Unstable Angina as a Component of Primary Composite Endpoints in Clinical Cardiovascular Trials: Pros and Cons

Background: Unstable angina (UA) is a component of acute coronary syndrome that is only occasionally included in primary composite endpoints in clinical cardiovascular trials. The aim of this paper is to elucidate the potential benefits and disadvantages of including UA in such contexts. Summary: UA comprises <10% of patients with acute coronary syndromes in contemporary settings. Based on the pathophysiological similarities, it is ideal as a part of a composite endpoint along with myocardial infarction (MI). By adding UA as a component of a primary composite endpoint, the number of events and feasibility of the trial should increase, thus decreasing its size and cost. Furthermore, UA has both economic and quality of life implications on a societal and an individual level. However, there are important challenges associated with the use of UA as an endpoint. With the introduction of high-sensitivity troponins, the number of individuals diagnosed with UA has decreased to rather low levels, with a reciprocal increase in the number of MI. In addition, UA is particularly challenging to define given the subjective assessment of the index symptoms, rendering a high risk of bias. To minimize bias, strict criteria are warranted, and events should be adjudicated by a blinded endpoint adjudication committee. Key Messages: UA should only be chosen as a component of a primary composite endpoint in cardiovascular trials after thoroughly evaluating the pros and cons. If it is chosen to include UA, appropriate precautions should be taken to minimize possible bias.


Introduction
Unstable angina (UA), non-ST-segment elevation myocardial infarction (NSTEMI), and ST-segment elevation myocardial infarction (STEMI) comprise the three main presentations of acute coronary syndromes This article is licensed under the Creative Commons Attribution 4.0 International License (CC BY) (http://www.karger.com/Services/ OpenAccessLicense). Usage, derivative works and distribution are permitted provided that proper credit is given to the author and the original publisher.

Pros Cons
With the introduction of high-sensitivity troponins and use of strict criteria, the number of patients diagnosed with unstable angina has decreased, leading to low event rates in clinical trials To account for discordance hospitalization for unstable angina should be adjudicated with strict criteria by a blinded endpoint adjudication committee, increasing costs and logistic challenges A diagnosis of unstable angina comes with high risk of ascertainment and reporting bias and limited information on symptom description are often available Unstable angina shares pathophysiological similarities with myocardial infarction and belongs in the spectrum of acute coronary syndromes The addition of unstable angina to a primary composite endpoint increases overall event rate and enhances statistical efficiency, leading to increased feasibility of the trial Unstable angina affects quality of life and has economic implications

Epidemiology of UA
During the last decades, better risk factor control, evolving MI definitions with widespread use of high-sensitivity troponins, and reperfusion strategies have changed the epidemiology and prognosis of patients diagnosed with UA. The 2002 Euro Heart Survey of ACS based on data from 25 countries reported that UA comprised 42% of patients with an ACS [9]. Conversely, in the era of high-sensitivity troponins, a Danish cohort study found that less than 10% of patients with ACS were diagnosed with UA [10]. Similar results have been reported in other contemporary studies [11,12]. Interestingly, patients diagnosed with UA generally appear to be younger than those with NSTEMI but present with a higher prevalence of most cardiovascular risk factors and more advanced coronary artery disease (CAD) [11,13,14].
Patients with UA also appear to derive less benefit from intensified antiplatelet therapy and an invasive strategy within 72 h, although invasive coronary angiography is still recommended in patients with a high likelihood of UA [3]. A contemporary Swedish registry-based study found that ∼88% of patients diagnosed with UA underwent invasive coronary angiography and ∼75% underwent revascularization [13]. In contrast, a contemporary German study found that coronary angiography was performed in ∼72%, but only ∼29% underwent revascularization [12]. The discrepant findings may reflect differences in the definition of UA as well as differing practices related to referral to invasive assessment, which may again complicate the use of UA as a component of a primary composite endpoint.

Prognosis of Patients with UA
Generally, the risks of subsequent death and cardiovascular events are lower in the UA setting compared with NSTEMI [3,13]. For example, one report based on two independent prospective multicenter studies (n = 8,992) found that the rate of incident nonfatal MI was similar in patients with UA and NSTEMI, but all-cause mortality was considerably lower in those with UA [11]. Another study comparing outcomes of percutaneous coronary intervention-treated patients with UA, stable angina, and NSTEMI (n = 7,187, 38% with a diagnosis of UA) also reported substantially better outcomes in patients with UA compared with NSTEMI [14]. Nevertheless, after adjustment for baseline differences, event risks were found to be comparable. According to a registrybased study including 3,204 patients with UA, 6.3% of the patients died within the first year after the diagnosis [13].

UA in the Era of High-Sensitivity Troponin Assays
High-sensitivity cardiac troponin assays have improved sensitivity and precision, enabling more rapid and accurate diagnosis in patients with suspected MI [15,16]. The assays have superior precision at the 99th percentile and enable these to be more accurately defined. Accordingly, the introduction of high-sensitivity cardiac troponins has altered the epidemiology of UA and MI [17]. Considering unselected patients presenting with suspected NSTE-ACS, the use of such assays has enabled the identification of smaller MI including type 2 MI and the incidence of MI has increased (4% absolute and 20% relative increase) with a reciprocal decrease in UA [18]. It is notable however, that the general incidence of MI is declining [19]. With this decrease in the incidence of UA, it has been argued that UA no longer belongs within the clinical spectrum of ACS but could be viewed as a subgroup of severe stable CAD, serving as an argument for  [11,20,21]. Conversely, a contemporary study found that the risk of death in UA patients was higher than in stable angina (10.5% of UA patients died within the first 3 years vs. 7.5% in patients with stable angina) [14]. Serial troponin testing with exclusion of a significant rise or fall in its concentration is key as some patients with UA may present with chronically elevated high-sensitivity cardiac troponin levels, e.g., patients with known chronic kidney disease, chronic coronary syndrome, or heart failure. Still, it can be difficult to distinguish UA from NSTEMI, e.g., in very late presenters of MI, where the troponin release curve has flattened. As opposed to the aforementioned position on grouping UA as a subgroup of stable CAD, other studies have stated that UA and NSTEMI should be grouped together and patients should be managed depending on their cardiovascular risk factors instead of the diagnosis [12,13]. This position supports the use of UA as an endpoint in clinical cardiovascular trials.

Composite Endpoints in Cardiovascular Clinical Trials
The ideal endpoint of a clinical trial is based on several considerations. It should be well-defined, easy to evaluate, clinically relevant to the intervention and the patient, and have a low risk of ascertainment and reporting bias. In addition, it should occur at a suitable frequency within a limited time frame for a trial to be feasible. All-cause mortality has fulfilled all these criteria. However, a declining mortality rate in the general population, including a declining proportion of mortality caused by cardiovascular disease, makes all-cause mortality less sensitive to interventions that primarily exert their beneficial effects through cardiovascular outcomes [22,23]. Improved population health and novel treatments have also reduced cardiovascular event rates. Consequently, a single-component primary endpoint is likely to result in an unreasonably long follow-up period and/or large sample size. Therefore, composite endpoints are almost exclusively used in contemporary trials. A composite endpoint increases the total number of events, statistical power, and precision. Such endpoints are particularly valuable if the intervention has a similar beneficial effect on outcomes of equal importance, and if they share a common pathophysiological mechanism such as MI and UA. Nevertheless, individual tailoring of composite end-points may lead to differing results and conclusions and potentially hamper proper comparisons of trials.
The differences in the event rate of each component of a composite endpoint must be as small as possible since the overall effect of an intervention will largely be determined by the dominant event [24,25]. Several nonfatal endpoints like MI, UA, and coronary revascularization have a higher incidence (and thus by default occur earlier) than more fatal endpoints as death or resuscitated cardiac arrest. Both the incidence and severity of the components of the combined endpoints need consideration. Methods have been proposed to weight the components of a composite endpoint to counter these issues [26,27]. Attempts to quantify weights have been based on expert judgement (i.e., clinical-investigator Delphi panels), preference methods (decisions made by individuals when confronted with different scenarios) and disability-adjusted life years (DALY) lost [28]. However, the process of assigning weights is not standardized in randomized clinical trials, can be time consuming, and may be costly.
A commonly used composite endpoint in cardiovascular trials is major adverse cardiovascular events (MACE). Although it has not been defined strictly, it usually comprises a composite of death from cardiovascular causes, MI, and stroke [29]. Indeed, in 2008, the US Food and Drug Administration (FDA) decided to recommend assessment of the cardiovascular safety profile of antidiabetic agents using this three-point MACE but stated that hospitalization for ACS (UA), urgent revascularization, and possibly other endpoints could be added [30]. An example of a trial adding UA and urgent revascularization to the traditional 3-point MACE for its primary composite endpoint was the Thrombin Receptor Antagonist for Clinical Event Reduction in Acute Coronary Syndrome (TRACER) trial [31]. The key secondary endpoint of the trial was the traditional 3-point MACE, consisting of a composite of death from cardiovascular causes, MI, or stroke. Curiously, it was found that the occurrence of the primary endpoint did not reach significance, whereas the secondary did.

The Health Care Burden of UA
As seen in Figure 1, one argument for expanding a composite endpoint with UA would be to gain a broader perspective of the symptom burden and the consequences for the individual and society. The economic burden of UA is almost as high as that of MI during the first year after diagnosis [32]. Additionally, a study found high hos- pitalization rates for both UA (∼89%) and NSTEMI (∼99%) [12]. On the patient level, UA negatively affects quality of life [33,34].

UA in Clinical Trials
An important caveat of using UA as an endpoint in cardiovascular trials is the extent of ascertainment and reporting bias, particularly with respect to symptom description. The clinical distinction between UA, NSTEMI, and noncardiac chest pain can be challenging, even when applying current guideline recommendations. Limited information on symptom description and duration of ischemic symptoms may lead to refuted events due to a lack of event-specific data [13]. Compared with MI, clinical subjectivity plays a greater role in its ascertainment, and interpretations may differ markedly between clinicians. A registry-based study found that the proportions of UA among those with NSTE-ACS ranged from 4% to 22% in the 49 different participating hospitals, suggesting variations in the local perception of UA [13]. Therefore, it is essential that cardiovascular trials have a clear, operational definition of UA with strict and objective criteria.  [30] Criteria for adjudication of hospitalization for UA UA requiring hospitalization is defined as 1. Ischemic discomfort (angina, or symptoms thought to be equivalent) ≥10 min in duration occurring at rest or in an accelerating pattern with frequent episodes associated with progressively decreased exercise capacity. AND 2. Prompting an unscheduled hospitalization within 24 hours of the most recent symptoms. Hospitalization is defined as an admission to an inpatient unit or a visit to an emergency department that results in at least a 24-hours stay (or a change in calendar date if the hospital admission or discharge times are not available). . and believed to be responsible for the myocardial ischemic symptoms/signs. c. Angiographic evidence of new or worse ≥70% lesion (≥50% for left main lesion) and/or thrombus in an epicardial coronary artery that is believed to be responsible for the myocardial ischemic symptoms/signs. d. Need for coronary revascularization procedure (PCI or CABG) for the presumed culprit lesion(s), as defined in 3c. This criterion would be fulfilled if revascularization was undertaken during the unscheduled hospitalization, or subsequent to transfer to another institution without interceding home discharge.    Table 2).
Use of an adjudication committee: yes  In 2017, the Standardized Data Collection for Cardiovascular Trials Initiative established by the FDA proposed definitions for cardiovascular endpoints including hospitalization for UA to simplify the design and conduct of clinical trials and to enhance the ability to aggregate and compare data from multiple trials [35]. The full criteria for hospitalization for UA are presented in Table 2. In brief, the criteria consist of 4 elements: (1) ischemic symptoms with a duration >10 min at rest or in an accelerating pattern; (2) prompting an unscheduled hospitalization within 24 hours of the most recent symptoms; (3) absence of cardiac biomarker elevation; and (4) one of the following: ECG changes, evidence of inducible myocardial ischemia, angiographic evidence of CAD believed to be responsible for the ischemic symptoms or signs, or need for a coronary revascularization procedure.
An example of a trial using these criteria is the Reduction of Cardiovascular Events with Icosapent Ethyl-Intervention Trial (REDUCE-IT), which randomized statintreated patients with elevated triglycerides to icosapent ethyl or placebo [36]. The composite endpoint was cardiovascular death, MI, stroke, coronary revascularization, or UA. When comparing the concordance between investigator-reported and adjudicated primary endpoints, it was ∼100% for death and coronary revascularization, 80-90% for MI and stroke, but only 53% for UA [37]. Investigator-reported events exceeded adjudicated rates of UA (283 vs. 108), often at the expense of underreporting MI. Other studies have reached the same conclusions. One study compared local interpretation and blinded evaluation of the admission ECG in patients with UA and non-Q-wave MI and found considerable differences [38]. Another study evaluated reported and adjudicated events in the Stabilization of Atherosclerotic Plaque by Initiation of Darapladib Therapy (STABILITY) trial [39]. It was found that 25.4% of the 1,407 reported UA events were classified as such by the adjudication committee, primarily due to crossover between MI and UA. Indeed, the risk of misinterpretation and heterogeneity with UA is particularly high in large, multinational randomized controlled trials in which endpoints are interpreted by several local investigators from different hospitals. The large degree of subjectivity associated with this diagnosis may lead to variation in adjudication. One way of maximizing concordance for a complex endpoint like UA would be to exclusively use a blinded endpoint adjudication committee. However, the use of a clinical endpoint adjudication committee increases the costs and logistic challenges of clinical trials. Recent evidence has challenged the unquestioned role of an endpoint adjudi- CI: 5-26%, p = 0.005) UA was defined as ischemic discomfort at rest for at least 10 min prompting rehospitalization, combined with one of the following: ST-segment or T-wave changes, cardiac-marker elevations that were above the upper limit of normal but did not meet the criteria for MI, or a second episode of ischemic chest discomfort lasting more than 10 min and that was distinct from the episode that had prompted hospitalization.
Use of an adjudication committee: not specified  cation committee with the use of other components of a primary endpoint than UA [40][41][42]. Consequently, the choice of UA as a primary endpoint might not be favorable, as the use of UA leads to an ubiquitous need for an adjudication committee.  Table 3. The International Study of Comparative Health Effectiveness with Medical and Invasive Approaches (IS-CHEMIA) is an example of a trial expanding the primary endpoint by including UA in an attempt to increase feasibility [43]. Interestingly, the original grant application had described a five-component primary endpoint consisting of cardiovascular death, nonfatal MI, resuscitated cardiac arrest, or hospitalization for UA or heart failure. Subsequently, the steering committee was allowed to alter this endpoint to cardiovascular death or MI, but with a contingency plan to switch back to the original endpoint to retain power in case of a lower-than-expected event rate [51]. Indeed, it is not uncommon for clinical trials to have lower than anticipated event rates. The specific reasons for the inclusion of hospitalization for UA were clinical relevance, quality of life and economic implications, and potential benefits of revascularization. A strict definition of UA resembling the one proposed by the Standardized Data Collection for Cardiovascular Trials Initiative was employed. However, despite the fact that high-sensitivity troponins were not systematically used by all trial sites, relatively few UA events were reported (48 UA events vs. 326 events with nonprocedural MI). Several trials using UA as a component of the primary endpoint have been faced with similar issues as presented in Table 3.

Clinical Trials Using Urgent Revascularization as a Component of the Primary Endpoint: The FAME-2 Example
An alternative approach might be to use revascularization, as UA is closely linked to urgent revascularization. As an endpoint, revascularization is clinically feasible and pathophysiologically relevant. It is easy to evaluate, has a higher event rate than UA, and the use of an adjudication committee may not be as essential. Examples of clinical trials using revascularization as a component of a primary endpoint are presented in Table 4. The Fractional Flow Reserve-Guided PCI versus Medical Therapy in Stable Coronary Disease (FAME-2) trial included unplanned hospitalization leading to urgent revascularization as part of the primary composite endpoint consisting otherwise of death from any cause or nonfatal MI [57]. This adjudicated endpoint was counted only if patients were hospitalized unexpectedly because of persisting or increasing complaints of chest pain (with or without ST-T changes) and underwent a revascularization procedure during the same hospitalization. This definition has substantial similarities to the definition of UA as described previously, but may be more objective. The results of the trial at both 7 months and 5 years were driven mainly by urgent revascularization [58,59]. The definition of urgent revascularization as well as the fact that it resulted in the premature termination of FAME-2 has been questioned [60]. Furthermore, unplanned and planned revascularization as endpoints in clinical trials may face challenges with respect to subjectivity. Indeed, the decision on revascularization as well as whether to classify it as planned or unplanned are influenced by many factors such as local practice, available resources, and geographical location. Especially in nonblinded trials, there is concern for selection bias as investigators may be more willing to recommend revascularization for patients in one of the treatment groups rather than managing the symptoms noninvasively.

UA as a Registry-Based Endpoint
Recently, there has been an increasing interest in randomized controlled trials with registry-based follow-up as they are more feasible, cheaper, and allow for longerterm follow-up [61]. The diagnosis of UA in registries is more likely to be independent of the trial and possible bias balanced in the intervention and control group. A systematic review of the validity of acute cardiovascular outcome diagnoses recorded in European electronic health records found that the positive predictive value (PPV) of UA varied considerably [62]. PPV as low as 20-30% [63,64] to as high as 88% have been reported [65]. Conversely, a higher PPV has consistently been reported for MI, at >90% [62]. Because of considerable betweenstudy heterogeneity, the use of UA as a registry-based endpoint in clinical trials should be done very cautiously and after proper validation. One possible way to increase DOI: 10.1159/000524948 specificity could be to define UA based on hospitalization and urgent invasive coronary angiography or revascularization.

Conclusions and Recommendations
This review summarizes the challenges and opportunities associated with using UA as a component of a primary composite endpoint in cardiovascular clinical trials. Arguments favoring the use of UA rely on its pathophysiological similarities to MI, the ability to enhance statistical power, the impact on quality of life, and the health care burden. However, widespread use of high-sensitivity troponins has reduced the incidence of UA. UA may also be subject to ascertainment bias, reporting bias, and discordance between clinicians. Therefore, we recommend that UA events should be adjudicated with strict criteria by a blinded endpoint adjudication committee, or alternatively be defined in combination with revascularization, resulting in a limited number of events. If UA is considered as an endpoint in clinical trials, we recommend evaluating our list of pros and cons in Figure 1 and the practical guidance and recommendations in Table 1. The use of UA as a component of a primary composite endpoint should be done with caution and after thorough consideration.