1

Introduction

Incremental improvements in pharmacotherapy for percutaneous coronary intervention (PCI) continue to reduce the incidence of adverse clinical events for the treatment of both stable coronary artery disease and acute coronary syndromes (ACS). In addition to anticoagulation with either heparin or bivalirudin, platelet inhibition with dual antiplatelet therapy represented a major advancement in reducing acute and late vessel thrombosis . However, platelet inhibition through adenosine diphosphate or P2Y12 inhibition may have delayed onset, especially in the setting of ST-segment elevation myocardial infarction (STEMI) . This is further highlighted by a recent study that demonstrated that morphine significantly delays platelet inhibition with ticagrelor in patients with acute myocardial infarction (MI) . Intravenous glycoprotein IIb/IIIa inhibitors (GPIs) were introduced to provide rapid platelet inhibition and reduce peri-procedural adverse ischemic events associated with PCI. However, meta-analyses of different anticoagulation strategies for PCI have raised concern that benefits associated with GPI use may be outweighed by the risk of bleeding . This reflects major changes in practice patterns for pharmacotherapy during PCI with a significant decrease in routine GPI use and greater use of GPI as a provisional or “bailout” therapy.

Although newer P2Y12 inhibitors are associated with more rapid platelet inhibition compared with earlier therapy, such as ticlopidine and clopidogrel, adequate platelet inhibition for PCI remains a significant concern because of the prothrombotic milieu in ACS. Thus, adequate platelet inhibition early in the setting of PCI for ACS remains a major concern as it requires balancing antiplatelet effects to avoid acute stent thrombosis without increasing the risk of adverse bleeding events. Concern for increased bleeding is especially raised when considering upstream initiation of GPI in the setting of ACS . Despite there being previous analyses of GPI use, there is a paucity of outcomes data comparing the different GPI regimens available for clinical use with other anticoagulation regimens, such as heparin or bivalirudin with a provisional GPI regimen. For example, it is critical that a high-dose bolus (HBD) of tirofiban (25 mcg/kg) be utilized in combination with anticoagulation as it was shown to be superior to the previously employed low-dose bolus regimen (10 mcg/kg) . There are also inherent differences between the various GPI. Abciximab is a Fab fragment of the monoclonal antibody 7E3 that permanently binds the glycoprotein IIb/IIIa receptor on platelets, while the small molecule GPI, including tirofiban and eptifibatide, has a more rapid offset but requires renal dosing. Given these differences between the different GPIs, we performed a network meta-analysis of randomized controlled trials to compare heparin with provisional GPI, bivalirudin with provisional GPI, anticoagulation with routine HDB tirofiban, anticoagulation with routine abciximab, and anticoagulation with routine eptifibatide as the regimen for PCI and their impact on clinical outcomes.

2

Methods

The primary objective of this network meta-analysis was to compare the following anticoagulation strategies for PCI: (1) routine use of HDB tirofiban in combination with standard anticoagulation (HDB tirofiban), (2) routine use of abciximab in combination with standard anticoagulation (abciximab), (3) routine use of eptifibatide in combination with standard anticoagulation (eptifibatide), (4) heparin with provisional GPI use (heparin), and (5) bivalirudin with provisional GPI use (bivalirudin). For simplicity, we will refer from this point forward as a single name for the therapies described above. Standard anticoagulation therapy was defined as the use of heparin to maintain a therapeutic active clotting time, appropriate dosing of low molecular weight heparin, or standard dosing of bivalirudin. Two independent reviewers (MJL and RCL) systematically searched (1998 to February 2016) MEDLINE/PubMed, Cochrane CENTRAL, and ClinicalTrials.gov , applying the search terms “abciximab” OR “eptifibatide” OR “tirofiban” OR “bivalirudin” OR “glycoprotein IIb/IIIa inhibitor” and limited the search to clinical trials. We apply the prespecified inclusion criteria to include only: (a) patients from randomized controlled trials (RCTs) comparing the previously defined anticoagulation strategies for PCI, (b) with routine use of dual antiplatelet therapy, and (c) provided outcome measures and associated definitions. We excluded any study or group that: (a) included tirofiban that did not utilize HDB tirofiban (25 mcg/kg), (b) did not compare outcomes of at least two of the included groups, (c) utilized other treatment regimens (mechanical devices or thrombolytic therapies), (d) compared different means of administering (other than intravenous), or (e) compared different timing of GPI administration such as comparing early vs. delayed PCI strategies. Data were abstracted by the same two investigators (MJL and RCL) in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (eTable 1 in the supplement). We assessed studies for internal validity and risk of bias (eTable 2 in the supplement). We collected outcomes data for all-cause mortality, MI, major adverse cardiovascular events (MACE), major bleeding, minor bleeding, need for transfusion, and thrombocytopenia. We preferentially utilized 30-day outcomes when available. We accepted the study definitions for recurrent MI, MACE, major and minor bleeding, and thrombocytopenia, and all studies had event adjudication by physicians.

Dichotomous variables are reported as percentages, while continuous variables were reported as mean ± standard deviation or median (interquartile range). Mixed treatment comparison model generation was performed to directly and indirectly compare dichotomous outcomes for therapies using GeMTC 0.14.3 software (GeMTC, http://drugis.org/mtc ). Bayesian hierarchical random-effects model with directed acyclic graph model for general-purpose Markov chain Monte Carlo analysis was performed with 50,000 tuning iterations and 100,000 simulation iterations (GeMTC). Data are presented as odds ratios (OR) [credible intervals (CI)]. Convergence was appraised graphically according to Gelman and Rubin . Data from a consistency model are presented, and direction of findings was confirmed with an inconsistency model along with non-informative priors to serve as a sensitivity analysis. Additional sensitivity analyses were performed with removal of one study at a time to confirm directionality and magnitude of findings.

2

Methods

The primary objective of this network meta-analysis was to compare the following anticoagulation strategies for PCI: (1) routine use of HDB tirofiban in combination with standard anticoagulation (HDB tirofiban), (2) routine use of abciximab in combination with standard anticoagulation (abciximab), (3) routine use of eptifibatide in combination with standard anticoagulation (eptifibatide), (4) heparin with provisional GPI use (heparin), and (5) bivalirudin with provisional GPI use (bivalirudin). For simplicity, we will refer from this point forward as a single name for the therapies described above. Standard anticoagulation therapy was defined as the use of heparin to maintain a therapeutic active clotting time, appropriate dosing of low molecular weight heparin, or standard dosing of bivalirudin. Two independent reviewers (MJL and RCL) systematically searched (1998 to February 2016) MEDLINE/PubMed, Cochrane CENTRAL, and ClinicalTrials.gov , applying the search terms “abciximab” OR “eptifibatide” OR “tirofiban” OR “bivalirudin” OR “glycoprotein IIb/IIIa inhibitor” and limited the search to clinical trials. We apply the prespecified inclusion criteria to include only: (a) patients from randomized controlled trials (RCTs) comparing the previously defined anticoagulation strategies for PCI, (b) with routine use of dual antiplatelet therapy, and (c) provided outcome measures and associated definitions. We excluded any study or group that: (a) included tirofiban that did not utilize HDB tirofiban (25 mcg/kg), (b) did not compare outcomes of at least two of the included groups, (c) utilized other treatment regimens (mechanical devices or thrombolytic therapies), (d) compared different means of administering (other than intravenous), or (e) compared different timing of GPI administration such as comparing early vs. delayed PCI strategies. Data were abstracted by the same two investigators (MJL and RCL) in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (eTable 1 in the supplement). We assessed studies for internal validity and risk of bias (eTable 2 in the supplement). We collected outcomes data for all-cause mortality, MI, major adverse cardiovascular events (MACE), major bleeding, minor bleeding, need for transfusion, and thrombocytopenia. We preferentially utilized 30-day outcomes when available. We accepted the study definitions for recurrent MI, MACE, major and minor bleeding, and thrombocytopenia, and all studies had event adjudication by physicians.

Dichotomous variables are reported as percentages, while continuous variables were reported as mean ± standard deviation or median (interquartile range). Mixed treatment comparison model generation was performed to directly and indirectly compare dichotomous outcomes for therapies using GeMTC 0.14.3 software (GeMTC, http://drugis.org/mtc ). Bayesian hierarchical random-effects model with directed acyclic graph model for general-purpose Markov chain Monte Carlo analysis was performed with 50,000 tuning iterations and 100,000 simulation iterations (GeMTC). Data are presented as odds ratios (OR) [credible intervals (CI)]. Convergence was appraised graphically according to Gelman and Rubin . Data from a consistency model are presented, and direction of findings was confirmed with an inconsistency model along with non-informative priors to serve as a sensitivity analysis. Additional sensitivity analyses were performed with removal of one study at a time to confirm directionality and magnitude of findings.

3

Results

Using MEDLINE/PubMed, Cochrane CENTRAL, and ClinicalTrials.gov , we identified 8125 citations using the previously defined search terms. Implementing our inclusion/exclusion criteria, we evaluated 328 abstracts, of which we assessed 167 as full-text publications. We excluded studies in addition to the criteria listed above because of duplication of data, studies with GPIs other than those previously mentioned , GPI use in studies outside of PCI, studies in patients with fibrinolytic therapy, studies with mechanical devices (aspiration thrombectomy) in addition to GPI or anticoagulation, studies utilizing several GPIs in addition to heparin in one treatment arm , and studies comparing early vs. late invasive strategies, or lack of dual antiplatelet therapy. Our search flow diagram can be found in Fig. 1 , and the network profile of studies can be found in eFig. 1 in the supplement. For our meta-analysis, we included 41 RCTs with 38,645 patients with 2654 randomized to HDB tirofiban , 6752 to abciximab , 1669 to eptifibatide , 16,500 to heparin , and 11,070 to bivalirudin . Baseline study characteristics can be found in Table 1 . There was robust comparison between the treatment arms with the least data comparison between bivalirudin and GPI therapies and the least number of studies with eptifibatide. Due to inclusion of randomized controlled trials, there was generally low risk of study bias. Among the included patients in our analysis, mean age was 64 ± 11 years, 75% were male, 61% had hypertension, 52% had hyperlipidemia, 25% had diabetes mellitus, 33% were smokers, and 20% had prior MI. Baseline patient characteristics can be found in Table 2 . As seen in Table 3 , 46% of included patients presented with STEMI, 74% presented with ACS, 91% underwent stenting during PCI, 72% of patients were treated with clopidogrel, 9% were treated with ticlopidine, 8% were treated with prasugrel, and 10% were treated with ticagrelor.

Study flow diagram.

Table 2

Patient characteristics of included RCTs.

Study	Therapies	Age	Male (%)	Hypertension (%)	Hyperlipidemia (%)	Diabetes Mellitus (%)	Smoker (%)	Prior MI (%)
3T/2R	T v H	68 ± 10	73	71	54	26	15	43
ADVANCE	T v H	69 ± 8	68	48	NR	49	20	49
Danzi et al.	T v A	60 ± 12	83	51	49	17	62	NR
Ernst et al.	T v A	61 ± 11	73	37	22	13	48	4
EVEREST	T v A	63 ± 11	70	59	49	16	26	21
FABOLUS PRO	T v H	68 ± 11	69	63	49	20	59	16
FATA	T v A	64 ± 13	75	55	47	18	40	4
Jia et al.	T v H	59 ± 10	78	52	NR	17	72	4
MR PCI	T v E	58 ± 12	85	42	45	35	27	50
MULTI-STRATEGY	T v A	64 ± 13	76	57	53	15	37	8
ON-TIME 2	T v H	62 ± 12	76	34	26	11	47	9
OPTIMIZE-IT	T v H	66 ± 7	72	78	70	100	NR	17
Salarifar et al.	T v H	57 ± 10	79	36	26	29	44	NR
STRATEGY	T v A	63 (55, 72)	73	53	NR	15	40	11
TAMIP	T v A	52	NR	42	NR	38	57	NR
TENACITY	T v A	63 ± 11	73	80	84	31	27	31
ACE	A v H	64 (54, 74)	77	47	40	18	NR	11
ADMIRAL	A v H	61 ± 13	82	38	38	18	42	11
BRAVE 3	A v H	62 ± 12	74	70	43	18	42	10
CADILLAC	A v H	60 (51, 69)	73	48	38	17	43	14
CACHET	A v B	63 ± 11	76	NR	NR	NR	NR	NR
EVA-AMI	A v E	61 ± 13	78	50	42	16	46	8
ISAR-2	A v H	61 ± 12	76	64	43	15	43	NR
ISAR-REACT	A v H	66 ± 10	76	54	53	20	17	33
ISAR-REACT 2	A v H	66 ± 11	75	63	61	27	22	24
ISAR-REACT 4	A v B	68 ± 11	77	85	69	29	24	20
ISAR-SWEET	A v H	68 ± 9	74	71	58	100	16	34
ASSIST	E v H	61 ± 12	76	48	36	16	42	12
CLEAR PLATELETS-2	E v B	64 ± 13	61	79	77	40	46	34
ESPRIT	E v H	62 (54, 71)	73	59	58	20	23	32
INSTANT	E v H	65 ± 10	81	74	71	31	54	30
ARMYDA-7 BIVALVE	H v B	70 ± 9	72	91	NR	63	15	36
ARNO	H v B	69 ± 11	76	59	49	22	17	39
BRAVE 4	H v B	61	77	65	54	16	62	9
BRIGHT	H v B	58 ± 14	82	42	37	21	62	5
EUROMAX	H v B	62	76	44	37	13	42	9
HEAT PPCI	H v B	63 (54, 74)	72	42	38	14	42	12
ISAR-REACT 3	H v B	67 ± 10	76	89	80	27	15	31
MATRIX	H v B	65 ± 12	76	62	44	22	36	14
NAPLES III	H v B	78 ± 4	65	76	64	100	21	45
Xiang et al.	H v B	58	83	NR	NR	NR	NR	NR

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Time to start implementing Lean and Six Sigma in the catheterization laboratory
2. Reply: Time to start implementing lean and six sigma in the catheterization laboratory
3. Pre-hospital ticagrelor in patients with ST-segment elevation myocardial infarction with long transport time to primary PCI facility
4. Unfavorable bioresorbable vascular scaffold resorption, a cause of restenosis?

Stay updated, free articles. Join our Telegram channel

Join