Impact of Different Platelet Glycoprotein IIb/IIIa Receptor Inhibitors Among Diabetic Patients Undergoing Percutaneous Coronary Intervention
Do Tirofiban and ReoPro Give Similar Efficacy Outcomes Trial (TARGET) 1-Year Follow-Up
Background— The platelet glycoprotein IIb/IIIa receptor inhibitor abciximab, a monoclonal antibody, has been shown to improve early and late outcomes among diabetic patients undergoing percutaneous coronary intervention (PCI). It is unknown whether small-molecule agents confer similar benefits.
Methods and Results— In 18 countries, 4809 patients undergoing PCI with stent implantation were randomized to tirofiban or abciximab. At the time of enrollment, patients were stratified according to diabetes status. As compared with non-diabetic patients, patients with diabetes (n=1117) showed similar 30-day ischemic outcomes, an increased incidence of any target vessel revascularization (TVR) at 6 months (10.3% versus 7.8%; P= 0.008), and a trend toward higher 1-year mortality (2.5% versus 1.6%; P=0.056). Among diabetic patients randomized to tirofiban (n=560), the incidence of death, myocardial infarction (MI), or urgent TVR at 30 days was 6.2%, and among those randomized to abciximab (n=557) it was 5.4% (hazard ratio [HR] 1.16; P=0.540). At 6 months, the composite of death, MI, or any TVR occurred in 15.7% and in 16.9% of tirofiban and abciximab patients, respectively (HR 0.93; P=0.610). Any TVR occurred in 9.5% and 11.1%, respectively (HR 0.84; P= 0.366). The 1-year mortality was 2.1% in the tirofiban group and 2.9% in the abciximab group (HR 0.74; P= 0.436).
Conclusions— Among diabetic patients undergoing PCI, tirofiban and abciximab were associated with comparable event rates, including similar rates of 6-month TVR and 1-year mortality. These findings suggest that the non-glycoprotein IIb/IIIa properties of abciximab do not translate into a discernible long-term clinical benefit among diabetic patients.
Received February 15, 2002; revision received March 28, 2002; accepted March 28, 2002.
Diabetes mellitus is associated with worse outcome after percutaneous coronary angioplasty and stenting.1–3⇓⇓ Compared with non-diabetic patients, those with diabetes have higher mortality rates, more exuberant restenosis, and greater de novo disease progression.1,3,4⇓⇓ Subgroup analyses of clinical trials have shown the platelet glycoprotein (GP) IIb/IIIa receptor inhibitor abciximab to be particularly beneficial among diabetic patients undergoing coronary stenting, potentially influencing the restenotic process.5,6⇓ Animal studies have supported the concept that the monoclonal antibody fragment abciximab may limit restenosis by inhibiting neointimal hyperplasia via blockade of the αvβ 3 (vitronectin) integrin on smooth muscle cells.7–9⇓⇓ In addition, the greatest relative risk reduction in 1-year mortality provided by abciximab versus placebo has been observed among patients with diabetes.10
It is unknown whether the clinical benefit associated with abciximab among diabetic patients also applies to small-molecule agents, such as tirofiban, which display pure αIIbβ 3 (GP IIb/IIIa) inhibition. In particular, it remains to be determined whether the non-GP IIb/IIIa properties of abciximab, such as vitronectin11 or Mac-1 receptor inhibition,12 favorably affect target vessel revascularization (TVR) or late mortality. The Do Tirofiban and ReoPro Give Similar Efficacy Outcomes Trial (TARGET)13,14a⇓ was performed to test whether tirofiban (Aggrastat, Merck) was comparable to abciximab (ReoPro, Johnson and Johnson) in patients undergoing non-emergent stent-based PCI. We report the outcome of the diabetic population enrolled in the trial to 1 year.
The design and methods of TARGET have been previously described.13,14a⇓ In brief, between December 1999 and August 2000, 4809 patients undergoing non-emergent PCI of a de novo or restenotic lesion in a native vessel or a bypass graft were enrolled. The protocol mandated intention to stent all treated sites. Patients were excluded in the presence of ST-segment elevation myocardial infarction (MI), cardiogenic shock, creatinine >2.5 mg/dL, bleeding diathesis, or life-limiting conditions. At time of enrollment, patients were stratified according to diabetes status to allow similar distribution of diabetic patients among the 2 treatment arms. Patients were classified as having diabetes based on history, and insulin requirements were recorded. Enrollment sites were located in North America, Australia, and 14 European countries.
All patients received aspirin before the procedure, and the administration of clopidogrel 300 mg orally at least 2 hours before cardiac catheterization was strongly recommended. Heparin was dosed to achieve an activated clotting time of ≥250 seconds. The study drug was administered immediately before revascularization in a double-blind, double-dummy manner. Accordingly, all patients received both the active formulation of one treatment and the placebo formulation of the other treatment. Tirofiban was administered as a 10 μg/kg bolus followed by a 0.15 μg · kg−1 · min− 1 infusion for 18 to 24 hours, and abciximab was administered as 0.25 mg/kg bolus and 0.125 μg · kg−1 · min−1 infusion (maximum 10 μg/min) for 12 hours. Aspirin and clopidogrel were continued for 30 days, and long-term aspirin was recommended.
The primary endpoint of the trial was a composite of death, nonfatal MI, or urgent TVR at 30 days. The 6-month composite of death, MI, and any TVR rate, as well as the 1-year mortality, were prespecified secondary endpoints, and all data were collected prospectively. Death was defined as all-cause mortality. The definition of MI during index hospitalization and to 30 days has been described.13 From 30 to 180 days, MI was defined by a 2-fold rise in the creatine kinase (CK)-MB isoform on 2 occasions or the appearance of new Q-waves on ECG. Any TVR was defined as either percutaneous revascularization or bypass surgery to a vessel that had been treated during the index procedure. The first event for each patient was recorded for the 30-day composite endpoint. Safety analysis included major and minor bleeding complications as defined in the Thrombolysis in Myocardial Infarction (TIMI) trials.14b
Data Collection and Statistical Analysis
Treatment analysis included all patients who received either study drug. Each enrollment center contacted patients for follow-up at 30 days, 6 months, and 1 year. When relevant, other medical records were collected (eg, death certificates, CK values, and revascularization reports). Endpoint events were investigator-reported, with 100% monitoring of source documentation. The double blinding of treatment assignment was maintained through 1-year follow-up. All endpoint events were reviewed and adjudicated by an independent Clinical Events Committee.
Percentages reported for demographic, procedural, and safety data were based on non-missing observations. Continuous baseline and procedural characteristics were reported as mean±SD. Categorical data were presented as percentages. Mortality at 1 year was reported as Kaplan-Meier estimates. Cox proportional-hazards model was used to calculate hazard ratios (HR) and CI based on time to event. The same method was used to obtain probability values for the primary endpoint, subgroup analyses, and secondary endpoints, as well as to test the interaction of diabetic status with treatment. Comparisons of TIMI bleeding events were performed using Fisher’s exact test. No adjustments were made for multiple statistical comparisons, and P<0.05 was considered statistically significance.
Treatment and Outcome According to Diabetes Status
Of the 4809 patients enrolled in the trial, 1117 patients (23%) had diabetes, and follow-up at 1 year was available for 99% of patients. When compared with non-diabetic patients, those with diabetes had comparable event rates at 30 days, higher mortality and TVR rates at 6 months, and a trend toward higher mortality at 1 year (Table 1).
Stratification according to diabetes status at enrollment led to uniform distribution of diabetic patients to tirofiban (n=560) and abciximab (n=557) treatment. Diabetic patients treated with tirofiban versus abciximab had similar baseline and procedural characteristics (Tables 2 and 3⇓). TIMI major bleeding events occurred in 0.5% of patients randomized to tirofiban and in 0.7% of patients randomized to abciximab (P=0.725). Minor bleeding was observed in 3.6% and 5.0% of cases, respectively (P=0.249). At 30 days, the incidence of ischemic events was similar (Table 4), and the composite of death, MI, or urgent TVR occurred in 6.2% of patients in the tirofiban group and 5.4% of patients in the abciximab group (HR 1.16; 95% CI 0.71 to 1.89; P=0.540). The outcomes among diabetic patients in the 2 treatment arms remained comparable at 6 months (Table 5). The triple endpoint (death, MI, or any TVR) occurred in 15.7% of patients in the tirofiban group and in 16.9% of patients in the abciximab group (HR 0.93; 95% CI 0.69 to 1.24; P=0.610). The incidence of TVR was 9.5% and 11.1%, respectively (HR 0.84; 95% CI 0.59 to 1.22; P=0.366). Twelve diabetic patients (2.1%) died at 1 year in the tirofiban group and 16 patients (2.9%) died in the abciximab group (HR 0.74; 95% CI 0.35 to 1.57; P=0.436; Figure 1).
Among non-diabetic patients (n=3692), those receiving tirofiban had higher events rates at 30 days compared with those receiving abciximab; the incidence of death, MI, or urgent TVR in the 2 groups was 7.9% and 6.2%, respectively (HR 1.29; 95% CI 1.01 to 1.64; P=0.043). At 6 months, 14.6% of patients in the tirofiban group and 13.5% of patients in the abciximab group had an ischemic event (HR 1.08; 95% CI 0.91 to 1.29; P=0.355). The incidence of any TVR was similar between the 2 groups (7.7% and 7.9%, respectively). At 1 year, the mortality rate was 1.8% and 1.4%, respectively (HR 1.32; 95% CI 0.79 to 2.20; P=0.288; Figure 2). No significant interaction was observed between diabetes status and treatment (P=0.719 and P=0.362 for the composite endpoint at 30 days and 6 months, respectively; P=0.214 for mortality at 1-year).
Treatment and Outcome According to Insulin Treatment
Overall, diabetic patients treated with insulin (n=503) and those not insulin treated (n=614) had similar outcomes. The 30-day incidence of death, MI, or urgent TVR was 5.6% and 6.0% (P=0.764), and at 6 months the death, MI, or any TVR composite occurred in 15.9% and 16.6% of patients (P=0.789), respectively. At 1 year, the mortality rate was 2.8% and 2.3% (P=0.604), respectively.
Among patients on insulin, the composite endpoint of death, MI, or urgent TVR occurred in 8.1% in the tirofiban group and 3.1% in the abciximab group at 30 days (Figure 3). The difference reached statistical significance (HR 2.65; 95% CI 1.17 to 6.02; P=0.020) and was driven by periprocedural MI, which occurred in 6.5% of patients randomized to tirofiban and in 2.3% of patients randomized to abciximab (HR 2.79; 95% CI 1.09 to 7.14; P=0.032). At 6 months, the difference in the triple endpoint was no longer significant (17.4% versus 14.4%, HR 1.25; 95% CI 0.81 to 1.94; P=0.317). At 1-year, the mortality was 2.0% and 3.5%, respectively (HR 0.58; 95% CI 0.19 to 1.73; P=0.329).
Among diabetic patients not receiving insulin, the incidence of death, MI, or TVR at 30 days was 4.8% in the tirofiban group and 7.3% in the abciximab group (HR 0.65; 95% CI 0.34 to 1.26; P=0.203; Figure 3). At 6 months, the incidence of the combined endpoint was 14.5% and 18.9%, respectively (HR 0.73; 95% CI 0.50 to 1.08; P=0.118). The 1-year mortality was 2.2% in the tirofiban group and 2.3% in the abciximab group (HR 0.95; 95% CI 0.33 to 2.70; P= 0.920).
Previous clinical trial work from our group and others has substantiated the benefit of abciximab compared with placebo for diabetic patients undergoing PCI.6,10⇓ It has remained uncertain, however, whether small-molecule platelet GP IIb/IIIa receptor inhibitors would achieve a similar benefit. In the TARGET trial, which included the largest reported cohort of diabetic patients undergoing coronary stenting, we found that tirofiban led to comparable results with respect to many endpoints when compared with abciximab, specifically to similar rates of clinical restenosis (any TVR) and late survival. With diabetic patients duly acknowledged as among those at highest risk for restenosis and late mortality after PCI, this finding has important clinical and biological implications.
Diabetes and Early Outcome After PCI
In vitro and ex vivo studies have shown that diabetic patients have increased platelet aggregability.15–17⇓⇓ In particular, diabetes mellitus has been associated with a larger pool of activated circulating platelets, 18 a higher density of GP IIb/IIIa receptors per platelet,18 and a greater expression of platelet P-selectin and thrombospondin. 19 These adhesion molecules are potential triggers of inflammatory response and thrombosis. Despite increased baseline platelet aggregability, the administration of abciximab has been shown to achieve a similar degree of platelet aggregation inhibition among diabetic and non-diabetic patients undergoing PCI. 20
In contrast to clinical expectation and data from the Evaluation of Platelet IIb/IIIa Inhibitor for Stenting Trial (EPISTENT), 6 diabetic patients enrolled in the TARGET trial had favorable outcomes at 30 days, with relatively fewer ischemic events as compared with non-diabetic patients (5.8% versus 7.1%; P=0.153). The lack of excess of ischemic complications, and in particular of periprocedural MI, observed among diabetic patients seems counterintuitive. However, similar findings were observed in the Eptifibatide in Planned Coronary Stent Implantation Trial (ESPRIT), such that at 30 days, diabetic patients had lower event rates than non-diabetic patients, both in the placebo and in the eptifibatide treatment arms.21 One plausible explanation for this discrepancy among the trials is that following the results of the Bypass Angioplasty Revascularization Investigation (BARI)22 and the Arterial Revascularization Therapy Study (ARTS),23 which showed a survival benefit associated with bypass surgery among diabetic patients with multivessel disease, the diabetic patients enrolled in ESPRIT and TARGET may have represented a subset with coronary anatomy more favorable for PCI. This observation remains speculative, however, as a comparison of demographic or anatomic characteristics among the trials’ diabetic populations has not been made. In addition, as TARGET was designed with 2 active treatment arms, no placebo group was available for comparison to the previous trials.
In TARGET, we observed that the use of tirofiban and abciximab among diabetic patients led to similar 30-day event rates, though the hazard ratio for the composite endpoint was consistent with the overall trial findings that favored abciximab. No significant interaction was detected between diabetic status and treatment.
Integrin Receptors and Target Vessel Revascularization
Diabetes mellitus is associated with worse intermediate- and long-term outcomes after percutaneous coronary angioplasty1 and stenting as compared with non-diabetic patient outcomes.2,3⇓ Characteristic among diabetic patients is the excess of TVR, a surrogate marker for angiographic restenosis. Intravascular ultrasound studies have shown that diabetic patients have more aggressive tissue proliferation, both in stented and non-stented lesions.4 Likewise, at long-term follow-up, the outcome of diabetic patients after PCI is marked by more rapid progression of native coronary artery disease than that in non-diabetics.1 As anticipated, the diabetic patients enrolled in TARGET had significantly higher TVR rates compared with non-diabetic patients. Nonetheless, the 6-month TVR observed in the present study (10.3%) is comparable to the rate reported among diabetic patients in the stent-abciximab group of EPISTENT (8.1%).6 This demonstrates that in the current PCI-stent-IIb/IIIa era, the low TVR rate observed in EPISTENT can be reproduced in a large-scale diabetic population.
Controversy over whether abciximab has anti-restenotic properties evolved from observed reductions in 6-month TVR among some,5,24⇓ but not all,25,26⇓ trials. The only prior PCI-stent-abciximab trial, EPISTENT, showed that the combination of abciximab and stenting was associated with a significant (51%) reduction in TVR at 6 months among diabetic patients.5 The hypothesis of an inhibitory effect of the drug on restenosis was supported by the angiographic substudy of the trial, which showed that among diabetic patients undergoing coronary stenting, randomization to abciximab was associated with a significant increase in net gain (the difference between minimal luminal diameter at follow-up and before intervention).5 Animal studies have reinforced the concept that the antibody fragment abciximab may play a role in restenosis prevention by inhibiting neointimal hyperplasia. Putative pathways involve inhibition of both the αIIbβ3 (GP IIb/IIIa) receptor, which is confined to platelets, and of the α vβ3 (vitronectin) receptor, 11 which is heavily expressed by several cell lines, including endothelial cells, monocytes, and smooth muscle cells.12,27⇓ Abciximab has been shown to directly inhibit smooth muscle cell migration and proliferation through binding of αvβ3 integrin of smooth muscle cells.7–9⇓⇓ Conversely, the small-molecule tirofiban, which binds exclusively to the αIIbβ 3 receptor, has demonstrated no effect on smooth muscle cells in animal models7 or on angiographic restenosis in angioplasty substudies.28 Additional properties of abciximab include cross-reactivity with the leukocyte αMβ 2 integrin (Mac-1), a mediator of inflammation after arterial injury, which may also be involved in restenosis.12 Despite the suggestive findings in EPISTENT and the plausible mechanistic explanations from animal studies, an intravascular ultrasound study designed to assess neointimal volume after PCI-stent failed to demonstrate a significant benefit from abciximab.26
Among diabetic patients enrolled in TARGET, those randomized to tirofiban had a nonsignificant 16% lower TVR rate at 6 months compared with patients randomized to abciximab. Despite the limitation that TARGET was not an angiographic study, and acknowledging that TVR is only a surrogate marker of restenosis, our findings suggest that the non-GP IIb/IIIa properties of abciximab do not translate into a clinically relevant inhibition of the restenotic process.
In the EPISTENT trial, abciximab was associated with a significant overall mortality reduction at 1 year as compared with placebo (2.4% to 1.0%; P=0.037) among PCI-stent patients.29 In addition, a pooled analysis of the Evaluation of c7E3 in Preventing Ischemic Complications (EPIC), Evaluation in PTCA to Improve Long-Term Outcome With Abciximab GP IIb/IIIa Blockade (EPILOG), and EPISTENT data sets, found that diabetic patients specifically (n=1462) derive a 44% 1-year mortality reduction associated with abciximab (4.5% to 2.5%; P= 0.031).10 In TARGET, the diabetic patients enrolled were at a higher risk for death compared with non-diabetics at 6 months, and the trend persisted at 1-year. Up to 1 year, no difference in diabetic mortality was observed between the 2 agents. This suggests that the long-term mortality benefit of platelet GP IIb/IIIa receptor inhibitors among diabetic patients may not be linked to abciximab-specific effects.
Insulin Requirement and Outcome
Overall, diabetic patients on insulin and those not requiring insulin had similar outcomes at 30 days and 6 months, as well as similar 1-year mortality rates. Among patients not treated with insulin, tirofiban and abciximab were associated with comparable event rates at 30 days and 6 months, as well as similar 1-year mortality rates. Among patients on insulin, those randomized to tirofiban had higher incidence of the triple endpoint at 30 days, driven by an excess of periprocedural MI. This difference lost statistical significance at 6 months, and at 1-year the mortality rates were similar. The isolated significant benefit associated with abciximab in terms of periprocedural MI among patients requiring insulin could be the result of chance or could indicate that these patients derive particular benefit from a presumably more potent and more broad (ie, non-GP IIb/IIIa mediated) platelet inhibition generated by abciximab. Although we do not have information regarding the etiology of diabetes among patients receiving insulin (type 1 insulin-sensitive versus type 2 insulin-resistant), the vast majority of the patients enrolled in the trial were likely to have type 2 diabetes. Insulin resistance appears to have a direct impact on hemostasis, as suggested by the correlation between insulin levels and both fibrinogen and plasminogen activator inhibitor (PAI)-1 levels.30,31⇓ Further studies are needed to assess whether insulin resistance may be used to identify diabetic patients at highest thrombotic risk during PCI, and accordingly to guide future studies and specific treatments.
This study is subject to the limitations of subgroup analysis because it was not powered to compare tirofiban with abciximab solely among diabetic patients. In addition, multiple statistical comparisons without adjustments were performed. This analysis, however, involves the largest reported cohort of diabetic patients undergoing coronary stenting. Moreover, the diabetic subgroup was prespecified in the study protocol for stratification at enrollment and statistical analysis. As TARGET was not an angiographic study, no conclusive statements can be made regarding differences between the 2 drugs with regard to restenosis. Based on TVR rates, however, it is unlikely that a clinically meaningful difference went undetected.
Within the first head-to-head comparison of 2 platelet GP IIb/IIIa inhibitors as adjunctive treatment for stent-based PCI, periprocedural administration of the small-molecule tirofiban and the antibody fragment abciximab led to overall similar outcomes among diabetic patients. In particular, the comparable TVR and mortality rates observed between the 2 groups suggest that the non-GP IIb/IIIa properties of abciximab do not translate into a discernible long-term clinical benefit among diabetic patients.
*The list of the International Steering Committee, the Executive Committee, and study investigators was published previously (N Engl J Med. 2001;344:1888– 1894).
The TARGET study was supported by Merck & Co, Inc, West Point, Pa. Drs DiBattiste, Harris, and Demopoulos are employees of Merck.
- 1.↵Stein B, Weintraub WS, Gebhart SP, et al. Influence of diabetes mellitus on early and late outcome after percutaneous transluminal coronary angioplasty. Circulation. 1995; 91: 979–989.
- 4.↵Kornowski R, Mintz GS, Kent KM, et al. Increased restenosis in diabetes mellitus after coronary interventions is due to exaggerated intimal hyperplasia: a serial intravascular ultrasound study. Circulation. 1997; 95: 1366–1369.
- 6.↵Marso SP, Lincoff AM, Ellis SG, et al. Optimizing the percutaneous interventional outcomes for patients with diabetes mellitus: results of the EPISTENT (Evaluation of platelet IIb/IIIa inhibitor for stenting trial) diabetic substudy. Circulation. 1999; 100: 2477–2484.
- 8.↵Baron JH, Moiseeva EP, de Bono DP, et al. Inhibition of vascular smooth muscle cell adhesion and migration by c7E3 Fab (abciximab): a possible mechanism for influencing restenosis. Cardiovasc Res. 2000; 48: 464–472.
- 12.↵Simon DI, Xu H, Ortlepp S, et al. 7E3 monoclonal antibody directed against the platelet glycoprotein IIb/IIIa cross-reacts with the leukocyte integrin Mac-1 and blocks adhesion to fibrinogen and ICAM-1. Arterioscler Thromb Vasc Biol. 1997; 17: 528–535.
- 14B.↵Antman EM. Hirudin in acute myocardial infarction: safety report from the Thrombolysis and Thrombin Inhibition in Myocardial Infarction (TIMI) 9A Trial. Circulation. 1994; 90: 1624–1630.
- 16.↵Davi G, Gresele P, Violi F, et al. Diabetes mellitus, hypercholesterolemia, and hypertension but not vascular disease per se are associated with persistent platelet activation in vivo: evidence derived from the study of peripheral arterial disease. Circulation. 1997; 96: 69–75.
- 20.↵Steinhubl SR, Kottke-Marchant K, Moliterno DJ, et al. Attainment and maintenance of platelet inhibition through standard dosing of abciximab in diabetic and nondiabetic patients undergoing percutaneous coronary intervention. Circulation. 1999; 100: 1977–1982.
- 23.↵Abizaid A, Costa MA, Centemero M, et al. Clinical and economic impact of diabetes mellitus on percutaneous and surgical treatment of multivessel coronary disease patients: insights from the Arterial Revascularization Therapy Study (ARTS) trial. Circulation. 2001; 104: 533–538.
- 26.↵The ERASER Investigators. Acute platelet inhibition with abciximab does not reduce in-stent restenosis (ERASER study). Circulation. 1999; 100: 799–806.
- 27.↵Murphy JF, Bordet JC, Wyler B, et al. The vitronectin receptor (alpha v beta 3) is implicated, in cooperation with P-selectin and platelet-activating factor, in the adhesion of monocytes to activated endothelial cells. Biochem J. 1994; 304: 537–542.
- 29.↵Topol EJ, Mark DB, Lincoff AM, et al. Outcomes at 1 year and economic implications of platelet glycoprotein IIb/IIIa blockade in patients undergoing coronary stenting: results from a multicentre randomised trial. EPISTENT Investigators Evaluation of Platelet IIb/IIIa Inhibitor for Stenting. Lancet. 1999; 354: 2019–2024.
- 30.↵Juhan-Vague I, Thompson SG, Jespersen J. Involvement of the hemostatic system in the insulin resistance syndrome: a study of 1500 patients with angina pectoris. The ECAT Angina Pectoris Study Group. Arterioscler Thromb. 1993; 13: 1865–1873.
- 31.↵Festa A, D’Agostino R, Mykkanen L, et al. Relative contribution of insulin and its precursors to fibrinogen and PAI-1 in a large population with different states of glucose tolerance: the Insulin Resistance Atherosclerosis Study (IRAS). Arterioscler Thromb Vasc Biol. 1999; 19: 562–568.