Clinical Trial Design and Endpoints for Cardiovascular Stem Cell Therapies
Implementation of state-of-the-art reperfusion strategies improved the survival of patients presenting with acute coronary disease. Yet, success of reperfusion interventions with improved survival in parallel with aging and changes in clinical pattern of chronic coronary artery disease contribute to increasing prevalence of heart failure (HF). Multimodal treatments have been introduced to alleviate the symptoms, attenuate or reverse the progression of HF. Besides the life-style management and pharmacological therapies, various device-based interventions have been proposed to target various mechanisms underlying the progression of HF. Despite this progress, therapeutic options for patients with advanced HF remain limited. Advances propelled by translational medicine or technology development led to novel, biologic-based, therapeutic interventions aiming to fill the therapeutic gap and halt or reverse the vicious cycle of HF progression (Fig. 1). This combination of unmet need and encouraging experimental studies led to rapid introduction of clinical trials in early 2000s. Collectively, these trials demonstrated the clinical safety of cell-based interventions in various clinical settings. However, the totality of evidence supporting the efficacy remains scarce due to inconsistencies across the trials given by differences in design, or suboptimal methodology and choice of endpoints (1-3).
Figure 1. Standard and new therapeutic options for patients with advanced HF. From Toth G, Vanderheyden M, Bartunek J. Novel device-based interventions in HF. In J; Bartunek and Vanderheyden, eds: Translational approach to HF, Springer 2014.
Introduction and refinements of randomized clinical trials 50 years ago laid the basic principles of modern clinical research (4) and facilitated rapid advances of medical science ever since. They have been at the basis of the current practice of evidence-based medicine relying on prospective and controlled evaluation of therapeutic strategies. The principles of high quality studies rely on use of randomization, use of placebo or appropriate controls, blinding, appropriate statistical power and adequate decision making process during the conductance of the trial. Endpoints are key element design in the randomized clinical trials. Conventional methodology of high quality trial requires one clearly defined primary end point, defined a priori before its initiation (5). Pivotal trials are designed to provide robust evidence that may support regulatory approvals, expand indications and consolidate therapeutic strategies. In this regard combination of various clinically relevant readouts in the composite endpoint can elevate the impact by reflecting a comprehensive picture. Whether the study use only one primary endpoint or composite endpoint, the right endpoint selection is a crucial step in order to get a definite answer as to whether the method is of clinical benefit to or not. That is, the endpoint(s) must be chosen to reflect the patient’s survival, actual clinical status and quality of life.
B. Methodology and Endpoints in Previous Cardiovascular Stem-Cell Trials
Previous cardiovascular stem cell trials are mostly phase I/II trials addressing the feasibility and safety of biologics under study. Selected trials in patients with acute myocardial infarction (AMI), chronic ischemic HF and chronic ischemia are reviewed in Tables 1,2 and 3.
Table 1. Selected clinical trials with stem cells in patients with acute/subacute myocardial infarction (intracoronary delivery).
Table 2. Selected clinical trials in patients with chronic ischemic HF.
Table 3. Selected clinical trials in patients with chronic myocardial ischemia.
Although several meta-analysis (1-3,6,7) indicate a beneficial effect of cell-based therapy, the clinical translation remains controversial. The beneficial effects appear to be mainly observed in younger patients and in those with significantly depressed left ventricular ejection fraction (LV EF) under 40%. While clinical experience has not raised concerns regarding the safety profile or delivery techniques, there is controversy concerning heterogeneity of cellular products and their manufacturing, patient selection, trial methodology and design as well as procedural aspects. In addition, reported benefits on the surrogate endpoints such as improvements in ejection fraction are perceived as minimal, questioning their clinical significance. This perception stands in contrast with the earlier data on other therapeutic and interventional strategies in cardiovascular disease where apparently minor changes in surrogate endpoints yielded clinical benefit in daily practice. It should be noted, that at the time of early clinical trials, the mechanism of action, functional and biological properties and potency of cellular products was limited.
Several lessons can be derived from early clinical experience as regards the design and choice of endpoints. Surrogate endpoints were widely used to evaluate biologic plausibility and feasibility in early trials, and to aid in hypothesis development for larger studies. Thus, the endpoint(s) should be sensitive enough to track functional and biological effects of the cell based biologics. Traditionally, LV EF has been the most common surrogate endpoint in cardiovascular stem cells trials. However, its assessment varied among the trials, using multiple methods including transthoracic echocardiography (TTE), left ventricular angiography and cardiac magnetic resonance imaging (MRI). Alternative surrogate endpoints such as extent of myocardial damage have also been tested.
In the STEMI setting, early experience in TOPCARE, the Boost trial and the Leuven study (8-10) laid the basis for the future framework of the future clinical trials. The pre-clinical and clinical program that led to Reinfusion of Enriched Progenitor Cell and Infarct Remodeling in Acute Myocardial Infarction (REPAIR-AMI) (11) now represents the benchmark in the clincal translation of cell therapy in the STEMI setting. It provided several clues for further clinical translation of the cardiac regenerative clinical trials:
- Effects of cell therapy on post infarct recovery is apparent in patients with large infarcts and significantly depressed left ventricular function.
- Microenvironment within the infarct tissue and timing of coronary cell transfer could affect the therapeutic outcome in early days after reperfusion.
- Procedural aspects and adjunct pharmacological compounds may alter biological and functional properties of cellular products.
- Though various imaging techniques are suited to track changes in regional and global LV function, cMRI should be the method of choice to assess myocardial damage and function.
Similar lessons can be drawn also from the studies in chronic myocardial ischemia or congestive HF where surrogate endpoints also included clinical parameters such as functional class, exercise time or physical performance assessed from the 6 min walk test.
Taken together, advances in stem cell biology and translational medicine are at the basis of stem-cell based interventions. Their clinical potential is supported by the consistent experimental and pre-clinical findings demonstrating the benefits of regenerative interventions. Yet, despite persisting unmet clinical needs, it is increasingly challenging to demonstrate added clinical benefit with novel therapies to ever evolving and improving standard of care. Experience from early randomized cardiovascular clinical trials reminds us that key points in design of the trial are to formulate an adequate hypothesis, select optimal patient population in need, cell type and device delivery best suited for the given cell type and clinical target. To translate the promise of biologic based interventions, it is essential to ensure the appropriate and rigorous trial design and choice of endpoint along the regulatory path of regenerative interventions. These trials are guided by recommendations of regulatory bodies in US or Europe that delineate generic and pathology-specific requirements as well as rigorous criteria of good clinical practice and clinical research. The following text will detail the thoughts and consideration relevant to stages and settings of regenerative clinical trials.
C. Principles and Recommendations for Phase I/II Cardiovascular Stem-Cell Trials
The regulatory and scientific processes deployed in the development of cell therapy are essentially the same as those used in drug trials. The process for generating evidence on the safety and effectiveness of an intervention commence with trials including limited number of subjects. Pending the feasibility and safety, it may then progress to confirmatory trials. Typically initial Phase I trials are healthy volunteers in order to characterize the properties and safety as well as pharmacokinetic profile of a given compound and to optimize methods of delivery. Phase II studies are used to assess the feasibility and therapeutic efficacy on clinically relevant surrogate endpoints in the limited patient’s cohort. By the time a study has reached Phase II, sufficient data from animals and from Phase I studies have been collected to determine the safety.
The execution of Phase I/II studies is governed by established standards of clinical trials (5) with appropriate trial structure and logistics. They should be based on prospectively declared, detailed protocols written according to ethical standards. Protocols should be analyzed in detail and closely inspected by Steering/Protocol Review committees comprising various stakeholders involved in the clinical care, Data Safety and Monitoring Boards, as well as by local review boards. It is recommended (12) that at this stage, a measure of success in these trials is to provide clinical direction. Safety readouts are the primary endpoint together with feasibility. Positive Phase II studies with surrogated endpoints do not lead to approval and wide-spread adoption. Positive Phase II studies will produce a dataset that includes a foundation of benefit and safety to aid in justifying subsequent studies with a clinically relevant objective. This stepwise process is meant to ensure that safest products with the strongest efficacy move forward to Phase III trial clinical testing in larger populations.
Specific Considerations for Design of Regenerative Trials
Given the specifics of stem cell based intervention, Phase I trials are not meaningful in healthy volunteers. Thus the initial feasibility and safety profile is in a limited number of patients. At this stage, information from the bench or animal testing also provides some data on biological safety as regards to toxicity or chromosomal alterations and tumorogenicity, mortality or major morbidity secondary to the treatment or its delivery. Cell products characteristics, such as stability, potency, purity and mechanisms of cell action may also have been addressed.
The European Medicine Agency (EMA) has released a statement paper (13) covering specific aspects related to stem cell’s based product (SCBP) with an intention for marketing authorization (MA) application. In first-in-man studies, specific safety endpoints may need to be defined based on theoretical considerations and to detect early any toxicity arising from potential contaminants in the final cell product. In cases where sufficient safety evidence cannot be established in the preclinical studies, for example, due to difficulties in finding an appropriate animal model, the evidence should be generated in clinical trials by including additional endpoints for efficacy and safety. The need for and duration of post-authorization long-term efficacy follow-up should also be identified during the clinical trials, taking into consideration results from nonclinical studies and the intended therapeutic effect. In the US, the Food and Drug Administration (FDA) has postulated that cells or tissues used for therapeutic purposes are codified under the Good Tissue Practice (GTP). Consequently, FDA issued guidance about how the biologic and device regulations apply to cellular and genetic therapies (14). Classification of stem cell based therapies is based on indication to be treated. Cell-based therapeutics “that are, minimally manipulated, labeled or advertised for homologous use only, and not combined with a drug or device” do not require FDA approval. In contrast, manipulated autologous cells for structural use meet the definition of somatic cell therapy products and require an “investigational new drug” (IND) exemption or the FDA license approval. Accordingly, IND application necessitates detailed study protocols describing the clinical plan as well as the preparation and testing of the therapeutic cell product.
Population Selection and Trial Design
As cell therapy aims to advance the therapeutic options, patient selection may be based on standards of care in given clinical setting, i.e STEMI, CHF and CMI patient populations.
Phase I/II studies may have the option to carry out hypothesis testing at a levels of 0.10, 0.05 and 0.01. Determining the sample size for the cell therapy trial requires knowledge of the event rate anticipated in the control group and an estimate of the reduction likely to be achieved by the intervention. In the initial phase when addressing the feasibility and safety signals, the number of patients is typically limited. Dose-escalation trial design with safety evaluation at the end of each dose is favored. Surrogate efficacy signals are observed and collected and should be accompanied by the appropriate clinical safety outcome readouts.
The sample size calculation for ensuing Phase II studies is critical and may be based on the projected impact of the chosen primary endpoints. In a severe HF population an intervention anticipated to improve the LV EF more than 5%, depending on the method used to assess this surrogate endpoint, may include a rather small sample size, (usually up to 50 patients for a trial with 95% power to detect the difference at an alpha of > 0.05%). Those numbers could be reduced by using composite endpoints such as placing the patients into categories of improved, worse or unchanged (see details later), and/or by an adaptive design that allows patients in a Phase II study to contribute to the numbers required for a Phase III trial.
Other considerations may address the pathophysiologic basis of the disease and baseline clinical characteristics of the patients including cardiovascular risk profile and interference with potency of the given autologous cell product. As in pharmacological trials at this stage, it is important to identify patients who are most likely to benefit. Therefore, Phase II ischemic regenerative studies preferably include patients with advanced disease. For instance, in the REPAIR-AMI (11) study, patients with a lower EF at (> 48.9%) showed a significant, 3-fold higher recovery in global LV EF than was seen in the converse group. In the REGENT study (15) significant functional benefit was observed only in cell-treated patients who had baseline LV EF less than 37%. In the BOOST trial (8) only patients with larger infarcts and greater infarct transmurality demonstrated sustained functional improvement at later time points. In the study by Janssens et al (9) cell therapy led to enhanced recovery of regional function only in the most severely infarcted myocardial segments (characterized by the greatest infarct transmurality).
In ischemic HF setting within the frames of the TOPCARE-CHD registry (16), NT-proBNP serum levels >735 pg/ml at baseline were an independent predictor of a favorable response 3 months after intracoronary administration of bone-marrow mononuclear cells (BMMNCs). In the STEMI setting, stratification based on infarct size or presence of microvascular obstruction can be utilized to define the range of the myocardial damage responsive to regenerative therapies. In addition, understanding the impact of microvascular obstruction on the efficacy of coronary cell transfer may yield better insights into the design of the future Phase III studies.
Randomization is imperative. A well-matched control group is critical to evaluate effects of treatments and to properly ascribe culpability to any major adverse cardiovascular event (MACE) seen in the cell therapy group. Modified cross-over design in the Phase I/II can be considered if initial safety readouts do not suggest increased clinical risk associated with the treatment. This may enrich the analysis of the functional outcome and facilitate the design and sample size of the larger Phase II study.
Any bias in a study should be eliminated. To ensure unambiguous endpoint readout in cardiovascular stem cell therapy trials, blinding and appropriate control arm is critical for trial interpretation. An analysis of previously published studies on therapeutic trials of treatments for AMI showed that unblinded-randomized studies had about 16% larger positive effect achieved compared with blinded-randomized trials (17). This shows that randomization alone may not guarantee valid trial results. Methodologic issues affecting the validity of randomized controlled trials can occur both before and after assignment and thus influence the objectivity of the trial. Blinding and placebo may be challenging in some settings such as severe cardiac dysfunction with direct myocardial delivery where risk exposure to either unknown product or procedural risk associated with delivery needs to be weighed against the conventional double-blind placebo controlled design applied in the pharmacological trials. In this regard, sham procedure mimicking the active intervention without factual endomyocardial delivery with separation of interventional and clinical teams throughout the trial course may be an acceptable alternative to placebo-controlled double-blind design. On the other hand, if there are no ethical, procedure-related concerns, randomized, placebo-controlled, double-blind design remains the standard choice for the cell therapy trial. This has been underlined by a meta-analysis (18) pointing that two thirds of conclusions favouring an intervention would lose support if trials with unclear or inadequate allocation concealment were excluded from the meta-analysis. As demonstrated, unblinded participants may be affected by biases in reporting their symptoms, willingness to continue in the study, use of other effective intervention methods, and placebo effects. Recent meta-analysis of cell therapy in AMI setting (19) also highlighted the importance of rigorous study design in randomized controlled trials. Thus to ensure the objectivity, in trials with autologous stem cell source, the cell harvest should be performed in all patients, including the control group. This allows additional patient stratification as regards their biological stem cell profile in active and control arms.
Where substantial prior information is available and the patient characteristics in the new trial are similar to those of prior studies, a Bayesian trial design has been suggested as facilitating the trial enrollment and completion (20). In clinical trials, traditional statistical methods may use information from previous studies only at the design stage. Then, at the data analysis stage, the information from these studies is considered as a complement to, but not part of, the formal analysis. In contrast, the Bayesian approach uses Bayes’s Theorem to formally combine prior information with current information on a quantity of interest with an idea to consider the prior information and the trial results as part of a continual data stream, in which inferences are being updated each time new data become available. In other words, it minimizes the need for excessive patient enrollment. However, it is essential to understand the strengths and limitations of various clinical trial design in order to best interpret the available data regarding management of these complex patients.
D. Endpoints in Phase I and II Trials
General requirements and safety/efficacy profiles for Phase I, II and III trials are summarized in Table 4. Given the above mentioned consideration and specifics of stem cell-based interventions, the clinical community is attempting to harmonize paths of clinical adoption and rationalize choice of relevant endpoints. Nevertheless, the choice of clinical endpoints with adequate sensitivity and specificity to evaluate incremental improvement above that conferred by current standard of care presents a challenge. The proposed path shares common and disease specific selection (Fig. 2).
Table 4. Requirements and safety/efficacy profiles recommendations for Phase I, II and III trials.
Figure 2. The proposed path of endpoints selection in cardiac stem cell trials.
Keeping with the stepwise approach to the clinical path of adoption and the role of Phase II studies the following recommendations for endpoints can be considered:
Primary objective of all initial studies is to define the safety profile. The European Society of Cardiology (ESC) Stem Cell Task Group (21) proposed that initial safety readouts should focus on the risk of tissue injury or abnormal growth and the possibility of arrhythmias, requiring Holter monitoring or interrogation of an implantable cardioverter/defribrillator in HF. In that regard, initial experience should focus on gathering information on the key pharmacodynamics and pharmacokinetics features of regenerative therapy in broad terms. In case of stem cell therapy, this aspect may include the assays addressing the mechanistic action, potency, interaction with disease markers, dependency on the mode of administration. Cell therapy may be associated with challenges in each of these areas that are distinct from traditional pharmacologic therapies. For example, the component of a cell product, which contributes to efficacy and from which a dose-response could be established, is often unknown, and potency can be quite variable and difficult to quantify. As alluded earlier, dose dependency of the safety profile should be addressed in parallel to the explorative analysis of the efficacy signals.
Consistent with the role of Phase I/II study in the regulatory path, efficacy signals should not determine the sample size or power of the trial. It is critical not to overemphasize the individual outcome of surrogate endpoints, but rather to look at the totality of evidence they provide about the potential clinical benefit. Positive Phase II studies do not lead to approval and widespread clinical use, but instead produce subsequent studies by providing foundational evidence and generating hypotheses. Surrogate endpoints are generally acceptable to evaluate biologic plausibility and feasibility in these early trials, and to aid in hypothesis development for larger studies. Surrogate endpoints parameters used to assess the effects of stem cell therapy on survival in chronic ischemic heart disease should be able to:
- address the safety of cell-based intervention.
- closely correlate with survival.
- changes should reflect changes in the prognosis.
- there should be a pathophysiological basis for both relations.
The latter is particularly important. Positive results in surrogate endpoint parameters supported by a known pathophysiological basis, can be viewed as a probability signal to achieve the later clinical benefit (22-24) and thus provide a scientific/mechanistic support to formulate the hypothesis in the pivotal Phase III trial.
The surrogate endpoints parameters typically used in cardiac stem cell therapy should include end points from various domains reflecting target organ changes within the clinical context:
- measures of left ventricular function and structural evaluation (i.e: ejection fraction, end-systolic and end-diastolic volumes and dimensions, pressure-volume relationships, stroke volumes and indexes, ventricular sphericity, infarct scar, perfusion defect, ischemia).
- Biomarkers reflecting the presence and severity of disease: for example natriuretic peptides, cardiac enzymes, C-reactive protein, transcriptomic biomarkers, cytokines.
- Functional capacity and symptoms relevant to the clinical setting: 6-minute walking distance, maximal O2 consumption (VO2max), New York Heart Association Class (NYHA class), angina score.
- Patient-reported outcome, such as quality of life – Living with HF questionnaires, dyspnea, alone or in combination.
All surrogate endpoints need to be be complemented by observational analysis of clinical outcome related to the given setting. Typically clinical endpoints include mortality and cardiovascular morbidity including myocardial (re)infarction [Re-MI], stroke and hospitalization due to acute HF decompensation. As Phase I/II cell trials are often unblinded, analysis of all-cause safety endpoints (especially all-cause mortality) may reduce possible bias in those trials (24). Taken together, Phase II studies are outlined to demonstrate concordance in surrogate endpoints from different domains and to set the base for the hypothesis in the Phase III.
Phase I/II Endpoints in Specific Clinical Settings
Table 5. summarizes proposed primary and secondary endpoints according to specific clinical setting in patients with ischemic heart disease as well as peripheral arterial disease (PAD). Although most trials in Phase II are powered for surrogate endpoints, all the way through the cell trial clinical endpoints must be followed and safety readouts must be collected on an observational basis. As safety is the primary objective at this stage, appropriate readouts may be implemented to evaluate procedural complications, such as the risk of perforation with intramyocardial injection, or cell specific issues such as immunogenicity in case of allogeneic cells. As regards to setting specific safety issues, potential pro-arhythmogenic effect in congestive HF or excessive recurrent HF hospitalizations should be closely followed while patients with MI should be monitored for the possible reintervention/revascularization. The need for repeat revascularization as well as persistent and recurrent chest pain should be followed in patients with recurrent angina.
Table 5. Proposed primary and secondary endpoints according to specific clinical setting in patients with ischemic heart disease and peripheral arterial disease.
a.) Acute/Subacute Myocardial Infarction
Besides conventional assessment of regional and global LV function in the STEMI setting, assessment of infarct size reduction may indicate a potential benefit in late mortality, as it reflects a pathophysiological mechanism that stem cells trigger in MI setting.
Recent technical advances make segmental strain analysis as an index of regional contractility, attractive and sensitive surrogate endpoint for evaluation of benefits of stem cell therapy, especially in the context of modest global LV function improvement. Although the clinical significance of improvement in segmental strain in the absence of a more global increase in LV contractility is yet to be determined; such changes may predict a beneficial response to therapy with respect to LV remodeling.
Global EF is load dependent and under the influence of different factors, including LV afterload, preload, and in the setting of LV dysfunction, neurohormonal activity. The assessment of LV contractility and load-independent indices such as obtained from the pressure-volume loops may be free from influences of all these factors and thus better evaluated through the efficacy signals of the cell therapy.
The attenuated ventricular remodelling as a surrogate endpoint may also be analyzed in the setting of MI. In this regard, endsystolic LV volume (ESV) is a surrogate endpoint of interest as it remains the primary predictor of survival after myocardial infarction, being superior to ejection fraction when ejection fraction is low (<50%) or when end-systolic volume is high (23).
Alternative endpoints include assessment of myocardial perfusion and microvascular circulation (through MRI, fractional flow reserve and TTE/TEE coronary flow reserve).
b.) Refractory Angina/Ischemia
In patients with chronic myocardial ischemia with no revascularization option, the therapeutic objective is to reduce angina frequency and to improve exercise tolerance and exercise duration (analyzed through stress test). In addition, as demonstrated in trial by Losordo D et al (25), stress myocardial perfusion defect and wall thickening could be used as surrogate readouts.
Recognition that therapies designed to repair the microcirculation may enhance cardiovascular function led to a paradigm shift in treatment strategies for acute and chronic ischemia. In that regard, the concept of reconstituting the microvasculature as part of a pathohysiology-based terapeutic strategy may be considered in the future phase cell therapy II trials by deploying appropriate interventional or imaging based surrogate endpoints.
c.) Peripheral Arterial Disease (PAD)
Multiple surrogate endpoints could be analyzed in patients with PAD. These endpoints include indexes of ischemia (change in ankle-brachial index, change in transcutaneous oxygen tension, pain free walking distance) and subjective parameters (level of pain). It is important that additional endpoints include vascular hard events as ulcer healing and amputation. In the context of peripheral disease and critical limb ischemia, avoiding possible amputation and disappearance of pain are important endpoints whose importance cannot be ignored even if the method does not lead to better survival.
d.) Chronic Heart Failure (HF)
In chronic HF, the goal is to track impact on LV remodeling and function. The assesment of ventricular remodeling (i.e., characteristic changes in ventricular volumes and wall thickness and shape, ventricular sphericity, pressure-volume loops) could point to a positive effect of stem cells in chronic HF. As reverse remodeling may lead to a parallel decline in systolic and diastolic volumes, it is possible that LV EF may show only minor changes. Changes in ejection fraction should be evaluated in parallel with other surrogate endpoints such as exercise tolerance or changes in humoral biomarkers. In this regard, the lessons should be taken from the previous studies addressing benefits of positive inotropic drugs showing that improved cardiac function did not translate into the clinical benefit. Thus, the continuation of the clincial path in the early stage is determined by the synergy between the observatory surrogate signals and mechanisms leading to clinical safety readouts including mortality, HF admissions or occurrence of life threatening arrhythmia (ie VT/VF) remains the primary safety endpoint.
E. Definition of the Clinical Meaningful Response of Surrogate Parameters
The following definition of the clinical meaningful response in surrogate endpoints can be considered:
- symptoms: a change of one in NYHA class;
- in the functional domain, an increase of 50 meters in six minute walk test or of 10% in VO2max;
- index of LV function remodeling, a change of 5% in absolute LV EF or of 20mL or 10% (whichever is the greater) in left ventricular end systolic volume; and a change of 35% or 300pg/mL in NT-proBNP;
- pain free walking distance improvement, and improvement in ankle-brachial index of 0.15.
Although the investigators of a Phase II study may not be able to predict a priori how big the effect will be (for example a change in LV EF), they certainly should be expected to determine the extent of effect in the end. In addition, a comparison between observed versus expected changes in different surrogate parameters might be one of the best measures of the success of a Phase II trial.
F. Imaging Techniques for Surrogate Endpoints
Cardiac imaging is instrumental to evaluate surrogate endpoints of cardiac structure and function in tracking biological and functional effects. Since sample size is directly related to standard deviation of the surrogate endpoints (which depends on accuracy of the imaging technique), the higher the resolution of given imaging modality the lower the variability (26-29). The ideal assessment method should be one that is able to analyze cardiac structural and functional changes, including infarct size, myocardial perfusion and microvascular function, regional wall motion, volumes, and chamber geometry in one setting. The strengths and weaknesses of the most often used imaging modalities in the context of stem-cell application are given in Table 6.
Table 6. Strengths and weaknesses of different imaging techniques for typical goals in the context of stem cell implantation.
Cardiac MRI is considered technique of choice for these purposes (27), as it provides imaging capability to detect wide range of dynamic physical and chemical processes including flow, motion, morphology and tissue composition. The application of this imaging modality enables accurate quantification of ventricular volumes, function and mass for longitudinal monitoring of therapeutic efficacy. Cardiac MRI, especially considering infarct measurement by late gadolinium enhancement, has provided a powerful diagnostic tool for the determination of scar burden and for defining the viability of injured myocardium, and therefore is a suitable tool to monitor the myocardium after stem cell therapy. Myocardial strain calculated from cardiac MRI tagging is currently regarded as the non-invasive gold standard for assessment of regional function. Though MRI has become a key method to demonstrate surrogate endpoints efficacy in early phase, its use may be limited due to devices implanted in the target patient population.
Techniques such as single-photon emission computed tomography (SPECT), contrast enhanced echocardiography, PET (positron-emission tomography) are helpful alternatives in assessment of infarct size and myocardial perfusion.
Clinically available techniques for evaluation of viability include MRI and nuclear imaging with PET (mainly using F18-FDG, evaluating glucose utilization) and SPECT (with F18-FDG or Tc-99m–labeled agents), or low-dose dobutamine echocardiography (assessing contractile reserve and wall motion). PET currently is a highly sensitive method for assessing myocardial viability and is underutilized in cardiac stem cell investigations. Contractile reserve can also be assessed by low-dose dobutamine MRI.
Variety of imaging modalities is used for assessment of LV function and volumes. Number of them may either pose a safety risk in patients with devices or have been criticized as being exposed to reproducibility issues, interference of atrial fibrillation and dependency on the image quality. In this regard use of contrast echocardiography in combination with 3-dimensional imaging may overcome these limitations. In this regard, echocardiography remains the most readily available technique to assess cardiac structure and function. It should be noted that it has been widely utilized in other pharmacological or interventional trials where changes in function tracked by echocardiography were directly related to the clinical outcome validating its broad use in the clinical setting.
The combined imaging such as PET/MR or PET/CT imaging may be deployed to provide not only robust information on cardiac structure and function, but contribute to mechanistic understanding of the cellular interventions (30) by assessing cellular retention and survival and thereby to provide information relevant to dosing and most optimal route of delivery.
Taken together, a variety of imaging tools are available to evaluate biological and structural changes induced by the cellular therapy. Cardiac MRI is the technique of choice as providing the most comprehensive technique able to evaluate multiple facets of cardiac function and structure. Benefiting from the routine use of contrast and recent advances in the imaging technology echocardiography can be considered as being readily available in large clinical trials. The choice of the imaging modality should depend also on the local expertise in individual sites to ensure the optimal quality and reproducibility of the measurements.
G. Phase III/IV Trials
Phase III trials are conventionally randomized, placebo-controlled, double-blind and designed to generate definitive conclusions about the clinical merits of applied method. The priority of regulatory bodies is to approve therapies and interventions that meaningfully improve quality of life and decrease cardiovascular morbidity and mortality. Hence, Phase III trials must be clinically driven, emphasizing hard clinical endpoints starting with all cause mortality/cause specific mortality (31). Then, depending on the clinical setting, endpoints include cardiovascular improvement/deterioration including re-infarction and need for revascularization, life threatening arrhythmias and HF worsening and re-admissions. The choice of either all-cause mortality and cardiovascular mortality remains debated (31); while all-cause mortality may appear more relevant from the regulatory perspective, it is associated with random noise diluting the therapeutic impact in high risk patients such as HF patients. In this particular setting, non-cardiovascular mortality is increased due to aging or co-morbidities and cardiovascular mortality may be preferred over the all-cause mortality (31). Nevertheless, all-cause mortality should serve as a safety readout and should be directionally concordant or neutral with the eventual improvement in cardiovascular mortality. The HF population remains the most challenging target for cell therapy (31). The HF Association of the ESC proposed in its consensus to define HF hospitalization as at least an overnight stay in-hospital caused by substantive worsening of HF symptoms and/or signs requiring the augmentation of i.v. HF therapy including inotrope, diuretics or vasodilators, ideally pre-defined in the critical events manual (31). Similar to endpoints using cardiovascular mortality, tracking the non-HF hospitalization is also of interest as their reduction or concordant direction with HF admissions provides the safety re-assurance and strenghtens the findings of the cardiac-related endpoints.
To determine the sample size for any pivotal trial of cell therapy requires knowledge of the event rate anticipated in the control group based on the standard of care and an estimate of the reduction likely to be achieved by the intervention based on the surrogate efficacy signals from Phase II. For example, in a severe HF population in which the probability of death or HF hospitalization in the course of a year is 35%, an intervention anticipated to reduce the event rate by 50% would require 330 patients for a trial with 90% power to detect the difference at an alpha of <0.01%. The number of patients that must be enrolled rises to 1500 if the intervention is expected to reduce the event rate by only 25%. If the primary endpoint is all-cause mortality sample-size will need to be even larger to account for the ‘random noise’ added by other deaths due to associated HF co-morbidities. Such numbers can be reduced by maneuvers such as defining composite endpoints that include classifying patients into categories of improved, worse or unchanged; and by an adaptive design that allows patients in a Phase II study to contribute to the numbers required for Phase III. Though being instrumental to project estimated rates of events, standard of care is a dynamic process and projected event rates based on the previous published data may be often different at the time of trial execution. This can hamper the expected enrollment rates or be associated with lower than expected number of events leading potentially to the trial termination. These risks are to be specifically considered in patients with chronic myocardial ischemia or critical limb ischemia where improvements in standard of care and technology advances can potentially reduce pool of patients previously qualifying for cell therapy interventions.
As in Phase II studies, sample size, design, control arm and blinding are affected by the clinical setting and delivery technique. The blinding is crucial as its quality is critical for objective evaluation of clinical endpoints such as quality of life, as well as end points such as re-admissions/re-hospitalizations. Additional challenge for adjudication of clinical events, such as re-admissions, are different in local clinical practices and increasing economic pressure can produce heterogeneity in the rates of admission for otherwise similar clinical presentation. Timing of obtaining the primary endpoint should also be carefully considered in the context of the target population and expected rates of clinical events. Typically one year follow-up is the most meaningful timing for the evaluation; this time can be shortened if the patient population is enriched to include patients with advanced disease or when using composite endpoints.
H. Composite Endpoints in Phase III Trials
Considering that individual “hard” endpoints often require large trials and lack statistical power when insufficient number of patients are included, composite endpoints might provide an acceptable alternative to assess the clinical effect of cell intervention. A composite outcome can prove helpful, enabling early clinical adoption in a high risk populaton or population with unmet need as it was demonstrated by the initial CRT regulatory approval. The appropriate a priori identification of composite endpoints can increase the statistical precision and efficiency of the trial and make it less costly. In addition, a composite outcome may be helpful in a situation where it is difficult to decide which outcome to elect as the primary clinical outcome measure by using the combination of various readouts from various domains. However, a caveat should be that adding more components should not make the interpretation of the results complex. A positive example of a Phase III trial with a composite endpoint is the CAPRICORN study (32) where an important component of the composite outcome has not been substantially modified and the combination of several endpoints proved the superiority of the active treatment. Fundamental condition for a composite outcome is that individual outcomes contributing to a composite outcome have to be associated with the primary objective. Likewise, their analysis is to be done in the hierarchical order where the most significant clinical outcome cancels out any other outcomes that are clinically less significant. An example of this is the LIFE trial where the procedure had a large effect on revascularization or stroke, but not on death or infarction (33). The limitation in terms of different clinical significance of the observed individual endpoints may be leveled by using a pre-defined analytical plan giving varying weight to the components; for example cardiovascular death 1, reinfarction 0.5 and target vessel revascularization 0.1 point.
Though the appropriateness of the assigned arbitrary values to each of endpoints could be debated, the methodology of hierarchical composite endpoint can reduce the sample size needed to show efficacy as compared to a single clinically-related endpoint. In this regard, Finkelstein & Schoenfeld analytical plan of the hierarchical evaluation may serve as an example. For instance, a trial in the HF setting incorporates various events, such as all cause death, HF events, change in 6 min walk distance and changes in ejection fraction or end-systolic volumes. The analysis is based on a hierarchy of endpoints and is constructed by comparing every subject in the active arm to every subject in the comparator arm and assigning +1, 0, -1 depending on whether the subject in the treatment arm did better, same or worse. These scores are assigned at the designated time of the primary readout, for instance at 52 or 104 weeks. In case of mortality, days alive out of the designated time period or number of HF events are counted; using pre-specified cut-off values. Other readouts are categorized as meaningful improvement, deterioration or no meaningful change. Then, mortality is evaluated first. If both died, the one dying later did better. If only one died, the one surviving did better. If neither died, the next point in the hierarchy is evaluated. The one with less HF events did better. If they are similar, a lower endpoint hierarchy such as 6 minute walk distance is evaluated. Subjects tied in all these endpoints will then be evaluated for other endpoints such as volumes or ejection fraction. In the final step, treatment groups are compared using a test statistic based on the sum of net scoress for all subjects.
Hence, choice of individual endpoints to compose the hierarchical analytical tree should be based on several considerations:
- does the composite endpoint measure a severity of the disease – does it cover its critical components;
- can its use solve the medical problem or it is just a statistical convenience;
- are the individual parameters valid, biologically plausible and of importance to patients;
- are the final results clear and meaningful, to they provide a basis for therapeutic decision, does each endpoint support overall result and may it indicate/predict improved mortality (based on state-of-the art knowledge).
Composite endpoints in stem cell trials may include: cardiovascular death, reinfarction, worsening HF with repeat hospitalizations or intravenous treatment, target vessel revascularization, stroke, implantable cardioverter device therapy, heart transplantation in ischemic disease HF, endovascular treatment, ulcer healing and leg amputation in case of patients with peripheral artery disease. Cardiovascular death includes death resulting from an acute myocardial infarction, sudden cardiac death, death due to HF, death due to stroke, and death due to other cardiovascular causes, including unwitnessed death without other cause of death. As alluded earlier, it is also important to emphasize that complex composites combining objective measures of mortality/morbidity with subjective measures (NYHA class) or mechanistic endpoints (ie BNP) are often difficult to interpret and thus are generally discouraged.
I. Future Directions of Cardiovascular Stem-Cell Trials
As our knowledge about cardiac stem cell therapies is further evolving, the insight into the significance of patient-specific disease management grows. In many cases it will be necessary to tailor the endpoints to meet the needs of the population under study. In this regard, systematic stratification of patients to adequately match clinical trials will be of paramount importance. Clinicians and scientists should find a way to tailor the appropriate therapy to patients in need. Therefore, the emphasis should be on delineating acute versus chronic disease to ensure proper target strategy, to identify responders and non-responders as well as pre-treatment of co-morbidities to limit modifiable confounding factors; this is especially important for stem cell treatment. In every case, the hard clinical endpoints remain necessary for future clinical trials, but standardization of outcome measures and cell manufacturing remain an imposed standard in the trial design. Standards of efficacy need to be established as continuous improvements in standard of care lead to better quality of life and survival; likewise regional differences in the health care policy need to be considered as they may impact the assessment of readouts as readmissions (31). Increased socio-economical pressure forces physicians and health care providers to shift the care towards in house and outpatient care and new endpoints reflecting this change in the socio-economic and health care policy and standards should be considered. In this regard, quality of life or individual patient independence may gain regulatory importance. Questions as to which endpoints confirm the viability of a new therapy for cardiac disease and its specific setting such as HF or post MI setting should be topic of further academic and regulatory dialogue.
In addition to examining the technology on an individual basis, additional data on patient subsets, anatomic and physiologic correlates, and the duration of effect are required. There will be patients who benefit more from stem cell therapy than others and there are non-responders. As the initial autologous stem cell-based cardiac regeneration therapy led to a rather rapid translation into early-phase clinical trials, the research has now reached the phase where conclusive answers and thoughtful fine-tuning are needed. Phase III, large-scale, multicenter, double-blind, randomized clinical trials performed under rigorous safety standards are necessary to definitely confirm the clinical benefit of cell products under study.
Together with above raised questions, new types of cells, new preparations of cell products (34), improvements in delivery methods, strategies to enhance cell potentiality and to improve myocardial microenviroment, the creation of bioartificial hearts and the potential benefits of cell transplantation in nonischemic HF pave the road to new cell-based biotherapeutics. Consensus on clinical endpoints and methodologies used to assess those endpoints are necessary to move forward and critically examine the quality of data gathered in future clinical investigations of cardiac stem cell therapy.
- Martin-Rendon E, Brunskill SJ, Hyde CJ, et al. Autologous bone marrow stem cells to treat acute myocardial infarction: A systematic review. Eur Heart J 2008; 29:1807-1818.
- Abdel-Latif A, Bolli R, Tieyjeh IM, et al. Adult bone marrow-derived cells for cardiac repair: A systematic review and meta analysis. Arch Intern Med 2007; 167:989-997.
- Wen Y, Meng L, Ding Y, et al. Autologous transplantation of blood-derived stem/progenitor cells for ischaemic heart disease. Int J Clin Pract 2011; 65:858-865.
- Hill AB. The environment and disease: association or causation? Proc R Soc Med 1965; 58:295-300.
- Stanley K. Design of randomized controlled trials. Circulation 2007; 115:1164-1169.
- Delewi R, Hirsch A, Tijssen JG, et al. Impact of intracoronary bone marrow cell therapy on left ventricular function in the setting of ST-segment elevation myocardial infarction: a collaborative meta-analysis. Eur Heart J 2014; 35(15):989-998.
- Kandala J, Upadhyay GA, Pokushalov E, at al. Meta-analysis of stem cell therapy in chronic ischemic cardiomyopathy. Am J Cardiol. 2013; 112(2):217-25.
- Meyer GP, Wollert KC, Lotz J et al. Intracoronary bone marrow cell transfer after myocardial infarction: Eighteen months’ follow-up data from the randomized, controlled BOOST (BOnemarrOw transfer to enhance ST-elevation infarct regeneration) trial. Circulation 2006; 113:1287–1294.
- Janssens S, Dubois C, Bogaert J et al. Autologous bone marrow-derived stem-cell transfer in patients with ST-segment elevation myocardial infarction: double-blind, randomized controlled trial. Lancet 2006; 367:113–121.
- Assmus B, Schachinger V, Teupe C, et al. Transplantation of progenitor cells and regeneration enhancement in acute myocardial infarction (TOPCARE-AMI). Circulation 2002; 106:3009-3017.
- Schachinger V, Erbs S, Elsasser A et al. Intracoronary bone marrow-derived progenitor cells in acute myocardial infarction. N Engl J Med 2006; 355:1210 –1221.
- Hare JM, Bolli R, Cooke JP, et al. Phase II clinical research design in cardiology: Learning the right lessons too well: observations and recommendations from the cardiovascular cell therapy research netwok (CCTRN). Circulation 2013; 127:1630-1635.
- George B. Regulations and guidelines governing stem cell based products: Clinical considerations. Perspect Clin Res 2011; 2(3):94–99.
- US Food and Drug Administration. Guidance for human somatic cell therapy and gene therapy, FDA Center for Biologics Evaluation and Research. 1998
- Tendera M, Wojakowski W, Ruzyllo W, et al. Intracoronary infusion of bone marrow-derived selected CD34+CXCR4+ cells and non-selected mononuclear cells in patients with acute STEMI and reduced LV EF: results of randomized, multicentre Myocardial Regeneration by Intracoronary Infusion of Selected Population of Stem Cells in Acute Myocardial Infarction (REGENT) Trial. Eur Heart J 2009; 30(11):1313-21.
- Assmus B, Fischer-Rasokat U, Honold J, et al. Transcoronary transplantation of functionally competent BMCs is associated with a decrease in natriuretic peptide serum levels and improved survival of patients with chronic postinfarction HF: results of the TOPCARE-CHD Registry. Circ Res 2007, 27; 100(8):1234-41.
- Chalmers T, Celano P, Sacks H, Smith H Jr. Bias in treatment assignment in controlled clinical trials. N Engl J Med 1983; 309:1358–1361.
- Pildal J, Hrojartsson A, Jogensen KJ, Hilden J, Altman DG, Gozsche PC. Impact of allocation concealment on conclusions drawn from meta-analyses of randomized trials. Int J Epidemiol 2007; 36:847–857.
- Jeong H, Woo Yim H, Cho Y, et al. The effect of rigorous study design in the research of autologous bone marrow-derived mononuclear cell transfer in patients with acute myocardial infarction. Stem Cell Research & Therapy 2013; 4:82.
- Lee JJ, Chu CT. Bayesian clinical trials in action. Stat Med 2012; 31(25):2955-72.
- Bartunek J, Dimmeler S, Drexler H, et al. The consensus of the task force of the European Society of Cardiology concerning the clinical investigation of the use of autologous adult stem cells for repair of the heart. Eur Heart J 2006; 27:1338-1340.
- Kramer DG, Trikalinos TA, Kent DM, et al. Quantitative Evaluation of Drug or Device Effects on Ventricular Remodeling as Predictors of Therapeutic Effects on Mortality in Patients With HF and Reduced Ejection Fraction: A Meta-Analytic Approach. J Am Coll Cardiol 2010; 56:392-406.
- White HD, Norris RM, Brown MA, et al. Left ventricular end-systolic volume as the major determinant of survival after recovery from myocardial infarction. Circulation 1987; 76:44-51.
- Wittes J, Lakatos E, Probstfield J. Surrogate endpoints in clinical trials: Cardiovascular diseases. Stat Med. 1989;8:415–425.
- Losordo DA, Schatz RA, White CJ, et al. Intramyocardial Transplantation of Autologous CD34+ Stem Cells for Intractable Angina. Circulation 2007; 115:3165-3172.
- Beeres SL, Bengel FM, Bartunek J, et al. Role of imaging in cardiac stem cell therapy. J Am Coll Cardiol 2007; 49(11):1137-1148.
- Budoff MJ, Cohen MC, Garcia MJ, et al. ACCF/AHA clinical competence statement on cardiac imaging with computed tomography and magnetic resonance; a report of the American College of Cardiology Foundation/American Heart Association/American College of Physicians Task Force on Clinical Competence and Training. J Am Coll Cardiol 2005; 46:383-402.
- Sanz-Ruiz R, Ibanes EG, Arranz AV, et al. Phases I-III clinical trials using adult stem cells. Stem Cells International 2010:579 142, doi:4061/2010/579142.
- Bellenger N.G., Davies L.C., Francis J.M., Coats A.J., Pennell D.J.; Reduction in sample size for studies of remodeling in HF by the use of cardiovascular magnetic resonance. J Cardiovasc Magn Reson 2000; 2:271-278.
- Auerbach MA, Schoder H, Hoh C, et al. Prevalence of myocardial viability as detected by positron emission tomography in patients with ischemic cardiomyopathy. Circulation 1999; 99:2921-2926.
- Zannad F, Garcia AA, Anker SD et al. Clinical endpoints in HF trials: a European Society of Cardiology HF Association consensus document. Eur J Heart Fail 2013:15:1082-1094.
- Dargie H, Colluci W, Ford H, et al. Effect of carvedilol on outcome after myocardial infarction in patients with left-ventricular dysfunction: the CAPRICORN randomised trial. Lancet 2001, 357(9266): 1385 – 1390.
- Dahlöf B, Devereux RB, Kjeldsen SE, et al. Cardiovascular morbidity and mortality in the Losartan Intervention For Endpoint reduction in hypertension study (LIFE): a randomised trial against atenolol. Lancet. 2002; 359(9311):995-1003.
- Behfar A, Crespo-Diaz R, Terzic A, Gersch B. Cell therapy for cardiac repair—lessons from clinical trials. Nature Reviews Cardiology 2014; 11:232-246.