Distal Microcirculatory Protection During Percutaneous Coronary Intervention in Acute ST-Segment Elevation Myocardial Infarction: Title and subTitle BreakA Randomized Controlled Trial FREE

Myocardial salvage is frequently suboptimal after percutaneous coronary intervention (PCI) in acute myocardial infarction (AMI), and mortality remains excessive in high-risk subgroups.^{1 - 4} Despite normal epicardial Thrombolysis in Myocardial Infarction (TIMI) flow after primary PCI, myocardial perfusion (as assessed either by angiographic capillary opacification [myocardial blush] or electrocardiographic ST-segment resolution [STR]) is abnormal in a significant proportion of patients, contributing to increased infarct size and reduced survival.^{5 - 15} Recent studies have emphasized the ubiquitous occurrence of distal embolization of atheromatous and thrombotic debris after primary PCI, potentially resulting in microcirculatory dysfunction, abnormal myocardial metabolism, and increased myonecrosis.^{16 - 18} Pilot studies and small randomized trials have demonstrated that distal protection of the microcirculation during primary PCI retrieves embolic debris in more than 70% of patients and may result in greater rates of normal TIMI flow and blush, complete STR, improved left ventricular function, and enhanced event-free survival compared with either historical or concurrent controls.^{19 - 25} We therefore performed a large-scale pivotal trial to evaluate the usefulness of distal microcirculatory protection during mechanical reperfusion therapy in AMI.

The Enhanced Myocardial Efficacy and Recovery by Aspiration of Liberated Debris (EMERALD) trial was a prospective, randomized, multicenter trial of distal protection during angioplasty in AMI (Figure 1). Investigators were selected after demonstration of proficiency in a pilot-phase study.²⁶ Individuals eligible for enrollment were consecutive patients aged 18 years or older with AMI presentation more than 30 minutes but less than 6 hours after symptom onset, with 2 mm or more of ST-segment elevation in 2 or more contiguous leads or with presumably new left bundle-branch block, in whom primary or rescue (ie, after failed thrombolysis) PCI was intended. Principal clinical exclusion criteria included major surgery or active bleeding within 6 weeks; aspirin, thienopyridine, or heparin allergy; neutropenia (<1000 neutrophils/mm³), thrombocytopenia (<100 000 platelets/mm³), hepatic dysfunction, or renal insufficiency (serum creatinine level >2.5 mg/dL [221 μmol/L]); cardiogenic shock; noncardiac condition with expected survival less than 1 year; and current participation in other investigations. The study was approved by the institutional review board at each center, and consecutive, eligible patients signed written informed consent.

Twenty-four–hour continuous ST-segment Holter monitoring was begun in the emergency department after consent was obtained. Aspirin, 324 mg (chewable); clopidogrel, 300 mg orally; heparin, 70 U/kg intravenous bolus; and intravenous β-blockers in the absence of contraindications were administered. Left ventriculography and coronary arteriography were performed to evaluate appropriateness for randomization to receive PCI with or without distal protection. Patients were eligible for randomization if PCI was indicated and if the vessel diameter at the infarct lesion was either known or expected to be 2.5 to 5.0 mm, without excessive tortuosity or lesion/vessel calcification, with 3 cm or more of distal vessel available to accommodate the device. Patients were excluded if multivessel intervention was required during the index procedure; if they had unprotected left main vessel disease; or if coronary artery bypass graft surgery would be required within 30 days.

Telephone randomization was performed in random blocks of 4 or 6 patients, stratified by the intention to use glycoprotein IIb/IIIa inhibitors and by primary vs rescue PCI. An activated clotting time of 300 seconds or greater was targeted before PCI (200-300 seconds if use of glycoprotein IIb/IIIa inhibitors was elected).

The GuardWire Plus (Medtronic Corp, Santa Rosa, Calif) consists of a 0.014-in guidewire incorporating a central inflation lumen distally attached to an elastomeric balloon (0.028-in crossing profile, 2.5- to 5.0-mm diameter range) which when inflated results in cessation of antegrade blood flow.^{27 - 28} Intervention is then performed over the wire, after which liberated debris suspended within a stagnant blood column is aspirated through the 5F monorail Export catheter. Operators were instructed to position the elastomeric balloon as close to the lesion as possible to minimize exposure of unprotected side branches to embolic debris. Aspiration prior to PCI was encouraged to diminish thrombus burden and improve distal visualization. All predilation and postdilation angioplasty balloon inflations, as well as stent implantation, were to be protected if possible. Thrombectomy devices were not permitted unless refractory thrombotic complications developed.

Cardiac creatine phosphokinase MB isoenzyme levels were measured at 8, 16, and 24 hours postprocedure. The Holter monitor was removed at 24 hours and sent to a core laboratory for analysis. Prespecified postprocedural medications consisted of aspirin, 325 mg by mouth daily indefinitely; clopidogrel, 75 mg by mouth daily for 4 weeks in patients with stents; oral β-blockers if not contraindicated; and angiotensin-converting enzyme inhibitors for patients with hypertension or left ventricular dysfunction. A resting technetium Tc 99m sestamibi scan was performed 5 to 14 days postprocedure, with raw data sent to a core nuclear laboratory. Clinical follow-up was performed at 1 and 6 months.

Independent on-site study monitors verified 100% of primary and secondary end point data. An independent data and safety monitoring board periodically assessed safety throughout the study, at each point recommending the study continue without change.

The EMERALD trial was designed to test the dual hypotheses that distal protection during mechanical reperfusion in AMI would result in enhanced reperfusion success, as evidenced by a greater rate of complete STR measured by continuous Holter monitoring 30 minutes after last contrast injection and by reduced infarct size as measured by Tc 99m sestamibi imaging. Secondary end points included final TIMI flow, corrected TIMI frame counts, and myocardial blush (after removal of all angioplasty equipment), and angiographic complications, analyzed at a core angiographic laboratory blinded to treatment assignment^{1 ,6 - 7} ; the 1- and 6-month composite and component rates of major adverse cardiac events related to left ventricular dysfunction (death, new-onset sustained hypotension, new-onset severe heart failure, and hospital readmission for left ventricular dysfunction); and the 1- and 6-month composite and component rates of major adverse cardiac events related to ischemia (death, reinfarction, ischemic target vessel revascularization, and disabling stroke).^{1 - 2}

ST-segment resolution was determined from continuous 24-hour digital 12-lead electrocardiograph recordings (Northeast Monitoring 180+, Boston, Mass). The instrumentation and core laboratory analysis methods used have been published previously.^{29 - 30} ST-segment levels were measured 60 ms after the J-point in the most abnormal lead. ST-segment recovery was determined by comparing electrocardiograms recorded at different time intervals after the last contrast injection with the most abnormal electrocardiogram in the data recorded between study enrollment and arrival in the catheterization laboratory. Changes during the interventional procedure were excluded from analysis. The percentage resolution of ST-segment elevation was categorized using the Schroder classification as complete (>70%), partial (30%-70%), or absent (<30%).^{31 - 32} The primary end point was prespecified as the dichotomous rate of complete STR measured 30 minutes after last contrast injection.

Prior to enrollment, single-photon emission computed tomography imaging equipment and acquisition procedures at each site were quality controlled using a cardiac phantom.³³ Raw patient data were processed by staff blinded to treatment assignment at a central core laboratory as previously described and validated.^{34 - 35} Infarct size was expressed as percentage involvement of the left ventricle. The infarct size for patients not undergoing a sestamibi scan because of death before the end of the day 5 to 14 data acquisition window was imputed as the largest infarct measured for any patient in the study (72% of the left ventricle).

All primary and secondary clinical end points were adjudicated by a clinical events committee blinded to treatment allocation after review of original source documentation. Reinfarction, ischemic target vessel revascularization, and disabling stroke were defined as previously described.¹ Severe heart failure was defined as heart failure with documented arterial partial pressure of oxygen less than 60 mm Hg or with pulmonary edema documented radiographically or requiring intubation, 100% oxygen, or insertion of an intra-aortic balloon pump. Severe hypotension was defined as systolic blood pressure less than 90 mm Hg lasting for more than 30 minutes and requiring either pressor support or insertion of an intra-aortic balloon pump.

The study size was powered to demonstrate both an increase in complete STR from an expected 50% (based on prior reports^{10 - 12 ,15} ) to 70% and a reduction in infarct size from an expected 15% (SD, 15%)^{4 ,36 - 38} to 10% (SD, 15%). The statistical method used to incorporate multiple primary end points was adapted from Benjamini and Hochberg.³⁹ The global primary (or principal) end point of the study would be considered positive if both end points were significant at α = .05 or if either end point were significant at α = .025. With data available in 400 patients (200 per group) and an α error of .05, 98.5% and 91.5% power would be available for the comparisons of STR and infarct size, respectively. The study would thus contain 90.1% power for both end points to be significant. At α = .025, 97% and 86.3% power would be present for the STR and infarct size end points, respectively. Thus, the overall power to demonstrate superiority in the trial would be 99.4%.

Categorical data were compared using the Fisher exact test. Continuous variables are presented as medians with interquartile ranges (IQRs) (except for infarct size, which was also expressed as mean [SD] per standard convention) and were compared by the Kruskal-Wallis nonparametric test. Time-to-event data are summarized as Kaplan-Meier estimates and compared with the log-rank test. Hazard ratios with 95% confidence intervals were generated using Cox regression after confirmation of the proportional hazards assumption for each end point. All analyses are by intention-to-treat, and all comparisons are 2-sided. All statistical analyses were performed with SAS version 8.2 (SAS Institute Inc, Cary, NC).

Between May 20, 2002, and November 21, 2003, 501 patients with AMI at 38 academic and community-based centers in 7 countries were randomized to receive either PCI with distal protection (n = 252) or PCI alone (n = 249). Patient flow is shown in Figure 1.

Baseline demographic and angiographic features were well matched between the 2 randomized cohorts, except for a slightly higher incidence of diabetes in the control group (Table 1 and Table 2). Of note, the inclusion criteria resulted in enrollment of patients with rapid times from symptom onset to hospital arrival (median, 77 minutes) and marked degrees of baseline ST-segment elevation (peak worst-lead median, 4.0 mm; IQR, 2.5-6.0 mm).

Among the 252 patients randomized to receive distal protection, aspiration was performed in 97% (242/251), and all angioplasty and stent balloon inflations were fully protected with occlusion balloon inflation in 79% (193/245). Some, but not all, inflations were protected in 14% (35/245) of cases, and no inflations were protected in 7% (17/245). Median aspirate volume was 50 mL (IQR, 37-75 mL), and visible debris was extracted from 73% (182/250) of patients. An example of atherothrombotic debris retrieved is shown in Figure 2.

As shown in Table 2, distal protection resulted in an additional median 21-minute delay to reperfusion, a 14-minute increase in procedural duration, and modest increases in fluoroscopy time and consumption of contrast media. The rates of angiographically documented thromboembolic complications were similar in patients treated with and without distal protection, though thrombectomy use for thrombotic procedural complications was required in slightly fewer patients receiving distal protection. Similarly, weak trends were present for more rapid postprocedural epicardial flow (ie, lower corrected TIMI frame counts) and greater microcirculatory filling (blush grade 3) with distal protection.

Site-reported intraprocedural complications occurred with similar frequency in the distal protection and control groups, including sustained ventricular tachycardia or fibrillation (7.9% [20/252] vs 6.8% [17/249], respectively; P = .73) and heart block or bradycardia requiring treatment (18.3% [46/252] vs 20.5% [51/249], respectively; P = .57). Emergency coronary artery bypass graft surgery was required in no distal protection patients and in 2 control patients. There were no intraprocedural deaths or strokes.

The coprimary end point of complete STR 30 minutes after the last contrast injection occurred in 63.3% (152/240) of patients receiving distal protection and in 61.9% (148/239) of control patients (absolute difference, 1.4% [95% confidence interval, –7.7% to 10.5%]; P = .78). There were no significant differences between the 2 groups in the rates of complete STR at any postprocedural time period (Figure 3). The mean percent STR in the distal protection and control groups was also similar at 30 minutes (74.2% vs 71.3%, respectively; P = .34) and at all other measured time periods. Of numerous prespecified subgroups examined, none was found in which a strong signal was present toward greater STR reperfusion success with distal protection (Table 3).

There was also no difference in the coprimary end point of Tc 99m sestamibi–assessed infarct size between the distal protection and control groups, whether measured as medians (12.0% [IQR, 2.0%-26.0%] vs 9.5% [0%-23.0%], respectively; P = .15) or means (18.1% [SD, 19.0%] vs 16.0% [18.7%], respectively; difference, 2.1% [IQR, –1.4% to 5.7%]; P = .24). Furthermore, no subgroup was identified in which infarct size was smaller with distal protection compared with no protection (Table 4). Of note, infarct size was increased with distal protection in thrombotic lesions and when the infarct artery was occluded prior to reperfusion.

In patients assigned to receive distal protection in whom all balloon inflations were protected vs those in whom some or none were protected, there were no differences in either the rates of complete STR at 30 minutes (62.9% [117/186] vs 64.6% [31/48], respectively; P = .87) or infarct size (median, 12.0% vs 11.5%, respectively; P = .74). Furthermore, while there was no difference in complete STR at 30 minutes in the distal protection group between patients in whom visible debris was vs was not removed (64.2% [113/176] vs 60.3% [38/63], respectively; P = .65), infarct size was actually increased in patients with visible debris extraction (median, 14.0% vs 8.0%; P = .02).

As seen in Table 5, adverse events related either to left ventricular dysfunction or ischemic complications at both 30 days and 6 months occurred with similar frequency with or without distal protection. Mortality at 6 months was also nearly identical in both groups.

Numerous small single-center experiences have demonstrated retrieval of macroscopic atherothrombotic debris from 70% to 97% of patients with AMI undergoing primary PCI with distal protection, often with improved angiographic outcomes, STR, and/or myocardial recovery compared with historical controls.^{19 - 30} Coupled with the intuitive belief that distal embolization is inherently deleterious, these reports have resulted in distal protection during primary PCI becoming routine in many centers and countries, despite the absence of a properly designed and powered prospective, randomized trial.

In the present multicenter study, representing academic and community medical centers from urban and rural environments in 7 countries, technical success rates with embolic protection during AMI were high. Aspiration was completed in 97% of patients, and some or all balloon inflations were protected in 93% of cases, reflecting the relatively small diameter (0.028 in) of the device and its ability, similar to that of a standard floppy guidewire, to traverse the native coronary circulation. Moreover, atheromatous debris, thrombotic debris, or both was retrieved in 73% of patients, similar to the 76% rate of histologically confirmed emboli extracted in the previously reported pilot phase of this study.²⁶

Nevertheless, despite these high technical success rates, distal microcirculatory protection did not improve epicardial blood flow or myocardial perfusion; did not result in more complete STR; did not reduce infarct size; and did not enhance survival free from adverse events related either to left ventricular dysfunction or ischemic complications of AMI. Thus, we did not identify clinical usefulness of distal microcirculatory protection in AMI, and its routine use in patients undergoing mechanical reperfusion therapy cannot be recommended.

Multiple explanations may be proposed as to why distal microcirculatory protection failed to enhance myocardial reperfusion success, reduce infarct size, or improve clinical outcomes in the present trial. First, the device may not have been efficient enough in aspirating liberated atherothrombotic debris. Despite visible debris extraction from approximately three quarters of patients, embolization may still have occurred at the time of initial lesion crossing and/or into unprotected proximal side branches (though there was no angiographically apparent increase in side-branch occlusion in the embolic protection group). Second, the additional median 21-minute delay to angioplasty in the distal protection group or the several extra minutes of additional balloon occlusion may have increased infarct size and worsened clinical outcomes, offsetting the potential benefit of emboli removal, though most prior studies have suggested that this degree of delay at a median reperfusion time of 3.6 hours would be unlikely to affect myocardial salvage or survival.^{40 - 42} Third, transmural infarction may similarly already be complete with little chance for additional myocardial recovery in most patients reperfused after 3 hours, as demonstrated in canine coronary occlusion models.⁴³ However, distal protection did not reduce infarct size, even in patients presenting within 1 hour of symptom onset. Fourth, while aspiration of atherothrombotic emboli is not beneficial in a nonrestrictive “all comers” population, selected subsets might exist in whom distal microcirculatory protection would be useful. However, a strong signal was not identified in any subset for usefulness of distal protection and, paradoxically, infarct size with distal protection was increased in several high-risk groups in which embolization would be most anticipated (ie, those with angiographically confirmed thrombus and totally occluded infarct vessels) and in patients in whom visible debris was extracted, suggesting the possibility of harm. And finally, the mechanism underlying reduced myocardial perfusion and suboptimal myocardial salvage after angioplasty in AMI is likely multifactorial. Thus, while distal microcirculatory embolization may occur during primary PCI and be preventable, such intervention may be “too little, too late” to achieve meaningful myocardial salvage in a disease state characterized by systemic and local mediators of inflammation and endothelial dysfunction, capillary leakage with interstitial edema, and reperfusion injury.^{44 - 45}

While the sample size and angiographic, electrocardiographic, and sestamibi imaging acquisition rates were more than sufficient to exclude clinically relevant differences between the 2 study groups in indices of myocardial perfusion and infarct size,^{46 - 47} the study was underpowered to exclude small differences in clinical events or possible benefit (or harm) in subgroups. Similarly, small differences in infarct size between the 2 groups cannot be excluded, especially in the extent of subendocardial infarction,⁴⁸ though the clinical significance of such a finding has not been established. Finally, acquisition of infarct size end point data was slightly more frequent in the distal protection group than the control group; relative treatment-related angiographic and clinical results, however, were similar in patients with and without sestamibi measures.

Our study demonstrates that though embolic debris is liberated and may be retrieved in most patients by use of the GuardWire during primary PCI in AMI, distal embolic protection with this device does not improve microvascular flow or function, reduce infarct size, or result in enhanced event-free survival. Its routine use in this setting cannot therefore be advocated. Moreover, these results fundamentally challenge the pathophysiologic importance of distal embolization during mechanical reperfusion therapy. Further investigation is required to determine whether these results apply to other microcirculatory protection devices such as filter-based catheters (for which pilot studies in AMI have been favorable)⁴⁹ and proximal balloon occlusion and aspiration systems,⁵⁰ which offer the theoretical advantages of side-branch protection and prevention of embolization during lesion crossing. Finally, it is important to emphasize that the disappointing results achieved with distal protection as an adjunct to mechanical reperfusion therapy in evolving AMI in the present study should not be generalized to other patient and lesion subtypes (eg, PCI of degenerated saphenous vein grafts), in which both a balloon occlusion and aspiration system²⁷ and a filter-based catheter²⁸ have been demonstrated to improve clinical and angiographic outcomes.