducting trials in this setting using mortality or MACE as endpoints. In this sense, the results of BAMI will be invaluable to the broad cardiovascular community for designing future studies of other therapies. BAMI was certainly not a negative study, and its outcome must not be misconstrued as evidence that cell therapy in general is futile. BAMI may well be the last large-scale study of cell therapy in acute MI, but it is not the end of cell therapy for heart disease. Rigorous clinical trials (some of which are ongoing) are needed to reach a verdict on cell therapy in HF and refractory angina. The most important thing in the post-BAMI era is to keep an open mind and let ourselves be guided by the evidence. Conflict of interest: none declared. The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology.