Abstract
To determine whether the low-molecular weight heparin enoxaparin remains favourable when compared with unfractionated heparin (UFH) among patients with acute coronary syndromes (ACS) when incorporating efficacy and safety of these adjunctive therapies using a net clinical endpoint.
We performed a meta-analysis of randomized trials of enoxaparin vs. UFH in ST-elevation-MI (STEMI) or non-ST-elevation-ACS (NSTEACS) (n = 49 088 patients in 12 trials). The net clinical endpoint was defined as death, MI, or major bleeding by 30 days. Death or myocardial infarction (MI) was significantly reduced with enoxaparin when compared with UFH (9.8 vs. 11.4%, OR 0.84, P < 0.001). The net clinical endpoint occurred less frequently with enoxaparin than UFH (12.5 vs. 13.5%, OR 0.90, P = 0.051). Major bleeding was higher with enoxaparin (4.3 vs. 3.4%, OR 1.25, P = 0.019). Among STEMI trials, the net clinical endpoint was significantly lower with enoxaparin (OR 0.84, P = 0.015), but there was no difference in NSTEACS trials (OR 0.97).
When compared with UFH, enoxaparin was associated with superior efficacy as adjunctive antithrombin therapy among >49 000 patients across the ACS spectrum. Although bleeding was increased with enoxaparin, this increase was offset by a reduction in death or MI. The net clinical benefit in favour of enoxaparin was evident among the STEMI population and was neutral among the NSTEACS population.
Introduction
Antithrombin therapy is an established, guideline-recommended treatment, and is central to the management of patients with acute coronary syndromes (ACS). However, the optimal antithrombin agent is still debated.1,2 Some, but not all large-scale trials in patients with both ST-segment elevation myocardial infarction (STEMI) and non-ST-elevation ACS (NSTEACS) have shown improved efficacy with the low-molecular weight heparin enoxaparin when compared with unfractionated heparin (UFH), but safety concerns of increased bleeding with enoxaparin have also been reported.3–10 The balance between efficacy and safety has been a focus of discussion regarding the choice of optimal antithrombin agent.11–13
The goal of this analysis was to determine whether enoxaparin remains favourable when compared with UFH among patients with ACS when incorporating the efficacy and safety profile of these adjunctive therapies by performing a meta-analysis using a composite net clinical endpoint.
Methods
We performed a PubMed search for randomized clinical trials comparing enoxaparin with UFH among patients with STEMI or NSTEACS. We also performed a hand search of references from the original manuscripts and prior meta-analyses. Inclusion criteria for the analysis were all trials that were: (i) randomized, (ii) compared enoxaparin with UFH, and (iii) were conducted in patients with STEMI or NSTEACS. Trials conducted exclusively in the STEMI population were restricted to those in which patients were treated with aspirin and fibrinolytic therapy (ASSENT 3,3 HART II,14 Baird et al.,15 ENTIRE-TIMI 23,16 ASSENT 3 Plus,17 and ExTRACT-TIMI 254). Figure 1 contains the study flow chart for the meta-analysis.
Study flow chart for trial review and selection for inclusion in the meta-analysis.
Study flow chart for trial review and selection for inclusion in the meta-analysis.
Data were abstracted for each trial from the manuscript and were sent to the corresponding author for verification, as well as clarification of any missing data. When data were not available from the corresponding author, the sponsoring pharmaceutical company was queried for missing data and clarification of data. The primary data sets for ESSENCE,18 TIMI 11B,19 ENTIRE-TIMI 23,16 A to Z,20 and ExTRACT-TIMI 254 were available at the TIMI Study Group.
Endpoints
The composite endpoint of the analysis was net clinical events, defined as death, non-fatal MI, or non-fatal major bleeding by 30 days, or the closest timepoint available to 30 days. If a subject met more than one part of the composite event, only one event was counted. The components of the net clinical composite endpoint were also evaluated individually. The net clinical endpoint was not available from the ACUTE II trial, but individual components of the endpoint as well as death or MI were available and included in all other analyses. Differences in the timepoints used in the individual trials are noted in Table 1. Eight of the 12 trials used the TIMI major bleeding criteria, defined as observed bleeding with a decrease in haemoglobin of more than 5 mg/dL or intracranial or pericardial bleeding. The definition of major bleeding for the remaining four trials is shown in Table 1. For trials that did not include intracranial hemorrhage (ICH) as part of the major bleeding definition in the main manuscript (INTERACT, Baird et al., ASSENT 3, and ASSENT 3 Plus), data were obtained from the lead authors to include ICH as a major bleeding event in order to provide a more comparable definition of major bleeding as well as a more comprehensive evaluation of safety as part of the net clinical endpoint.
Trial designs
| Trial | Population | n | Year published | Blinding | Randomization arms | Endpoint description | |
|---|---|---|---|---|---|---|---|
| Enoxaparin | UFH | ||||||
| ST-elevation MI | |||||||
| ASSENT 33 | STEMI | 4075 | 2001 | Open-label | 30 mg bolus; 1 mg/kg bid for ≤7 days | 60 U/kg bolus, 12 U/kg/h for 48 h to aPTT 50–70 s | Death 30 days; MI in-hospital; major bleeding (requiring transfusion or intervention due to haemodynamic compromise or ICH) in-hospital |
| HART II14 | STEMI | 400 | 2001 | Open-label | 30 mg bolus, 1 mg/kg bid for ≥3 days | 4000–5000 U bolus, 15 U/kg/h for ≥3 days to aPTT 2.0–2.5 × control | Death 30 days; MI 30 days; TIMI major bleeding in-hospitala |
| Baird et al.15 | STEMI | 300 | 2002 | Open-label | 40 mg bolus, 40 mg tid for 4 days | 5000 U bolus, 30 000 U infusion over 24 h for 4 days to aPTT 2.0–2.5 × control | Death 90 days; MI 90 days; major bleeding (clinically significant haemorrhage or ICH) on study drug |
| ENTIRE-TIMI 2316 | STEMI | 242 | 2002 | Open-label | 0 or 30 mg bolus; 1 mg/kg bid for ≤8 days | 60 U/kg bolus, 12 U/kg/h for ≤3 days to aPTT 1.5–2.5 × control | Death 30 days; MI 30 days; TIMI major bleeding 30 daysa |
| ASSENT 3 Plus17 | STEMI | 1635 | 2003 | Open-label | 30 mg bolus; 1 mg/kg bid for ≤7 days | 60 U/kg bolus, 12 U/kg/h for 48 h to aPTT 50–70 s | Death 30 days; MI in-hospital; major bleeding (requiring transfusion or intervention because of haemodynamic compromise or ICH) in-hospital |
| ExTRACT-TIMI 254 | STEMI | 20 479 | 2006 | Double-blind | 30 mg bolus (if age <75); 1 mg/kg bid (if age <75) or 0.75 mg/kg bid (if age ≥75) for ≤8 days | 60 U/kg bolus (omitted if open-label UFH received within 3 h), 12 U/kg/h for ≥3 days to aPTT 1.5–2.0 × control | Death 30 days; MI 30 days; TIMI major bleeding 30 daysa |
| Non-ST elevation acute coronary syndromes | |||||||
| ESSENCE18 | NSTEACS | 3171 | 1997 | Double-blind | 1 mg/kg bid for ≤4 days | 5000 U bolus, infusion dose adjusted to aPTT 55–85 s for ≥48 h | Death 30 days; MI 30 days; TIMI major bleeding 30 daysa |
| TIMI 11B19 | NSTEACS | 3910 | 1999 | Double-blind | 30 mg bolus, 1 mg/kg bid for ≤8 days | 70 U/kg bolus, 15 U/kg/h for ≥3 days to aPTT 1.5–2.5 × control | Death 30 days; MI 30 days; TIMI major bleeding 30 daysa |
| ACUTE II31 | NSTEACS | 525 | 2002 | Double-blind | 1 mg/kg bid for ≤4 days | 5000 U bolus, 1000 U/h for ≤4 days to aPTT 1.5–2.5 × control | Death 30 days; MI 30 days; TIMI major bleeding 30 daysa |
| INTERACT23 | NSTEACS | 746 | 2003 | Open-label | 1 mg/kg bid for 48 h | 70 U/kg bolus, 15 U/kg/h for 48 h to aPTT 1.5–2.0 × control | Death 30 days; MI 30 days; major bleeding (retroperitoneal haemorrhage, or bleeding at a specific site accompanied by a 3 g/dL drop in haemoglobin or resulting in death or ICH) 30 days |
| A to Z20 | NSTEACS | 3618 | 2004 | Open-label | 1 mg/kg bid for ≤5 days | 4000 U bolus, 900 U/h if ≥70 kg; 60 U/kg bolus, 12 U/kg/h if <70 kg; aPTT 50–70 s | Death 30 days; MI 30 days; TIMI major bleeding through 24 h after tirofiban discontinuationa |
| SYNERGY6 | NSTEACS | 9975 | 2004 | Open-label | 1 mg/kg bid | 60 U/kg bolus, 12 U/kg/h to aPTT 1.5–2.0 × control or 50–70 s | Death 30 days; MI 30 days; TIMI major bleeding in-hospitala |
| Trial | Population | n | Year published | Blinding | Randomization arms | Endpoint description | |
|---|---|---|---|---|---|---|---|
| Enoxaparin | UFH | ||||||
| ST-elevation MI | |||||||
| ASSENT 33 | STEMI | 4075 | 2001 | Open-label | 30 mg bolus; 1 mg/kg bid for ≤7 days | 60 U/kg bolus, 12 U/kg/h for 48 h to aPTT 50–70 s | Death 30 days; MI in-hospital; major bleeding (requiring transfusion or intervention due to haemodynamic compromise or ICH) in-hospital |
| HART II14 | STEMI | 400 | 2001 | Open-label | 30 mg bolus, 1 mg/kg bid for ≥3 days | 4000–5000 U bolus, 15 U/kg/h for ≥3 days to aPTT 2.0–2.5 × control | Death 30 days; MI 30 days; TIMI major bleeding in-hospitala |
| Baird et al.15 | STEMI | 300 | 2002 | Open-label | 40 mg bolus, 40 mg tid for 4 days | 5000 U bolus, 30 000 U infusion over 24 h for 4 days to aPTT 2.0–2.5 × control | Death 90 days; MI 90 days; major bleeding (clinically significant haemorrhage or ICH) on study drug |
| ENTIRE-TIMI 2316 | STEMI | 242 | 2002 | Open-label | 0 or 30 mg bolus; 1 mg/kg bid for ≤8 days | 60 U/kg bolus, 12 U/kg/h for ≤3 days to aPTT 1.5–2.5 × control | Death 30 days; MI 30 days; TIMI major bleeding 30 daysa |
| ASSENT 3 Plus17 | STEMI | 1635 | 2003 | Open-label | 30 mg bolus; 1 mg/kg bid for ≤7 days | 60 U/kg bolus, 12 U/kg/h for 48 h to aPTT 50–70 s | Death 30 days; MI in-hospital; major bleeding (requiring transfusion or intervention because of haemodynamic compromise or ICH) in-hospital |
| ExTRACT-TIMI 254 | STEMI | 20 479 | 2006 | Double-blind | 30 mg bolus (if age <75); 1 mg/kg bid (if age <75) or 0.75 mg/kg bid (if age ≥75) for ≤8 days | 60 U/kg bolus (omitted if open-label UFH received within 3 h), 12 U/kg/h for ≥3 days to aPTT 1.5–2.0 × control | Death 30 days; MI 30 days; TIMI major bleeding 30 daysa |
| Non-ST elevation acute coronary syndromes | |||||||
| ESSENCE18 | NSTEACS | 3171 | 1997 | Double-blind | 1 mg/kg bid for ≤4 days | 5000 U bolus, infusion dose adjusted to aPTT 55–85 s for ≥48 h | Death 30 days; MI 30 days; TIMI major bleeding 30 daysa |
| TIMI 11B19 | NSTEACS | 3910 | 1999 | Double-blind | 30 mg bolus, 1 mg/kg bid for ≤8 days | 70 U/kg bolus, 15 U/kg/h for ≥3 days to aPTT 1.5–2.5 × control | Death 30 days; MI 30 days; TIMI major bleeding 30 daysa |
| ACUTE II31 | NSTEACS | 525 | 2002 | Double-blind | 1 mg/kg bid for ≤4 days | 5000 U bolus, 1000 U/h for ≤4 days to aPTT 1.5–2.5 × control | Death 30 days; MI 30 days; TIMI major bleeding 30 daysa |
| INTERACT23 | NSTEACS | 746 | 2003 | Open-label | 1 mg/kg bid for 48 h | 70 U/kg bolus, 15 U/kg/h for 48 h to aPTT 1.5–2.0 × control | Death 30 days; MI 30 days; major bleeding (retroperitoneal haemorrhage, or bleeding at a specific site accompanied by a 3 g/dL drop in haemoglobin or resulting in death or ICH) 30 days |
| A to Z20 | NSTEACS | 3618 | 2004 | Open-label | 1 mg/kg bid for ≤5 days | 4000 U bolus, 900 U/h if ≥70 kg; 60 U/kg bolus, 12 U/kg/h if <70 kg; aPTT 50–70 s | Death 30 days; MI 30 days; TIMI major bleeding through 24 h after tirofiban discontinuationa |
| SYNERGY6 | NSTEACS | 9975 | 2004 | Open-label | 1 mg/kg bid | 60 U/kg bolus, 12 U/kg/h to aPTT 1.5–2.0 × control or 50–70 s | Death 30 days; MI 30 days; TIMI major bleeding in-hospitala |
aTIMI major bleeding defined as a decrease in haemoglobin of more than 5 mg/dL or intracranial or pericardial bleeding.
Statistical analysis
A meta-analysis was performed of the relative odds based on random-effects models using the method by Der Simonian and Laird.21 A test of heterogeneity, which evaluates variability in the treatment effects, was performed using the Mantel–Haenszel method.22 Results are presented as odds ratios (OR) with their 95% confidence intervals (CIs) and P-values. Event rates for each trial individually and the pooled data are presented as frequencies. A P-value of <0.05 was considered statistically significant. All statistical analyses were performed using Stata/SE, version 9.0 (StataCorp, College Station, TX, USA).
Results
The design and endpoint descriptions for each of the 12 trials (n = 49 088) are shown in Table 1. Among the STEMI trials (n = 27 131), one (ExTRACT-TIMI 25), involving the majority of subjects (n = 20 479) was double-blind and the remaining five were open-label (Table 1). Use of an enoxaparin bolus and the dosing of the bolus varied across trials (Table 1). Enoxaparin subcutaneous injection dosing was consistent at 1 mg/kg bid, with the exception of the trial by Baird et al.,15 which used a 40 mg tid dose. In the ExTRACT-TIMI 25 trial, the maintenance injection was reduced to 0.75 mg/kg in subjects age ≥75 years and to once daily in patients with creatinine clearance <30 mL/min.4 The duration of enoxaparin treatment ranged from 3 to 8 days in the trials. UFH bolus and infusion were constant at 60 U/kg bolus and 12 U/kg/h infusion, with the exception of the HART II trial,14 which used a 15 U/kg/h infusion.
Among the NSTEACS trials (n = 21 945), three were double-blind and three were open-label (Table 1). Enoxaparin dosing was 1 mg/kg bid in all trials, with the duration of treatment ranging from 2 to 8 days. The UFH dosing varied, with earlier trials having slightly higher bolus and infusion doses.
The majority of patients in the STEMI trials were treated with fibrin-specific lytics (Table 2). Among the NSTEACS trials, SYNERGY,6 A to Z,20 and INTERACT23 trials enrolled the highest risk patients, with more than 80% biomarker positive (Table 3).
Baseline characteristics: ST-elevation MI
| ASSENT 33 | HART II14 | Baird et al.15 | ENTIRE-TIMI 2316 | ASSENT 3 Plus17 | ExTRACT-TIMI 254 | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Enox n = 2040 | UFH n = 2038 | Enox n = 200 | UFH n = 200 | Enox n = 149 | UFH n = 151 | Enox n = 160 | UFH n = 82 | Enox n = 818 | UFH n = 821 | Enox n = 10 256 | UFH n = 10 223 | |
| Age (years) | 61 (12) | 61 (13) | 60 | 61 | 62 (12) | 62 (10) | 57 (10) | 57 (10) | 62 (13) | 62 (13) | 60 (12) | 60 (12) |
| Females, n (%) | 463 (23) | 478 (23) | 44 (22) | 52 (26) | 40 (27) | 41 (27) | 26 (16) | 13 (16) | 194 (24) | 184 (22) | 2415 (24) | 2368 (23) |
| Diabetes, n (%) | 381 (19) | 363 (18) | 31 (15) | 23 (11) | 16 (11) | 13 (9) | 24 (15) | 13 (16) | 115 (14) | 128 (16) | 1545 (15) | 1515 (15) |
| Heart rate (b.p.m.) | 75 (17) | 74 (17) | N/A | N/A | N/A | N/A | 71 (16) | 72 (16) | 74 (19) | 74 (19) | 76 (17) | 76 (16) |
| Systolic blood pressure (mmHg) | 134 (22) | 133 (23) | N/A | N/A | 130 (24) | 129 (24) | 134 (19) | 140 (21) | 134 (25) | 132 (25) | 133 (21) | 134 (21) |
| Fibrin-specific lytic (%) | 2012 (99) | 2002 (98) | 200 (100) | 200 (100) | 52 (35) | 48 (32) | 160 (100) | 82 (100) | 802 (98) | 806 (98) | 8143 (80) | 8141 (80) |
| ASSENT 33 | HART II14 | Baird et al.15 | ENTIRE-TIMI 2316 | ASSENT 3 Plus17 | ExTRACT-TIMI 254 | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Enox n = 2040 | UFH n = 2038 | Enox n = 200 | UFH n = 200 | Enox n = 149 | UFH n = 151 | Enox n = 160 | UFH n = 82 | Enox n = 818 | UFH n = 821 | Enox n = 10 256 | UFH n = 10 223 | |
| Age (years) | 61 (12) | 61 (13) | 60 | 61 | 62 (12) | 62 (10) | 57 (10) | 57 (10) | 62 (13) | 62 (13) | 60 (12) | 60 (12) |
| Females, n (%) | 463 (23) | 478 (23) | 44 (22) | 52 (26) | 40 (27) | 41 (27) | 26 (16) | 13 (16) | 194 (24) | 184 (22) | 2415 (24) | 2368 (23) |
| Diabetes, n (%) | 381 (19) | 363 (18) | 31 (15) | 23 (11) | 16 (11) | 13 (9) | 24 (15) | 13 (16) | 115 (14) | 128 (16) | 1545 (15) | 1515 (15) |
| Heart rate (b.p.m.) | 75 (17) | 74 (17) | N/A | N/A | N/A | N/A | 71 (16) | 72 (16) | 74 (19) | 74 (19) | 76 (17) | 76 (16) |
| Systolic blood pressure (mmHg) | 134 (22) | 133 (23) | N/A | N/A | 130 (24) | 129 (24) | 134 (19) | 140 (21) | 134 (25) | 132 (25) | 133 (21) | 134 (21) |
| Fibrin-specific lytic (%) | 2012 (99) | 2002 (98) | 200 (100) | 200 (100) | 52 (35) | 48 (32) | 160 (100) | 82 (100) | 802 (98) | 806 (98) | 8143 (80) | 8141 (80) |
Baseline characteristics: non-ST-elevation acute coronary syndromes
| ESSENCE18 | TIMI 11B19 | ACUTE II31 | INTERACT23 | A to Z20 | SYNERGY6 | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Enox n = 1607 | UFH n = 1564 | Enox n = 1953 | UFH n = 1957 | Enox n = 315 | UFH n = 210 | Enox n = 380 | UFH n = 366 | Enox n = 2026 | UFH n = 1961 | Enox n = 4993 | UFH n = 4985 | |
| Age (years) | 63 (12) | 64 (11) | 64 (12) | 64 (11) | 65 (12) | 64 (13) | 64a | 64a | 60 (11) | 61 (11) | 68a | 68a |
| Females, n (%) | 528 (33) | 531 (34) | 677 (35) | 701 (36) | 108 (34) | 69 (33) | 121 (32) | 112 (31) | 580 (29) | 564 (29) | 1696 (34) | 1684 (34) |
| Diabetes, n (%) | 360 (22) | 339 (22) | 385 (20) | 393 (20) | 75 (24) | 45 (21) | 84 (22) | 85 (23) | 395 (20) | 356 (18) | 1424 (29) | 1502 (30) |
| ECG changes, n (%)8 | 897 (56) | 895 (57) | 1611 (83) | 1626 (83) | N/A | N/A | 88 (23) | 79 (22) | 1430 (70) | 1410 (72) | 3904 (78) | 3941 (79) |
| Biomarker positive, n (%)8 | N/A | N/A | 738 (38) | 775 (40) | 187 (59) | 122 (58) | 311 (82) | 312 (85) | 1627 (80) | 1563 (80) | 4198 (84) | 4190 (84) |
| Cardiac catheterization performed, n (%) | 612 (38) | 646 (41) | 794 (41) | 840 (43) | 187 (59) | 126 (60) | 236 (62) | 237 (65) | 1224 (60) | 1210 (62) | 4600 (92) | 4588 (92) |
| PCI performed, n (%) | 161 (10) | 206 (13) | 243 (12) | 271 (14) | 89 (28) | 66 (31) | 103 (27) | 111 (30) | 528 (26) | 514 (26) | 2323 (47) | 2364 (47) |
| CABG performed, n (%) | 99 (6) | 105 (7) | 118 (6) | 137 (7) | 49 (16) | 40 (19) | 48 (13) | 45 (12) | 178 (9) | 198 (10) | 965 (19) | 899 (18) |
| ESSENCE18 | TIMI 11B19 | ACUTE II31 | INTERACT23 | A to Z20 | SYNERGY6 | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Enox n = 1607 | UFH n = 1564 | Enox n = 1953 | UFH n = 1957 | Enox n = 315 | UFH n = 210 | Enox n = 380 | UFH n = 366 | Enox n = 2026 | UFH n = 1961 | Enox n = 4993 | UFH n = 4985 | |
| Age (years) | 63 (12) | 64 (11) | 64 (12) | 64 (11) | 65 (12) | 64 (13) | 64a | 64a | 60 (11) | 61 (11) | 68a | 68a |
| Females, n (%) | 528 (33) | 531 (34) | 677 (35) | 701 (36) | 108 (34) | 69 (33) | 121 (32) | 112 (31) | 580 (29) | 564 (29) | 1696 (34) | 1684 (34) |
| Diabetes, n (%) | 360 (22) | 339 (22) | 385 (20) | 393 (20) | 75 (24) | 45 (21) | 84 (22) | 85 (23) | 395 (20) | 356 (18) | 1424 (29) | 1502 (30) |
| ECG changes, n (%)8 | 897 (56) | 895 (57) | 1611 (83) | 1626 (83) | N/A | N/A | 88 (23) | 79 (22) | 1430 (70) | 1410 (72) | 3904 (78) | 3941 (79) |
| Biomarker positive, n (%)8 | N/A | N/A | 738 (38) | 775 (40) | 187 (59) | 122 (58) | 311 (82) | 312 (85) | 1627 (80) | 1563 (80) | 4198 (84) | 4190 (84) |
| Cardiac catheterization performed, n (%) | 612 (38) | 646 (41) | 794 (41) | 840 (43) | 187 (59) | 126 (60) | 236 (62) | 237 (65) | 1224 (60) | 1210 (62) | 4600 (92) | 4588 (92) |
| PCI performed, n (%) | 161 (10) | 206 (13) | 243 (12) | 271 (14) | 89 (28) | 66 (31) | 103 (27) | 111 (30) | 528 (26) | 514 (26) | 2323 (47) | 2364 (47) |
| CABG performed, n (%) | 99 (6) | 105 (7) | 118 (6) | 137 (7) | 49 (16) | 40 (19) | 48 (13) | 45 (12) | 178 (9) | 198 (10) | 965 (19) | 899 (18) |
aMedian.
Clinical outcomes
The funnel plot shown in Figure 2 demonstrates the treatment effect found in each trial plotted against the size of the trial. The plot shows general symmetry with the exception of the ENTIRE-TIMI 23 trial, which had a small sample size and a strong treatment effect for enoxaparin when compared with UFH. The ExTRACT-TIMI 25 trial was the largest study but did not have a disproportionate treatment effect in relation to the majority of the trials. Although the ENTIRE-TIMI 23 trial was an outlier, the trial was included in the main analysis. A sensitivity analysis in which each of these two trials was excluded showed consistent findings with the main analysis.
Funnel plot demonstrating the treatment effect found in each trial plotted against the size of the trial. The plot shows general symmetry with the exception of the ENTIRE-TIMI 23 trial, which had a small sample size and a strong treatment effect for enoxaparin compared with unfractionated heparin (UFH). The ExTRACT-TIMI 25 trial was the largest study but did not have a disproportionate treatment effect in relation to the majority of the trials.
Funnel plot demonstrating the treatment effect found in each trial plotted against the size of the trial. The plot shows general symmetry with the exception of the ENTIRE-TIMI 23 trial, which had a small sample size and a strong treatment effect for enoxaparin compared with unfractionated heparin (UFH). The ExTRACT-TIMI 25 trial was the largest study but did not have a disproportionate treatment effect in relation to the majority of the trials.
Across the entire spectrum of ACS (STEMI and NSTEACS; n = 49 088), the composite efficacy endpoint of death or non-fatal MI was reduced among enoxaparin subjects when compared with UFH subjects (9.8 vs. 11.4%, OR 0.84, 95% CI 0.76–0.92, P < 0.001; Figure 3). The composite net clinical endpoint of death, non-fatal MI, or non-fatal major bleeding also occurred less frequently with enoxaparin when compared with UFH (12.5 vs. 13.5%, OR 0.90, 95% CI 0.81–1.003, P = 0.051; Figure 4). For the net clinical endpoint, evidence of heterogeneity between trials was observed (P = 0.006), as well as heterogeneity between STEMI and NSTEACS syndromes (P = 0.005).
Enoxaparin vs. UFH for the comparison of death or non-fatal myocardial infarction (MI), displayed using a random effects model. Black squares represent odds ratios (ORs), the size of which reflects the statistical weight of a trial in calculating the OR. The horizontal lines represent 95% confidence intervals (CIs). There was evidence of heterogeneity between ST-segment elevation myocardial infarction (STEMI) and non-ST-segment elevation acute coronary syndromes (NSTEACS) (P = 0.005).
Enoxaparin vs. UFH for the comparison of death or non-fatal myocardial infarction (MI), displayed using a random effects model. Black squares represent odds ratios (ORs), the size of which reflects the statistical weight of a trial in calculating the OR. The horizontal lines represent 95% confidence intervals (CIs). There was evidence of heterogeneity between ST-segment elevation myocardial infarction (STEMI) and non-ST-segment elevation acute coronary syndromes (NSTEACS) (P = 0.005).
Enoxaparin vs. UFH for the comparison of death, non-fatal myocardial infarction, or non-fatal major bleed, displayed using a random effects model.
Enoxaparin vs. UFH for the comparison of death, non-fatal myocardial infarction, or non-fatal major bleed, displayed using a random effects model.
Among the STEMI cohort (n = 27 131), the net clinical composite event rate was lower with enoxaparin (11.1 vs. 12.9%, OR 0.84, 95% CI 0.73–0.97, P = 0.018), with no significant evidence of heterogeneity between trials (P = 0.143). The reduction in death, MI, or major bleeding among the STEMI cohort was evident even when excluding individual trials in a sensitivity analysis. When excluding the largest trial, ExTRACT-TIMI 25, consistent results were observed for the net clinical composite event rate (11.5 vs. 13.2%, OR 0.80, 95% CI 0.62–1.04, P = 0.09).
There was no difference in the net clinical event rate in the NSTEACS trials (14.1 vs. 14.3%, OR 0.97, 95% CI 0.86–1.09, P = 0.607), with no significant evidence of heterogeneity between trials (P = 0.132).
Individual endpoints
Considering individual endpoints for the entire ACS spectrum, mortality was not significantly different with enoxaparin when compared with UFH (5.0 vs. 5.3%, OR 0.94, 95% CI 0.87–1.02, P = 0.14; Figure 5); MI was significantly lower with enoxaparin (5.5 vs. 6.9%, OR 0.75, 95% CI 0.65–0.86, P < 0.001; Figure 6); major bleeding was significantly higher with enoxaparin (4.3 vs. 3.4%, OR 1.25, 95% CI 1.04–1.50, P = 0.019; Figure 7).
Enoxaparin vs. UFH for the comparison of death, displayed using a random effects model.
Enoxaparin vs. UFH for the comparison of death, displayed using a random effects model.
Enoxaparin vs. UFH for the comparison of myocardial infarction, displayed using a random effects model.
Enoxaparin vs. UFH for the comparison of myocardial infarction, displayed using a random effects model.
Enoxaparin vs. UFH for the comparison of major bleed, displayed using a random effects model.
Enoxaparin vs. UFH for the comparison of major bleed, displayed using a random effects model.
Results were similar in the STEMI cohort for the comparison of enoxaparin with UFH, respectively, with mortality of 6.6 and 7.1% (OR 0.92, 95% CI 0.84–1.01, P = 0.097); MI 3.4 and 5.1% (OR 0.64, 95% CI 0.52–0.78, P < 0.001); and major bleeding 2.6 and 1.8% (OR 1.45, 95% CI 1.23–1.72, P < 0.001). Death or MI occurred in 9.6% of enoxaparin subjects and 11.7% of UFH subjects (OR 0.78, 95% CI 0.67–0.91, P = 0.002).
In patients with NSTEACS, there was no difference in mortality between enoxaparin and UFH (3.0% each, OR 0.99, 95% CI 0.83–1.18, P = 0.890). MI occurred significantly less frequently in the enoxaparin group (8.0 vs. 9.1%, OR 0.87, 95% CI 0.79–0.96, P = 0.005), as did the composite of death or non-fatal MI (10.0 vs. 11.0%, OR 0.90, 95% CI 0.81–0.996, P = 0.043). Major bleeding did not differ between treatment groups (6.3% with enoxaparin vs. 5.4% with UFH, OR 1.13, 95% CI 0.84–1.54, P = 0.419).
Discussion
When compared with UFH, enoxaparin was associated with superior efficacy as adjunctive antithrombin therapy among more than 49 000 patients across the ACS spectrum. Although major bleeding was increased with enoxaparin, this increase was offset by a significant reduction in death or non-fatal MI.
Antithrombotic therapy acts in part by reducing the risk of reocclusion of initially successfully reperfused infarct arteries in the setting of STEMI and preventing further thrombus formation in NSTEACS by inhibiting thrombin generation and/or activity.24–27 Enoxaparin intervenes more proximally in the coagulation cascade to a greater extent than UFH (greater anti-factor Xa to anti-factor IIa activity),24–27 presumably resulting in a reduction in the amount of thrombin generated in the culprit artery. In addition, enoxaparin offers a more stable level of anticoagulation, which eliminates the need for aPTT monitoring and allows for subcutaneous dosing rather than intravenous infusions, making it a more convenient strategy that may also reduce the cost of care.28
Prior meta-analyses have examined individually the efficacy and safety of enoxaparin when compared with UFH in STEMI7,10 and in NSTEACS,8,9 but the present analysis is the first to (i) evaluate a net clinical endpoint; (ii) assess the entire ACS spectrum; and (iii) include the ExTRACT-TIMI 25 trial, thereby increasing the total population of patients available for analysis by 72% (from n = 28 609 to n = 49 088). Given the consistent efficacy of reduced death or MI with enoxaparin, use of a net clinical endpoint incorporating bleeding can be justified to characterize the clinical profile of the therapy; had no efficacy been observed, use of a net clinical endpoint would not be warranted.
There was a consistent efficacy benefit of reduced death or MI with enoxaparin in both STEMI and NSTEACS trials. Net clinical benefit was evident among the STEMI population and was neutral among the NSTEACS population. Several factors may have contributed to these findings. In the NSTEACS trials, the major bleeding rate was higher overall than in the STEMI trials (5.8% in NSTEACS trials vs. 2.2% in STEMI trials), and therefore contributed more events to the net clinical endpoint. Although the absolute major bleeding event rate was higher in the NSTEACS trials than the STEMI trials, there was no significant relative increase in major bleeding between enoxaparin and UFH in the NSTEACS trials (OR 1.13, P = NS) in contrast to the STEMI trials. Conversely, the relative reduction in death or MI was somewhat greater in the STEMI trials (OR 0.78) than in the NSTEACS trials (OR 0.90). These differences in both efficacy and bleeding may be explained in part by the greater use of revascularization procedures in the NSTEACS trials, differences in concomitant therapies such as thienopyridines and pre-randomization antithrombin use, and different durations of enoxaparin therapy in the trials. In addition, NSTEACS patients tend to be a more heterogeneous population than STEMI patients, with therapies often showing different degrees of efficacy and safety in unstable angina patients and in NSTEMI patients. Additional trials evaluating lower doses of enoxaparin in certain populations at high-risk for bleeding, including the elderly and those with reduced renal function, should be considered to improve the safety profile of enoxaparin. Another consideration to reduce bleeding rates in the ACS population undergoing PCI is increased use of radial artery catheterization, where studies have shown reductions in major bleeding when compared with femoral access.29,30
Limitations
As with all meta-analyses, differences in trial designs should be considered when interpreting the overall results. Timing and definitions of endpoints for MI and major bleeding varied somewhat across trials, as did the use of adjudication of events. However, point estimates for MI consistently fell to the left of the line of unity (favouring enoxaparin). Point estimates for major bleeding are more varied, but the majority of trials fell to the right of the line of unity (favouring UFH). Duration and dose of study drugs also differed between trials, as did the use of concomitant therapies, revascularization, and risk profiles. Given the heterogeneity in the analysis of the entire ACS spectrum for the net clinical endpoint, the results of the STEMI and the NSTEACS cohorts are reported individually as well as pooled. Patient-level data were not available for all 12 trials, so additional subgroup analyses could not be performed. Long-term data beyond 30 days were not available in most trials so late mortality was not evaluated.
Clinical implications
Compared with UFH, adjunctive antithrombin therapy with enoxaparin was associated with significantly superior efficacy benefit across the ACS spectrum among more than 49 000 patients. Among STEMI patients, 21 death or MI events were prevented for every 1000 patients treated with enoxaparin, at the cost of an increase of four non-fatal major bleeds; among NSTEACS patients, nine death or MI events were prevented for every 1000 patients treated with enoxaparin, at the cost of an increase of eight non-fatal major bleeds. These data provide evidence in favour of enoxaparin adjunctive antithrombin regimen to support ACS therapy over the standard strategy of UFH, especially among STEMI patients.
Acknowledgements
We thank Leah S. Teoh from Sanofi-Aventis, Karen Pieper from Duke Clinical Research Institute, Kris Bogaerts and Ann Belmans from Biostatistical Centre K.U. Leuven, and Mary Tan from the Canadian Heart Research Centre for assistance in providing missing data.
Conflict of interest: The TIMI study group has received research grants from Sanofi-Aventis. Gibson, Morrow, Van de Werf, Cohen, Antman, and Braunwald have received research grants and speaking honoraria from Sanofi-Aventis. Menown has received speaker honoraria from Sanofi-Aventis. Mahaffey has received research grant support and speaker honoraria from Sanofi-Aventis, Schering-Plough, and The Medicines Company. Goodman has received research grant support and speaker honoraria from Sanofi-Aventis, Glaxo Smith Kline, and The Medicines Company.
Comments