Intervention for Non–ST-Segment Elevation Acute Coronary Syndromes

Ischemic Risk Stratification

Several risk stratification tools have been developed that can be applied at the time of the diagnosis of NSTE-ACS, which are associated with short-term and long-term outcomes, including death, recurrent myocardial infarction (MI), and need for urgent revascularization. In addition to providing prognostic value, risk stratification may help guide the therapeutic decision-making process, including recommendations regarding the use of antithrombotic and antiplatelet agents, and the need for early revascularization. The importance of early risk stratification has also been highlighted in cardiovascular society guidelines, and currently has a class I recommendation in the 2015 European Society of Cardiology (ESC) guidelines and a class IIa recommendation in the 2014 American Heart Association (AHA) and American College of Cardiology (ACC) guidelines. ^, While many risk scores have already been well validated, each is limited by its individual drawbacks and should be interpreted in the context of these limitations.

The Thrombolysis in Myocardial Infarction (TIMI) risk score utilizes seven clinical variables and is designed to be easily calculated at the bedside to predict the 14-day risk of all-cause mortality, MI, or recurrent ischemia requiring revascularization (C-statistic = 0.65) ( Fig. 19.1 ). The variables are: age ≥ 65 years, at least three coronary artery disease (CAD) risk factors, known CAD with stenosis on angiography of at least 50%, use of aspirin in the past 7 days, severe angina with two or more episodes within 24 hours, ST-segment changes of at least 0.5 mm, and positive cardiac biomarker. The TIMI risk score was initially derived from cohorts of two phase 3 randomized trials for unfractionated heparin (UFH) versus enoxaparin in NSTE-ACS. The test cohort included patients randomized to the UFH group in the TIMI 11B trial ( n = 1957), whereas the three validation cohorts included the remainder of patients in the merged data; that is, patients randomized to the enoxaparin group in the TIMI 11B trial ( n = 1953) and the two treatment groups in the ESSENCE trial (Efficacy and Safety of Subcutaneous Enoxaparin in Unstable Angina and Non–Q-Wave MI) ( n = 3171). The presence of each risk factor corresponds to 1 point, and the simple arithmetic sum of the number of variables present correlates with clinical outcomes. Accordingly, a TIMI risk score of 0 correlates with a 4.7% event rate through 14 days, whereas a ten-fold gradient event rate (40.9%) occurs among patients with a score of 6 to 7.

Following its internal validation, the TIMI score was then applied to other randomized clinical trials, and externally validated in other real-world registries for long-term outcomes occurring through 30 days, 6 weeks, and 12 months. In the platelet receptor inhibition for ischemic syndrome management in patients limited by unstable signs and symptoms (PRISM-PLUS) trial for tirofiban ( n = 11,915), the TIMI risk score demonstrated a strong prognostic discriminatory capacity in identifying patients who were more likely to benefit from more aggressive pharmacologic therapy using platelet glycoprotein (GP) IIb/IIIa inhibitors in addition to heparin and aspirin. The TIMI risk score modified the treatment effect, where patients with a TIMI risk score of 4 or greater were at higher risk of developing adverse events, and had a greater relative risk reduction as compared with those with lower TIMI risk scores. Similarly, in the TACTICS-TIMI 18 trial for tirofiban (Treat Angina with Aggrastat and Determine Cost of Therapy with an Invasive or Conservative Strategy), NSTE-ACS patients with either intermediate risk (TIMI score of 3 or 4) or high risk (TIMI score of 5 to 7) derived a significant 30-day benefit from the use of early invasive strategy for revascularization, whereas low-risk patients (TIMI risk score of 1 or 2) did not.

In a similar fashion to the TIMI risk score, the PURSUIT risk score was developed using a cohort of patients who were randomized in the PURSUIT trial for eptifibatide (Platelet glycoprotein IIb/IIIa in UA: Receptor Suppression Using Integrilin (eptifibatide) Therapy) ( n = 9461). The risk models in PURSUIT identified seven patient characteristics that were associated with either all-cause death alone (C-statistic = 0.81) or the composite of all-cause death or MI (C-statistic = 0.67) through 30 days following hospital admission. These characteristics included patient age, gender, highest Canadian Cardiovascular Society (CCS)-class in the previous 6 weeks prior to admission, heart rate, systolic blood pressure, signs of heart failure, and ST-segment depression on presenting electrocardiogram (ECG). Points were given for each predictive factor that was present, which were then added to provide a final risk score. This score was then plotted on a graph that illustrated the variation of 30-day mortality (y-axis) as a function of the risk score (x-axis). Interestingly, the score relied heavily on patient age (i.e., age contributes to 8 to 14 points out of the total 20 points), whereas baseline biomarkers were not included.

In contrast to both TIMI and PURSUIT risk scores, the Global Registry of Acute Coronary Events (GRACE) risk score was developed using the patient cohort included in the GRACE registry ( n = 43,810) ( Fig. 19.2 ). The GRACE risk score models predicted in-hospital risk, as well as the cumulative 6-month risk, of the same end points studied in the PURSUIT risk score: either all-cause death alone (C-statistic = 0.81) or the composite of all-cause death or MI (C-statistic = 0.73). The score was then validated using the cohort of patients randomized in the GUSTO IIb trial (Global Use of Strategies to Open Occluded Open Arteries) ( n = 12,142). However, it was thought that the original risk models were too complex, consisting of 14 variables for the in-hospital all-cause death outcome and 12 variables for the composite outcome. Due to their complexity, a simplified nomogram was eventually developed that reduced the score to the most substantial eight variables that provided the highest predictive capacity for both the in-hospital and 6-month outcomes. These variables included patient age, heart rate, systolic blood pressure, serum creatinine, Killip class for heart failure, cardiac arrest at admission, ST-segment deviation on presenting ECG, and baseline elevation in cardiac biomarkers.

Afterwards, the GRACE 2.0 score was developed to evaluate risk of outcomes through longer periods of time, that is, 1 year and 3 years, and was then validated externally against the French Registry on Acute ST-Elevation and Non–ST-Elevation Myocardial Infarction (FAST-MI) registry. Compared with the initial GRACE score, the GRACE 2.0 score excluded serum creatinine and Killip class, which might not be readily available at the time of clinical evaluation, and replaced them with renal failure and diuretic use, respectively. More importantly, GRACE 2.0 incorporated non linear statistical models that could estimate risks more accurately than the original GRACE models. Notably, the robustness of the GRACE risk score remained strong even when outcomes were evaluated at the 5-year follow-up mark (C-statistic = 0.77 for all-cause mortality).

More recently, the CRUSADE (Can Rapid Stratification of Unstable Angina Patients Suppress Adverse Outcomes with Early Implementation of the ACC/AHA Guidelines) risk score was developed using the CRUSADE registry data. The patient population specifically included older NSTE-ACS patients of at least 65 years of age, and risk modeling was performed based on long-term outcomes through 1, 2, and 3 years ( n = 43,239). In the validation cohort, a total of 13 baseline clinical variables were independently associated with the outcome of all-cause mortality through 1 year and were further evaluated in the validation cohort (C-statistic = 0.75). Variables included patient age, gender, heart rate, systolic blood pressure, weight, signs of heart failure on presentation, prior history of heart failure, history of stroke, history of diabetes mellitus, history of peripheral arterial disease, hematocrit, serum creatinine, and troponin on admission. The CRUSADE models identified age as the most significant predictor of mortality, a finding previously reported in the PURSUIT risk score. However, when compared to the 5-year GRACE cohort, the CRUSADE population was significantly older (mean age of 77 years in CRUSADE vs. 67 years in the GRACE). The advantages of the CRUSADE risk score is its strong discriminatory capacity; nonetheless, it is newer than the TIMI, PURSUIT, and GRACE scores and is yet to be validated as extensively as the other scores in external cohorts.

When reconciling all these risk scores, it is evident that they are not only clinically useful, but also suffer from substantial limitations, as each has been individually criticized for fundamental design, statistical, and clinical pitfalls. Their predictive power is moderately good, where the C-statistic is reported to be low as 0.65 when the internal validation cohorts were evaluated, and occasionally lower with other external validations. The positive predictive value of these scores may also be modest. Despite their limitations, these tools—particularly the TIMI and GRACE scores—have withstood the test of time and proven useful when extensively validated in real-world observational studies, as well as when used as stratification methods for investigational therapies in randomized clinical trials. Nonetheless, novel assessment models are under development in an effort to overcome these limitations, but these scores have so far been based on modest sample sizes and, albeit promising, will require extensive validation prior to becoming clinically useful.

Bleeding Risk Stratification

Although patients with NSTE-ACS benefit from antithrombotic therapy at the time of diagnosis, antiplatelet and anticoagulant therapies are associated with an increased risk of bleeding. Balancing the risk of ischemia and the risk of bleeding is challenging, and there are no prediction models that simultaneously incorporate the two. Furthermore, bleeding scores have not been as extensively employed in the stratification of patients in clinical trials as ischemic scores. However, baseline prediction of the bleeding risk may complement the estimation of ischemic risk and may help guide a balanced approach to the risks and benefits of adjunctive pharmacologic therapies in NSTE-ACS.

The CRUSADE bleeding score was developed among 71,277 patients enrolled in the CRUSADE quality improvement initiative. It is comprised of eight clinical variables that include baseline hematocrit, creatinine clearance, gender, signs of congestive heart failure on presentation, history of diabetes mellitus, history of prior vascular disease, heart rate, and systolic blood pressure (C-statistic = 0.71). The score quantifies risk for in-hospital major bleeding and categorizes patients by quintiles of bleeding risk ranging from very low risk (3.1% risk of bleeding) to very high risk (19.5% risk of bleeding). Remarkably, all eight variables in the CRUSADE bleeding score have been included in ischemic risk scores, highlighting that patients with high bleeding risk are also those with high risk of ischemia.

Immediate Angiography

Immediate angiography within 2 hours is recommended for patients who present with at least one of the following features: hemodynamic instability due to either cardiogenic shock or mechanical complication of the MI event, overt heart failure, persistent angina despite pharmacologic therapy, or unstable ventricular arrhythmias. There are no randomized trials that compare a strategy of immediate angiography to other strategies in this subgroup of patients, and these recommendations are therefore based on expert opinion. Overall, the prognosis in these clinical scenarios is thought to be poor; and accordingly, these patients would likely benefit from immediate revascularization if they have no other severe life-threatening co-morbidities that would preclude either surgical or endovascular intervention.

Invasive Versus Ischemia-Guided Strategy

An early invasive strategy for revascularization is favored among patients with NSTE-ACS. Compared with an ischemia-guided strategy, an early invasive strategy has been associated with a reduced rate of recurrent ischemic events. Historically, prior to the development of novel antithrombotic therapies and new devices, trials comparing invasive versus ischemia-guided strategies showed mixed, inconclusive results. The FRISC-II study (Fragmin and Fast Revascularization during Instability in Coronary Artery Disease) was the first randomized trial to compare an early invasive strategy to ischemia-guided strategy in the era of stenting and GP IIb/IIIa inhibitors. An early invasive strategy within 7 to 10 days significantly reduced the composite of all-cause death or MI through 6 months (9.4% vs. 12.1%, P = .031) and 1 year (10.4% vs. 14.1%, P = .005) ( n = 2457) as compared with a conservative strategy of heparin administration for 3 months. Although the 6-month analysis demonstrated no reduction in death at 6 months in the overall population, there was a significant 43% relative risk reduction in the rate of death when comparison was made at 1 year. Further subgroup analysis also demonstrated a gender effect, where men derived the most benefit with an invasive strategy, an effect that was not observed among women ( P -interaction = .012 for the composite end point of death and MI, and P -interaction = .029 for the end point of death alone).

More recently, additional pharmacologic and mechanical interventions have been introduced; since the introduction of the strategies, the balance between an invasive and ischemia-guided strategy has increasingly favored early intervention. In the TACTICS-TIMI-18 randomized trial (Treat Angina with Aggrastat and Determine Cost of Therapy with an Invasive or Conservative Strategy–Thrombolysis in Myocardial Infarction 18), NSTE-ACS patients were randomized to either early invasive strategy within 4 to 48 hours or conservative strategy ( n = 2220). Only high-risk patients treated with the GP IIb/IIIa inhibitor tirofiban who also had either ST-segment or T-wave changes, elevated biomarkers, or history of CAD were enrolled. Compared with conservative strategy, early invasive strategy was associated with a significant reduction in ischemic outcomes through 6 months (15.9% vs. 19.4%, P = .025). Similarly, the rates of the composite of all-cause death or nonfatal MI were significantly lower with the invasive strategy when evaluated at 30 days (4.7% vs. 7.0%) and at 6 months (7.3% vs. 9.5%).

As the definition of MI evolved, the relevance of TACTICS-TIMI-18 was questioned given its use of earlier definitions of MI. Subsequently, the RITA-3 trial (third Randomized Intervention Treatment of Angina) randomized NSTE-ACS patients to either early intervention or conservative strategy ( n = 1810), but the definition of MI was consistent with the new third universal definition. Early intervention in RITA-3 was likewise associated with reduction in the primary end point of all-cause death, nonfatal MI, or refractory angina through 4 months compared with ischemia-guided strategy (9.6% vs. 14.5%, 0.66, 95% confidence interval [CI] 0.51-0.85; P = .001). This reduction was driven mainly by differences in rates of refractory angina.

In contrast to other contemporary trials, the ICTUS trial (Invasive Versus Conservative Treatment in Unstable Coronary Syndromes) showed no benefit with routine invasive strategy ( n = 1200). ICTUS enriched for high-risk subjects and enrolled only NSTE-ACS patients with positive troponin and either a history of CAD or new ischemic ECG changes. Patients were randomized to either an early routine invasive strategy or a selective invasive strategy (i.e., medical management and subsequent revascularization only if persistent symptoms or life-threatening features). Both strategies had similar rates of the primary end point of all-cause death, nonfatal MI, or rehospitalization for anginal symptoms at 1 year. Although an early invasive strategy was associated with a significantly reduced rate of rehospitalization (7.4% vs. 10.9%, P = .04), this was offset by a higher rate of nonfatal MIs (15.0% vs.10.0%, P = .005), particularly in the peri-procedural timeframe. This high rate of MI was attributed to the protocol’s definition of MI, which used a very low threshold for MI using creatine-kinase myocardial bands (CK-MB), as well as a substantially high sampling frequency for CK-MB of every 6 hours postpercutaneous coronary intervention (PCI). Although this definition was consistent with the contemporary society guidelines at the time, it was different from the definitions used in any of the other trials.

Unfortunately, individual trials lacked statistical power to demonstrate efficacy. In a pooled meta-analysis of FRISC-II, ICTUS, and RITA-3 ( n = 5467), a routine invasive strategy was associated with 19% reduced risk of the composite of cardiovascular (CV) death or nonfatal MI as compared with selective invasive strategy through 5 years (14.7% vs. 17.9%, hazard ratio [HR] = 0.81, 95% CI 0.71 to 0.93; P = .002). This reduction was primarily driven by a 23% reduced rate of nonfatal MIs (10.0% vs. 12.9%), with numerical trends also favoring routine invasive strategy for CV death and all-cause death. The benefit of an early invasive strategy did not attenuate over time among patients followed through 5 years. In another pooled analysis of seven randomized trials including 8375 NSTE-ACS patients, the early invasive strategy was similarly associated with a 25% relative risk reduction for all-cause death (4.9% vs. 6.5%, HR = 0.75, 95% CI 0.63 to 0.90; P = .001) and 17% for nonfatal MIs (7.6% vs. 9.1%, HR = 0.83, 95% CI 0.72 to 0.96; P = .012) through 1 year.

Among NSTE-ASC patients, an early invasive strategy is generally preferred over an ischemia-guided strategy. High-risk patients, namely patients with troponin elevation and ST-segment change, are the subgroup of patients most likely to benefit from an early invasive strategy.

Timing of the Intervention: Early Versus Delayed Invasive Strategy

While it was postulated that early angiography leads to early revascularization and resolution of tissue ischemia, delayed angiography beyond 24 hours was thought to result in stronger plaque stabilization, the onset of pharmacologic efficacy of orally administered antiplatelet agents, and a reduced rate of periprocedural thromboembolic events. Results from smaller trials were inconsistent, and data from larger trials were not helpful as the median timing of the PCI in these trials ranged widely between 19 hours in ICTUS and 96 hours in FRISC-II. In small trials, namely VINO (Value of First Day Coronary Angiography/Angioplasty in Evolving Non-ST Segment Elevation Myocardial Infarction) ( n = 131), ISAR-COOL (Intracoronary Stenting With Antithrombotic Regimen Cooling Off among NSTE-ACS patients) ( n = 410), and RIDDLE-NSTEMI (Randomized study of Immediate versus Delayed Invasive Intervention in patients with NSTEMI) ( n = 323), an early invasive strategy, when performed within 2 to 24 hours, was associated with a significant reduction in the rate of all-cause death or recurrent MI as compared with a delayed strategy. In contrast, there was an increased rate of MI with immediate PCI in the OPTIMA trial ( n = 142), and no benefit with early angiography when performed within 24 hours in ABOARD (Angioplasty to Blunt the Rise of Troponin in Acute Coronary Syndromes Randomized for an Immediate or Delayed Intervention) ( n = 352), enzyme-linked immunosorbent assay (ELISA) ( n = 220), ELISA-3 ( n = 542), and LIPSIA-NSTEMI (Leipzig Immediate versus early and late PercutaneouS coronary Intervention triAl in NSTEMI) ( n = 401).

All these trials were primarily limited by their small sample size, low event rates, and occasionally the use of surrogate measures assess clinical outcomes. The comparison between early intervention within 24 hours and delayed intervention beyond 36 hours was eventually performed in the adequately powered, large-scale TIMACS randomized trial (The Timing of Intervention in Acute Coronary Syndromes) ( n = 3031). Early intervention did not reduce the primary end point of all-cause death, MI, or stroke, but was associated with a 28% relative risk reduction in the secondary outcome of death, MI, or refractory ischemia (9.5% vs. 12.9%, HR = 0.72, 95% CI 0.58 to 0.89; P = .003) and also appeared to be particularly superior among high-risk patients who had a GRACE score of more than 140 points on admission.

Finally, a meta-analysis of eight randomized trials, including TIMACS, demonstrated that early intervention was associated with a numerical reduction in mortality, but this benefit did not reach statistical significance (HR = 0.81, 95% CI 0.64 to 1.03; P = .088). Patients with elevated biomarkers at baseline, history of diabetes, age of more than 75 years, and high GRACE score were identified as subgroups who derived significant benefit with an early invasive strategy within 24 hours.

In conclusion, based on adequately sized randomized trials and meta-analyses of smaller trials, an early invasive strategy, preferably within 24 hours of diagnosis, is favored. This hypothesis may likely require periodic reevaluation with the introduction of earlier and more potent pharmacologic interventions in the future.

Antiplatelet Therapy