Comparison of Angioplasty With Infusion of Tirofiban or Abciximab and With Implantation of Sirolimus-Eluting or Uncoated Stents for Acute Myocardial Infarction:  The MULTISTRATEGY Randomized Trial FREE

The use of abciximab, a potent intravenous antiplatelet agent, and uncoated-stent implantation in the infarct-related lesion is a complementary treatment strategy that reduces major adverse cardiac events (MACE) in patients undergoing angioplasty (percutaneous coronary intervention [PCI]) for ST-segment elevation myocardial infarction.¹ Compared with placebo, abciximab reduces mortality and reinfarction rates^2- 3 while, compared with balloon angioplasty, primary stenting of the culprit lesion reduces the need for reinterventions.⁴ In a recent individual patient-level data meta-analysis of 3 European trials, the estimated cumulative hazard rate for death or reinfarction was reduced to 12.9% with abciximab vs 19.0% with placebo with an overall estimated number needed to treat of 19.

The value of tirofiban at optimal bolus dose^5- 6 as an alternative to abciximab in patients undergoing angioplasty for ST-elevation myocardial infarction (STEMI) is uncertain. This may carry relevant clinical and economical implications.

Drug-eluting stents further reduce the risk of target-vessel revascularization after elective PCI compared with uncoated stents. However, the use of drug-eluting stents in this patient population is discouraged⁷ because of conflicting efficacy results reported in randomized trials^8- 9 and safety concerns reported by registry studies.¹⁰

We therefore performed a multicenter, prospective, randomized trial to determine the optimal pharmaco-mechanical reperfusion strategy in patients with evolving STEMI.

We enrolled patients at 16 centers in Italy, Argentina, and Spain to participate in the Multicentre Evaluation of Single High-Dose Bolus Tirofiban vs Abciximab With Sirolimus-Eluting Stent or Bare Metal Stent in Acute Myocardial Infarction Study (MULTISTRATEGY). The study protocol was approved by the ethics committee at each institution and was conducted according to the principles of the Declaration of Helsinki. All patients gave written informed consent before enrollment.

The design of the study has been detailed in a previous publication.¹¹ Briefly, patients were randomly assigned with the use of a 2 × 2 factorial design to 1 of 4 interventional strategies of reperfusion: abciximab with an uncoated stent, abciximab with a sirolimus-eluting stent, tirofiban with an uncoated stent, or tirofiban with a sirolimus-eluting stent.

The inclusion criteria were (1) chest pain for longer than 30 minutes with an electrocardiographic ST-segment elevation of 1 mm or greater in 2 or more contiguous electrocardiogram leads, or with a new left bundle-branch block, and (2) admission either within 12 hours of symptom onset or between 12 and 24 hours after onset with evidence of continuing ischemia. The exclusion criteria included administration of fibrinolytics in the previous 30 days, major surgery within 15 days, and active bleeding or previous stroke in the last 6 months. Immediately after eligibility criteria were met and before the visualization of coronary arteries through angiography, the treating physician at each investigational site performed open-label assignments of study treatments via sealed envelopes. Randomization was achieved with a 1:1:1:1 computer-generated random sequence supplied by an academic statistician, without stratification, in blocks of 30.

Either tirofiban or abciximab was administered at first medical contact, before arterial sheath insertion during the angiography procedure. Tirofiban was given as a bolus of 25 μg/kg, followed by an 18- to 24-hour infusion at 0.15 μg/kg/min. Abciximab was administered as a bolus of 0.25 mg/kg, followed by a 12-hour infusion at 0.125 μg/kg/min. Stenting was the default strategy in patients with a reference vessel diameter of 2.5 mm or larger at visual estimation. Patients were randomized to either sirolimus-eluting stent or any uncoated-stent type approved by the regulatory agency. Crossover from a sirolimus-eluting stent to other stent types was allowed only after failure of a sirolimus stent implantation attempt or when no available stent sizes matched the coronary reference diameter. Heparin was given at 40 to 70 U/kg, targeting an activated clotting time of at least 200 seconds. Patients received aspirin (160-325 mg orally or 250 mg intravenously, followed by 80-125 mg/d orally indefinitely) and clopidogrel (300 mg orally and then 75 mg/d for at least 3 months).

A 12-lead electrocardiogram was recorded before the procedure and 90 minutes after the last balloon inflation in the infarct-related artery. Follow-up visits were scheduled at 1, 4, and 8 months.

Two primary hypotheses were prespecified: tirofiban is noninferior to abciximab for the cumulative ST-segment resolution, expressed as the proportion of patients who achieve at least 50% recovery within 90 minutes after intervention, and sirolimus-eluting stent is superior to uncoated stent in terms of the composite of MACE, defined as death from any cause, reinfarction, and clinically driven target-vessel revascularization within the first 8 months. Key secondary end points for the entire follow-up period included each component of the composite end point, stent thrombosis, and bleedings according to the criteria of the Thrombolysis in Myocardial Infarction (TIMI) trials.¹²

Deaths from all causes are reported. Reinfarction was defined as follows: (1) within 24 hours of randomization: recurrent ischemic symptoms with new, persistent ST elevation greater than 1 mm in at least 2 contiguous leads or new persistent ST depression greater than 1 mm in at least 2 contiguous leads not due to changes from evolution of the index myocardial infarction; (2) between 24 hours and 7 days of randomization: ischemic symptoms greater than 20 minutes and either a creatine kinase level greater than twice the upper limit of normal or further elevations more than 50% above the previous lowest level in patients with already elevated enzyme levels; (3) after 7 days of randomization: either typical increase and decrease of levels of biochemical markers of myocardial necrosis to greater than the upper limit of normal or, if markers are already elevated, further elevation of a marker to greater than 50% of the lowest recovery level from the index myocardial infarction with either ischemic symptoms or other ischemic changes on the electrocardiogram.

Clinically driven target-vessel revascularization is defined as any coronary artery bypass graft surgery, or a second PCI of the original target vessel, driven by clinical symptoms of myocardial ischemia with either a positive stress test or electrocardiographic evidence of ischemic changes at rest attributable to the target vessel and the presence of luminal stenosis of more than 70% of the reference luminal diameter by visual estimate. Stent thrombosis was classified as definite, probable, or possible in keeping with recently proposed Academic Research Consortium classification.¹³

Clinical data were prospectively collected at each site by research nurses or treating physicians. Independent study monitors employed by University of Ferrara verified 100% of the data in the case-report forms. The data for all patients with primary end-point events were reviewed by an independent adjudication committee whose members were blinded to treatment assignments. Events adjudication was performed separately by 2 members, and in case of disagreement, the opinion of the third member was obtained and the final decision taken by consensus. The committee was also responsible for the adjudication of all clinical events according to the Academic Research Consortium.¹³

Changes in the ST-segment of the electrocardiogram were evaluated cumulatively before and 90 minutes after intervention. ST-segment elevation was measured to the nearest 0.5 mm at 60 milliseconds after the J point by a single experienced cardiologist who was blinded to treatment assignments. The intraobserver agreement was 94.1% (κ = 0.82) in identifying the recovery by at least 50% of ST-segment elevation in 217 randomly selected patients (30% of all interpretable electrocardiograms). Quantitative angiographic analyses were performed with a validated edge-detection system (CAAS II; Pie Medical, Maastricht, the Netherlands), and coronary flow was classified according to TIMI criteria. Angiographic analyses and TIMI grading were performed by 1 independent cardiologist who was blinded to treatment assignments.

Discrete data were summarized as frequencies, and comparisons were made with the likelihood-ratio χ² test or Fisher exact test. Continuous data were expressed as mean (SD) or median and interquartile range according to their distribution; comparisons were made with a 1-way analysis of variance or the Kruskal-Wallis test.

Comparisons Between Drug Groups. A total of 580 patients was required for greater than 85% power in detecting a 9% absolute difference, 0.89 in terms of relative risk, between groups in the proportion of patients who attained at least 50% resolution of ST-segment elevation, which corresponds to the 50% previously observed absolute difference between abciximab and placebo,¹⁴ with a 2-sided 2.5% significance level and an 85% expected event rate in the control group based on previous findings.¹⁴ The noninferiority test was computed with the continuity-corrected χ² of Dunnett and Gent on the entire patient cohort. This was based on both intention-to-treat and per-protocol principles and was applied to an exploratory analysis across several prespecified subgroups. The Cochran-Mantel-Haenszel χ² test was performed to evaluate possible imbalances of the relative risk among different recruiting centers.

Comparisons Between Stent Groups. Based on the STRATEGY trial,⁶ the tirofiban/sirolimus-eluting stent group experienced a 44% relative risk reduction in the occurrence of MACE at 8 months as compared with the combination of abciximab and uncoated stent, with an absolute risk reduction of 14% (from 32% to 18%). Because this was almost entirely driven by the different rate of reintervention in the 2 groups, we assumed this to be a reasonably good estimate of the effect of sirolimus-eluting stents compared with uncoated stents in the setting of STEMI. To correct for the “oculostenotic reflex,” which may artificially increase the MACE rate in the study population by affecting the need for target-vessel revascularization, we anticipated a lower event rate in the uncoated-stent group of around 27% in keeping with recent data, with a 40% relative risk reduction for MACE in the sirolimus-eluting–stent group (from 27% in the standard-stent to 16% in the active-stent group). Thus, a minimum of 600 patients was required for at least an 80% power to detect the prespecified difference using a 2-sided 2.5% level of significance (with Bonferroni correction).

Event-free survival curves were generated by the Kaplan-Meier method and survival differences between groups were compared using the log-rank test. To test whether initial differences between the 2 treatment groups influenced the difference in outcomes in terms of the primary end point, all variables with a P < .16 in the log-rank test were evaluated with the multivariate Cox proportional hazard model.¹⁵ The assumption of proportionality was tested using Schoenfield residuals. A model selection procedure was applied using the Akaike information criteria linked to a bootstrap approach.¹⁶ An extension of the standard Cox model that allows for random effects was also applied to estimate the heterogeneity between centers. All analyses were performed using Stata version 9.2 (Stata Corp, College Station, Texas).

Between October 2004 and April 2007, 1030 patients were screened; of these, 745 were enrolled and randomly assigned to the 4 treatment groups (Figure 1). After randomization and treatment, 1 patient in the abciximab/sirolimus-stent group withdrew from the study. Patient characteristics were similar among all 4 groups except for a slightly higher prevalence of prior transient ischemic attacks in the tirofiban/uncoated-stent group (Table 1).

The procedural results were similar in the 4 groups with the exception of a slightly smaller maximal stent size in the sirolimus-eluting–stent group—reflecting the limited available range of stent diameters—which was compensated for by a higher pressure of stent deployment (Table 2). In the uncoated-stent group, 96% of the implanted stents had less than 0.10 mm strut thickness. Coronary thrombus was unequivocally identified in more than two-thirds of the cases, and flow rates of grade 3 were present in 20% of patients before the procedure. Reference vessel and minimal lumen diameters did not differ among the 4 treatment groups. Normal flow grade was achieved in 91.4% to 95.7% of all patients.

Among the 722 patients (97%) who had an interpretable electrocardiogram (Figure 1), at least 50% recovery from ST-elevation occurred in 302 of 361 patients (83.6%) and 308 of 361 patients (85.3%) in the abciximab and tirofiban groups, respectively, in the intention-to-treat analysis (relative risk for tirofiban vs abciximab, 1.020; 97.5% confidence interval, 0.958-1.086; P value <.001 for noninferiority). The per-protocol analysis yielded similar results (relative risk, 1.020; 97.5% confidence interval, 0.959-1.086; P value <.001 for noninferiority). There was no interaction between the different glycoprotein IIb/IIIa inhibitors and stent types (P = .60). The efficacy of the tirofiban infusion did not show heterogeneity among different recruitment sites (χ², 6.22, P = .72) and was consistent among multiple prespecified subgroups (Figure 2).

At 30 days, the incidence of the primary clinical end point—a composite of death, reinfarction, or revascularization of the target vessel—was 4.3% vs 4.0% (P = .85) in the abciximab and tirofiban groups, respectively, and 5.1% vs 3.2% (P = .20) in the uncoated- and sirolimus-eluting–stent groups, respectively. The incidence of major and minor bleedings did not differ (7.8% in the abciximab vs 7.2% in the tirofiban group, P = .89), but the incidence of severe or moderate thrombocytopenia was more common with abciximab compared with tirofiban (4.0% vs 0.8%, P = .004) (Table 3).

At 8 months, the MACE rate was similar among those who received tirofiban (9.9%) and those who received abciximab (12.4%; P = .30) but was higher among those who were treated with the uncoated stent (54 patients, 14.5%) compared with those who were treated with the sirolimus-eluting stent (29 patients, 7.8%; P = .004) (Figure 3); there was no evidence of heterogeneity among sites (χ², 3.95; P = .18) and no evidence for interaction between drug and stent types at any time frame (P = .95 for interaction on MACE at 8 months and glycoprotein IIb/IIIa inhibitors and stent types). The composite of death or reinfarction and the incidence of stent thrombosis (Table 3) were similar across the 4 groups, but revascularization was reduced from 10.2% with the uncoated stent to 3.2% with the sirolimus-eluting stent (P < .001) (Figure 4). In multivariate analysis, patients in the sirolimus-eluting–stent group were less likely than those in the uncoated-stent group to have MACE (hazard ratio, 0.53; 97.5% confidence interval, 0.33-0.83; P = .006).

In patients with STEMI, platelet reactivity is associated with the severity of myocardial damage¹⁷ and strongly correlates with various measures of myocardial reperfusion, including ST-segment recovery after treatment.^18- 19 In the most recent study of abciximab vs placebo in patients undergoing primary angioplasty, the degree of ST-segment resolution was significantly improved with abciximab,²⁰ as was the mortality rate at 12 months.¹⁴ Tirofiban belongs to the same class of antiplatelet agents as abciximab, namely glycoprotein IIb/IIIa inhibitors. However, tirofiban differs from abciximab in terms of both pharmacodynamic and pharmacokinetic profiles.²¹

The first head-to-head comparison between abciximab and tirofiban was powered based on the preservation of a difference of at least 50% in the effect of abciximab as compared with that of placebo.²² In that study, abciximab was superior to tirofiban with respect to the prespecified combined end point.²² This result was driven by a higher rate of periprocedural myocardial infarction in the tirofiban group, suggesting inadequate early platelet inhibition with the bolus regimen (10 μg/kg) used.²² Subsequent dose-ranging studies showed that increasing the tirofiban bolus dose from 10 to 25 μg/kg provided an optimal level of platelet inhibition,²³ and several independent pharmacokinetics studies suggested that tirofiban, at increased dose, might even lead to a more consistent platelet inhibition than abciximab.^6,24- 25 To date, 3 small single-center investigations^6,26- 27 and 1 prematurely stopped multicenter randomized study²⁸ have compared high-dose tirofiban with abciximab in 719 patients undergoing PCI; however, none of these studies had adequate power to evaluate the comparison between the 2 drugs.

In the present study, we sought to determine whether tirofiban, at proper dosing, would achieve at least 50% of the anticipated effect of abciximab on the recovery of ST-segment elevation.

Tirofiban yielded noninferior recovery from ST-segment elevation after coronary intervention in comparison with abciximab; this result was consistent across different recruiting centers and multiple prespecified subgroups. Similarly, the rate of MACE or bleeding events did not differ between the tirofiban or abciximab groups, but the incidence of severe or moderate thrombocytopenia was lower in the tirofiban group compared with the abciximab group, a finding of potential clinical relevance.²⁹ This may be related to lower propensity of tirofiban to elicit an antibody response.³⁰

Similar to abciximab, tirofiban inhibits platelet activity through glycoprotein IIb/IIIa platelet receptor blockade, but unlike abciximab, tirofiban exerts a competitive and rapidly reversible antagonism and does not inhibit other β3 integrins, such as the vitronectin receptor, at the surface of vascular cells or the activated Mac-1 receptor on leukocytes.³¹ These have traditionally been regarded as crucial targets to explain abciximab effects especially on microcirculation in the setting of ongoing myocardial infarction.³² The results of our study may question this paradigm while emphasizing the relevance of the potency and consistency of platelet inhibition in explaining the differential clinical effects of different antiplatelet agents, which is in keeping with the findings from a recent comparison between 2 oral P2Y12 receptor antagonists.³³

Two medium-sized multicenter studies have compared drug-eluting vs uncoated stents in patients with STEMI.^8- 9 In the Trial to Assess the Use of the Cypher Stent in Acute Myocardial Infarction Treated with Balloon Angioplasty (TYPHOON), the use of sirolimus-eluting stents improved outcomes compared with uncoated stents in terms of the prespecified primary end point in 712 patients.⁹ That study, however, recruited a highly selected patient population based on findings in coronary angiography and mandated angiographic follow-up in a large proportion of patients.³⁴ In contrast, the Paclitaxel-Eluting Stent vs Conventional Stent in Myocardial Infarction with ST-Segment Elevation (PASSION) trial had less stringent angiographic exclusion criteria and mandated clinical follow-up only. In that trial, use of paclitaxel-eluting stents in 619 patients failed to provide a significant benefit compared with uncoated stents at 6 or 12 months.⁸ The present study was designed to avoid confounders driven by findings in coronary angiography; we recruited patients before coronary angiography to avoid selection biases, and our protocol mandated clinical follow-up to circumvent the artificial increase in reinterventions due to the “oculostenotic reflex.”³⁵

With this approach, we were able to recruit 72% of consecutive patients presenting with STEMI at study sites during enrollment, which favorably compares with the previous 2 studies.^8- 9 We observed a significantly lower rate of MACE at 8 months in patients assigned to sirolimus-eluting stent as compared with uncoated stent. This finding was driven by an almost 70% reduction in the incidence of target-vessel revascularization. Our protocol mandated dual antiplatelet treatment for at least 3 months after stent placement. Accordingly, more than 60% of patients in the sirolimus-eluting stent group were no longer taking thienopyridines by 8 months. Yet, there were no differences between the 2 stent groups in the overall mortality rate, the composite incidences of death or reinfarction, or the cumulative incidence of stent thrombosis.

Our randomized controlled data complement recent observational findings on the use of drug-eluting stents for off-label indications³⁶ and indicate that sirolimus-eluting stent implantation in the context of largely thrombotic lesions, such as those observed in consecutive patients with ongoing myocardial infarction, may be as safe as, and possibly more effective than, the use of uncoated stents at midterm follow-up.

Several limitations of the present study deserve comment. First, the open-label design may have introduced the potential for bias. We attempted to minimize this potential with the requirement that all ischemic and hemorrhagic events were adjudicated by independent committees unaware of the treatment assignments. Second, the selection of ST-segment resolution as a primary end point for the drug comparison may be regarded as suboptimal compared with clinical outcomes. The recovery from ST-segment elevation, however, represents a crucial target for managing this patient population because this finding reflects myocardial salvage^37- 38; may guide the intensity of treatment^39- 40 (eg, triggers activation of rescue PCI after a failed thrombolysis)⁴¹; and closely correlates with infarct size, left ventricular function recovery, and mortality outcomes.^42- 43 Finally, our study did not address the long-term durability of the safety and efficacy profile of sirolimus-eluting stents in patients with STEMI; this remains a critical issue that warrants further investigation.

In summary, our study provides evidence that in a broad population of largely unselected patients undergoing PCI for STEMI, tirofiban therapy is associated with a noninferior resolution from ST-segment elevation at 90 minutes postintervention compared with abciximab, and at 8-month follow-up, MACE are approximately halved by sirolimus-eluting stent implantation compared with uncoated stents.