OUP user menu

Impact of anticoagulation levels on outcomes in patients undergoing elective percutaneous coronary intervention: insights from the STEEPLE trial

Gilles Montalescot, Marc Cohen, Genevieve Salette, Walter J. Desmet, Carlos Macaya, Philip E.G. Aylward, Ph. Gabriel Steg, Harvey D. White, Richard Gallo, Steven R. Steinhubl
DOI: http://dx.doi.org/10.1093/eurheartj/ehn008 462-471 First published online: 14 February 2008

Abstract

Aims To determine the relationship between anticoagulation levels during percutaneous coronary intervention, and ischaemic events and bleeding.

Methods and results A sub-analysis from the STEEPLE trial was conducted. Pre-defined target anticoagulation levels were achieved in 86% of patients receiving enoxaparin, compared with 20% receiving unfractionated heparin (UFH) (P < 0.001). A significant relationship was observed between anti-Xa levels > 0.9 IU/mL and covariate-adjusted rate of non-coronary artery bypass graft-related major and minor bleeding [odds ratio (OR) 1.6, 95% CI 1.0–2.5 for each unit of anti-Xa; P = 0.03]; anti-Xa levels and covariate-adjusted incidence of death, myocardial infarction, or revascularization showed no significance (P = 0.47). Major bleeding increased significantly with an activated clotting time (ACT) > 325 s (OR 1.6, 95% CI 1.1–2.2 per 100 s; P = 0.04). A significant relationship with increasing ischaemic events was observed when ACT was < 325 s (OR 0.7, 95% CI 0.2–0.8 per 100 s; P = 0.006) indicating a narrow therapeutic window.

Conclusion Target anticoagulation levels were achieved more readily in patients receiving enoxaparin. An anti-Xa level of up to 0.9 IU/mL has a good safety and efficacy profile; poor achievement of target ACT with UFH makes assessing the optimal range difficult.

Keywords
  • Anticoagulation
  • Enoxaparin
  • Percutaneous coronary intervention
  • Unfractionated heparin

Introduction

Anticoagulant therapy administered during percutaneous coronary intervention (PCI) is the standard of care in order to prevent thromboembolic complications within the catheters or coronary arteries. Currently, several different anticoagulants are being used: unfractionated heparin (UFH);1 low-molecular-weight heparins (LMWHs) such as enoxaparin;221 and direct antithrombins, such as bivalirudin. Current guidelines recommend the use of activated clotting time (ACT) in order to guide the dose of UFH in PCI.22 However, despite these recommendations, uncertainty remains with regard to the correlation between the effects of UFH, as measured by ACT, and peri-procedural ischaemic and haemorrhagic complications.23,24 Additional complexity is added by the relatively weak correlation between the UFH dose administered and ACT, the use of different ACT targets according to the use of glycoprotein (GP) inhibitors IIb/IIIa, and by the variability in instruments used to measure ACT.

While there is a good correlation between anti-Xa levels after subcutaneous enoxaparin doses, and clinical outcomes in acute coronary syndrome patients,25 limited data are available on optimal anti-Xa levels when intravenous enoxaparin is used in PCI. Most data come from small, non-comparative trials.2,1618,20,2629

The recently completed SafeTy and Efficacy of Enoxaparin in PCI patients, an internationaL randomized Evaluation (STEEPLE) trial, a prospective, open-label randomized trial in 3528 patients showed that intravenous bolus enoxaparin was associated with a significant reduction in major bleeding, and more predictable anticoagulation levels compared with UFH in patients undergoing elective PCI.30 ACTs were prospectively measured with a standardized Hemochron device (ITC, Edison, NJ, USA), and anti-Xa levels were measured by a central core laboratory for both enoxaparin treatment arms, providing a wide range of anticoagulation levels with the two different doses. Thus, the STEEPLE trial provides an opportunity to assess the correlation between a broad range of anticoagulation levels, and ischaemic and bleeding outcomes with two different anticoagulants used in PCI with modern adjunctive pharmacology.

Methods

The design of the STEEPLE trial has been described previously.30 In brief, patients were enrolled if they were aged 18 years or older, scheduled to have elective PCI with a femoral approach, and did not meet any of the exclusion criteria, including recent thrombolysis, planned staged procedure, increased bleeding risk, treatment with a parenteral antithrombotic agent prior to PCI, or known hypersensitivity to study drugs.

Patients were randomized to intravenous enoxaparin (0.5 or 0.75 mg/kg) or ACT-adjusted UFH, stratified by GP IIb/IIIa inhibitor use. All patients received aspirin (75–500 mg/day at the discretion of the physician) and thienopyridines according to local practice.

Patients assigned to either enoxaparin group received a single intravenous bolus of enoxaparin immediately prior to PCI. When procedures were prolonged by more than 2 h, an additional bolus of enoxaparin (half of the original dose) was recommended.

Patients randomized to UFH who were not receiving concurrent GP IIb/IIIa inhibitors were given an initial intravenous bolus of 70–100 IU/kg to achieve a target ACT of 300–350 s. A lower initial bolus of UFH of 50–70 IU/kg was given to patients who received concurrent GP IIb/IIIa inhibitors in order to achieve a target ACT of 200–300 s. Additional boluses of UFH were given before the start of PCI and re-administered during the procedure in patients who did not reach the lower limit of the ACT target. All centres measured ACT with a standardized Hemochron device.

Outcome analyses

The occurrence of non-coronary artery bypass graft (non-CABG)-related major and minor bleeding at 48 h after the index PCI was evaluated according to pre-specified definitions (Table 1). Outcomes were adjudicated by a blinded independent event committee. The composite ischaemic endpoint was: all-cause mortality; non-fatal myocardial infarction (MI), defined as new significant Q wave in ≥ 2 leads or elevation of creatine kinase MB (or total creatine kinase) ≥ 3 times the upper limit of normal (if only total creatine kinase was measured, troponin I/T had to be positive); or urgent target-vessel revascularization at 30 days.

View this table:
Table 1

Non-coronary artery bypass graft-related major bleeding and minor bleeding definitions

Bleeding was classified as major if at least one of the following applied
 Resulted in death
 Retroperitoneal, intracranial, or intraocular
 Resulted in haemodynamic compromise requiring specific treatment
 Required intervention (surgical or endoscopic)
 Clinically overt with transfusion of at least one unit of PRBCs or whole blood
 Clinically overt with a decrease in haemoglobin of ≥3 g/dL (or when haemoglobin was not available, a decrease in haematocrit of ≥10%)
Bleeding was classified as minor if it did not meet the criteria for major bleeding and if at least one of the following applied
 Gross haematuria not associated with trauma
 Epistaxis that was prolonged, repeated, or required plugging or intervention
 Gastrointestinal haemorrhage
 Haemoptysis
 Sub-conjunctival haemorrhage
 Haematoma >5 cm or led to prolonged or new hospitalization
 Clinically overt with a decrease in haemoglobin ≥ 2 to ≤ 3 g/dL
 Required protamine sulfate administration for uncontrolled bleeding
  • PRBCs, packed red blood cells.

In all patients receiving enoxaparin, anti-Xa levels were measured at the beginning of PCI, 10 min after administering the enoxaparin bolus of either 0.5 or 0.75 mg/kg, and at the end of PCI. Having two different doses of enoxaparin provided a wider range of anti-Xa values for correlation with ischaemic and bleeding events. Target anti-Xa levels were pre-defined in the STEEPLE study protocol as 0.5–1.8 IU/mL. The lower limit was based on prior studies of target anti-Xa levels in patients with non-ST-elevation acute coronary syndrome,25,31 and in patients undergoing PCI.2,10,18,32 The upper value corresponds to the 75th percentile of peak anti-Xa values in patients treated with 1.25 mg/kg every 12 h who did not experience major haemorrhage in the Thrombolysis In Myocardial Infarction (TIMI) 11A study.33 In addition, the anti-Xa levels were categorized into: < 0.5, 0.5–1.2, 1.2–1.8, and > 1.8 IU/mL. The highest anti-Xa at any time was used for its relationship to bleeding, and the lowest anti-Xa at any time for its relationship to ischaemic events for each patient.

Similarly, ACT levels were measured at the beginning and end of the procedure in all patients receiving UFH, and target ACT was pre-defined as 200–300 s if a GP IIb/IIIa inhibitor was used, or 300–350 s in the absence of GP IIb/IIIa inhibitors.4 ACT was also categorized into the following levels: < 200, 200–300, 300–350, and > 350 s. The highest ACT at any time was used for its relationship to bleeding and the lowest for its relationship to ischaemic events for each patient.

Statistical analyses

The relationship between anticoagulation levels, treated as categorical variables, and bleeding and ischaemic endpoints was assessed by univariate analysis using Cochran–Mantel–Haenszel χ2-test for trend. The correlation between anticoagulation levels, treated as continuous variables, and bleeding and ischaemic endpoints was assessed using multivariable logistic regression analyses, adjusted for baseline characteristics; models were built using a backward approach (significance level for staying variables is equal to 0.05). Final models were checked for accuracy using Hosmer–Lemeshow tests and c-statistics. For bleeding, the factors considered for adjustment were: age ≥75 years, sex, obesity, diabetes, hypertension, current smoker, hypercholesterolemia, renal insufficiency (defined as a creatinine clearance ≤ 60 mL/min using the Cockcroft–Gault formula34); peripheral arterial disease; family history of coronary artery disease; unstable angina or MI within the previous 7 days (by report of the investigator); haemoglobin at entry (≤10 g/dL for women, or 11 g/dL for men); platelets at entry (<80,000 g/mm3); number of diseased arteries; sheath size; use of GP IIb/IIIa inhibitor; use of other antiplatelet agents; country; use of enoxaparin or other LMWH, or UFH or direct thrombin inhibitor between 7 days and 24 h prior to enrolment; and use of warfarin or other vitamin K antagonists within the previous week. For the ischaemic endpoint, the following variables were also considered (in addition to those listed above): TIMI flow grade (pre-procedure), use of β-blockers, use of calcium channel blockers, use of angiotensin-converting enzyme inhibitors, use of angiotensin receptor blockers, and use of statins within the previous week. Natural spline cubic transformation of anti-Xa and ACT was applied in all multivariable models, using the method of Stone and Koo.35

Two-tailed tests of significance are reported. A P-value of < 0.05 was considered statistically significant. Analyses were performed using SAS statistical software version 8.2 (SAS Institute Inc., Cary, NC, USA).

Results

The overall results of the STEEPLE trial showed that enoxaparin significantly reduced the incidence of non-CABG-related bleeding compared with UFH, while the incidence of the composite ischaemic endpoint was similar between enoxaparin and UFH.30

Baseline and procedural characteristics were comparable for patients receiving intravenous enoxaparin (n = 2298), or UFH (n = 1230) (Table 2). At least one additional bolus was given to 16.5% of patients receiving UFH, because of a low ACT. Of patients receiving enoxaparin, 0.4% received at least one additional bolus of enoxaparin for prolonged procedures (>2 h).

View this table:
Table 2

Baseline and procedural characteristics of patients enrolled in the STEEPLE trial

Enoxaparin (n = 2298)aUFH (n = 1230)a
Baseline characteristics
 Mean age±SD, years63.5 ± 10.463.5 ± 10.2
 Age ≥75 years (%)360 (15.7)184 (15.0)
 Male (%)1733 (75.4)910 (74.0)
 Mean weight±SD, kg84.1 ± 16.883.3 ± 16.0
 Creatinine clearance (%)
  >30 to ≤60 mL/min403 (18.0)217 (18.0)
  ≤30 mL/min20 (0.9)19 (1.6)
 Diabetes (%)682 (29.7)380 (30.9)
 Prior MI (%)
  >48 h to ≤7 days41 (1.8)24 (2.0)
  ≤48 h13 (0.6)3 (0.2)
 Prior unstable angina (%)
  >48 h to ≤7 days158 (6.9)82 (6.7)
  ≤48 h127 (5.5)60 (4.9)
 Prior PCI (%)819 (35.7)479 (38.9)
 Prior CABG (%)310 (13.5)186 (15.1)
Concomitant medications on day of PCI (%)
 GP IIb/IIIa inhibitor932 (40.6)491 (39.9)
 Thienopyridine2165 (94.3)1170 (95.1)
 Aspirin2133 (93.1)1159 (94.4)
PCI variables
 Stent procedure (%)2100 (93.8)1133 (94.0)
 Drug-eluting stent (%)1277 (57.0)675 (56.0)
 Target vessel (%)
  LAD1050 (46.9)546 (45.3)
  LCX691 (30.9)375 (31.1)
  RCA832 (37.1)484 (40.2)
  LMA30 (1.3)12 (1.0)
  Saphenous or artery graft92 (4.1)43 (3.6)
 Multi-vessel (≥2 vessels) intervention (%)345 (15.4)207 (17.2)
 Closure device used (%)878 (39.3)478 (39.7)
 Median time from end of PCI to sheath removal, h:min (IQR)02:22 (00:01, 04:38)02:24 (00:01, 04:14)
  • CABG, coronary artery bypass graft; GP, glycoprotein; IQR, interquartile range; LAD, left anterior descending coronary artery; LCX, left circumflex coronary artery; LMA, left main coronary artery; MI, myocardial infarction; PCI, percutaneous coronary intervention; RCA, right coronary artery; SD, standard deviation; UFH, unfractionated heparin.

  • aBecause of missing values, the % cannot be calculated on 2298 and 1230 patients.

Pre-specified target anticoagulation levels were achieved at both the beginning and end of the procedure in significantly more patients receiving enoxaparin compared with patients receiving UFH (85.8 vs. 19.7%, P < 0.001). For patients receiving enoxaparin, the median anti-Xa at the start and end of PCI was 0.94 (IQR: 0.78–1.14) and 0.82 (IQR: 0.65–0.97) IU/mL, respectively. For patients receiving UFH only, the median ACT at the start and end of PCI was 336 s (IQR: 302–378) and 292 s (IQR: 258–331), and for patients receiving UFH with GP IIb/IIIa inhibitors 300 s (IQR: 261–354) and 255 s (IQR: 218–297), respectively.

Anti-Xa levels and outcomes in patients receiving enoxaparin

Target anticoagulation levels were achieved by 85.8% of patients receiving enoxaparin. Of those who did not meet target levels, 11.7% patients had at least one anti-Xa value below 0.5 IU/mL and 2.6% of patients had at least one value above 1.8 IU/mL.

Univariate analysis of anti-Xa levels when ranked in four ranges of values is shown in Table 3. Multivariable logistic regression analysis confirmed that anti-Xa levels were an independent predictor of non-CABG-related major and minor bleeding: there was no correlation between anti-Xa levels and non-CABG-related major and minor bleeding up to anti-Xa levels of 0.9 IU/mL (P = 0.93), but with anti-Xa levels above 0.9 IU/mL, a significant correlation was found [odds ratio (OR) 1.6, 95% confidence interval (CI) 1.0–2.5 for each unit of anti-Xa; P = 0.03 (Figure 1A); Hosmer–Lemeshow test, P = 0.3156, c-statistic 0.698]. The same relationships between anti-Xa level and non-CABG-related major and minor bleeding were observed on the curves obtained from subjects without (P = 0.047 [Figure 1B]; Hosmer–Lemeshow test, P = 0.1045, c-statistic 0.691), or with (P = 0.30; Figure 1C) concomitant use of GP IIb/IIIa inhibitors. Other independent predictors for higher non-CABG-related major and minor bleeding rates were age ≥ 75 years, female gender, use of sheath size ≥7F, and use of a GP IIb/IIIa inhibitor.

Figure 1

Covariate-adjusted proportion of non-CABG-related major and minor bleeding up to 48 h (with 95% CI ---) in: (A) patients receiving enoxaparin; (B) patients receiving enoxaparin without GP IIb/IIIa inhibitors; and (C) patients receiving enoxaparin with GP IIb/IIIa inhibitors. CABG, coronary artery bypass graft; GP, glycoprotein.

View this table:
Table 3

Relationship between anti-Xa levels and bleeding up to 48 h, and ischaemic outcomes up to 30 days, in patients receiving enoxaparin

Anti-Xa levels (IU/mL)P-value
<0.50.5–<1.21.2–1.8>1.8
Total non-CABG-related bleeding, n/N (%)3/49 (6.1)95/1591 (6.0)32/422 (7.6)7/55 (12.7)0.048
Non-CABG-related major bleeding, n/N (%)1/49 (2.0)18/1591 (1.1)8/422 (1.9)1/55 (1.8)0.34
Composite ischaemic endpoint of 30-day death, non-fatal MI, and UTVR, n/N (%)22/249 (8.8)105/1757 (6.0)8/104 (7.7)1/7 (14.3)0.45
  • CABG, coronary artery bypass graft; MI, myocardial infarction; UTVR, urgent target-vessel revascularization.

No significant correlation was found between anti-Xa levels and the composite ischaemic endpoint of all-cause mortality, non-fatal MI, and urgent target vessel revascularization in the overall patient group on enoxaparin (P = 0.47; Figure 2A) or in the subgroup not receiving concomitant GP IIb/IIIa inhibitors (P = 0.38; Figure 2B). A low haemoglobin level at study entry and the use of a GP IIb/IIIa inhibitor were independent predictors of a higher ischaemic endpoint incidence. In patients with concomitant GP IIb/IIIa inhibitor use, anti-Xa values above 0.9 IU/mL (but not below 0.9 IU/mL; P = 0.94) were significantly correlated to the ischaemic endpoint (P = 0.004 [Figure 2C]; Hosmer–Lemeshow test, P = 0.4851, c-statistic 0.716), although the small number of events and patients (15 events among 262 patients) did not allow the calculation of OR and 95% CI.

Figure 2

Covariate-adjusted proportion of death, non-fatal MI or UTVR up to 30 days (with 95% CI ---) in: (A) patients receiving enoxaparin; (B) patients receiving enoxaparin without GP IIb/IIIa inhibitors; and (C) patients receiving enoxaparin with GP IIb/IIIa inhibitors. GP, glycoprotein; MI, myocardial infarction; UTVR, urgent target-vessel revascularization.

Activated clotting time levels and outcomes in patients receiving unfractionated heparin

Only 19.7% of patients receiving UFH achieved the pre-specified target anticoagulation level at both the beginning and end of PCI; 39.8% had at least one ACT value below the target range, whereas 57.9% had at least one value above the target range.

Univariate analysis with anticoagulation levels ranked in four ranges of values is reported in Table 4. The multivariable logistic regression analysis showed that ACT levels were an independent predictor of non-CABG-related major and minor bleeding (OR 1.3, 95% CI 1.0–1.7 per 100 s; P = 0.049; Hosmer–Lemeshow test, P = 0.7998, c-statistic 0.636). Other independent predictors of more non-CABG-related major and minor bleeding were female gender and use of a GP IIb/IIIa inhibitor (Figure 3A). Although non-significant, similar trends were observed in subgroups of patients with concomitant GP IIb/IIIa inhibitors (OR 1.3, 95% CI 1.0–1.7 per 100 s; P = 0.09), or without concomitant GP IIb/IIIa inhibitors (OR 1.3, 95% CI 0.9–2.0 per 100 s; P = 0.20). An ACT level higher than 325 s was an independent predictor of non-CABG-related major bleeding (OR 1.6, 95% CI 1.1–2.2 per 100 s; P = 0.04 [Figure 3B]; Hosmer–Lemeshow test, P = 0.5664, c-statistic 0.716).

Figure 3

Covariate-adjusted proportion (with 95% CI ---) of: (A) non-CABG-related major and minor bleeding up to 48 h; (B) non-CABG-related major bleeding up to 48 h; and (C) death, non-fatal MI, or UTVR up to 30 days, in all patients receiving UFH. CABG, coronary artery bypass graft; MI, myocardial infarction; UFH, unfractionated heparin; UTVR, urgent target-vessel revascularization.

View this table:
Table 4

Relationship between activated clotting time levels and bleeding up to 48 h, and ischaemic outcomes up to 30 days, in all patients receiving unfractionated heparin

ACT levels (s)P-value
<200200–<300300–350>350
Total non-CABG-related bleeding, n/N (%)24/268 (9.0)16/299 (5.4)58/567 (10.3)0.29
Non-CABG-related major bleeding, n/N (%)5/268 (1.9)8/299 (2.7)20/567 (3.5)0.17
Composite ischaemic endpoint of 30-day death, non-fatal MI, and UTVR, n/N (%)12/114 (10.5)40/682 (5.9)5/195 (2.6)8/143 (5.6)0.06
  • ACT, activated clotting time; CABG, coronary artery bypass graft; MI, myocardial infarction; UTVR, urgent target-vessel revascularization.

Multivariable logistic regression analysis showed that up to 325 s, ACT was an independent predictor of the composite ischaemic endpoint (OR 0.7, 95% CI 0.2–0.8 per 100 s; P = 0.006; Hosmer–Lemeshow test, P = 0.2486, c-statistic 0.638); above this value, the association became non-significant (P = 0.86; Figure 3C). Similar but non-significant trends were seen in the two subgroups of patients receiving UFH with concomitant GP IIb/IIIa inhibitors (ACT ≤ 325 s, P = 0.17; ACT > 325 s, P = 0.58) or without (ACT ≤ 325 s, P = 0.14; ACT > 325 s, P = 0.77). ACT was not a predictor of the composite ischaemic endpoint in either of these subgroups.

Procedural complications and anticoagulation levels

The rate of procedural complications was low in this patient population. There was no significant difference in the number of patients who experienced catheter thrombosis (1/1041 and 2/1202 patients receiving 0.5 mg/kg and 0.75 mg/kg enoxaparin, respectively vs. 4/1223 UFH; P = 0.64), or acute vessel closure or stent occlusion (2/1041 and 5/1202 patients receiving 0.5 mg/kg and 0.75 mg/kg enoxaparin, respectively, vs. 1/1223 UFH; P = 0.21 for comparison between the three groups) between treatment groups. All patients who experienced acute vessel closure or stent occlusion, while receiving either dose of enoxaparin (0.5 or 0.75 mg/kg) had anti-Xa levels within the specified anticoagulation range. Of the two patients in the enoxaparin 0.75 mg/kg group who experienced a catheter thrombosis, one had anti-Xa levels above the target range. None of the five patients receiving UFH who experienced a procedural complication were within the pre-defined ACT range; four were above and one was below the ACT range.

Discussion

Results of this analysis of the STEEPLE trial showed that target ranges for anticoagulation levels were achieved in 86% of patients receiving enoxaparin compared with 20% receiving UFH; the majority of patients receiving UFH (58%) had ACT levels above the ACT target range.

In the current practice of elective PCI with the use of dual antiplatelet therapy (aspirin and clopidogrel), an anti-Xa level of approximately 0.9 IU/mL in patients receiving enoxaparin was associated with an optimal risk:benefit ratio as shown by a low incidence of non-CABG-related major and minor bleeding, and low incidence of ischaemic events. Similarly, an ACT level of 325 s in patients receiving UFH (with and without concomitant GP IIb/IIIa inhibitor use) was associated with an optimal risk:benefit ratio as shown by a significant increase in non-CABG-related major bleeding above 325 s, and a significant increase in the incidence of the composite ischaemic endpoint with decreasing ACT values below 325 s. The lowest incidence of major bleeding and ischaemic outcomes was observed with UFH and ACT levels between 300 and 350 s. This narrow therapeutic window for patients receiving UFH, combined with the poor achievement of target anticoagulation levels, highlights the limitations of the standard UFH anticoagulation in PCI.

The therapeutic anticoagulation range is less well defined for enoxaparin compared with UFH. Therefore, the target anti-Xa levels were defined according to the lower range of effective levels for the treatment of venous disorders and the upper threshold level considered to be associated with an increased risk of bleeding.31,33 In addition, anti-Xa levels < 0.5 IU/mL are associated with an increase in 30-day mortality in ACS patients compared with patients with anti-Xa levels 0.5–1.2 IU/mL (P = 0.004).30

The use of a wide target anticoagulation range in patients receiving enoxaparin allowed us to perform a continuous evaluation of the impact of anti-Xa levels on clinical outcomes. This provides a greater degree of accuracy in assessing the relationship between anti-Xa levels and clinical outcomes in PCI patients, compared with previous studies which analysed only categorized anti-Xa levels.30

The results from our study are consistent with the findings from a previous single-centre study in which 95% of patients undergoing elective PCI achieved a peak anti-Xa level in the pre-defined target range of 0.5–1.5 IU/mL following treatment with single intravenous bolus of enoxaparin (0.5 mg/kg).2 Similar results were also found in unselected acute coronary syndrome patients receiving the standard subcutaneous enoxaparin regimen, with 93 and 94% of patients achieving target anti-Xa levels in two different studies.25,36

The 0.5 and 0.75 mg/kg intravenous doses appear to produce levels of anticoagulation similar to those obtained with subcutaneous enoxaparin (1 mg/kg twice daily). With intravenous dosing, dose-dependent peak anti-Xa levels are rapidly achieved within 5 min after administration, falling to below 0.5 IU/mL within 2–7 h after a single intravenous bolus of enoxaparin, depending on the dose (1, 0.75, or 0.5 mg/kg).18,20,26,32 These data were highly consistent across studies and demonstrated the predictable nature of anticoagulation with enoxaparin.

In our study, anticoagulation levels were maintained within the target range throughout the PCI procedure for the majority of patients treated with enoxaparin. Slightly different results were reported recently by Zalc et al.37 where anti-Xa levels < 0.5 IU/mL at the start of PCI were found in 12 of 54 elective PCI patients receiving 0.5 mg/kg enoxaparin. However, the results from the STEEPLE trial with more than 2200 patients receiving enoxaparin confirmed the good predictability of anticoagulation, and found anti-Xa levels > 0.5 IU/mL at the start of procedure in more than 90% of patients with 0.5 mg/kg enoxaparin.

Our study demonstrated that the use of a GP IIb/IIIa inhibitor was independently associated with higher bleeding rates in the enoxaparin treatment group, and there was a significant increase in bleeding complications with increasing anti-Xa levels beyond 0.9 IU/mL. This trend towards increased bleeding was the same with or without GP IIb/IIIa inhibitors (Figure 1), confirming the concept that the same enoxaparin dose is appropriate with or without GP IIb/IIIa inhibitors use. The evidence from this, and other studies, has also shown that the combination of enoxaparin (up to 0.75 mg/kg intravenously) with a GP IIb/IIIa inhibitor was not associated with an excess in the rate of bleeding complications compared with low-dose UFH and a GP IIb/IIIa inhibitor.17,29,30

Overall, we did not observe any significant correlation between anti-Xa levels and the composite ischaemic endpoint of all-cause mortality, non-fatal MI and UTVR. However, there was a significant correlation between the composite ischaemic endpoint and anti-Xa levels >0.9 IU/mL in patients receiving GP IIb/IIIa inhibitors. Interestingly, recent data from the GRACE registry suggests that LMWH is associated with lower mortality rates and increased bleeding risk compared with UFH only in cases where GP IIb/IIIa inhibitors are not used.38 This suggests that interactions may occur between these inhibitors and LMWH in non-STEMI ACS patients who undergo PCI. However, it should be noted that the use of GP IIb/IIIa inhibitors in the STEEPLE trial was left to the discretion of the investigators. Therefore, this result may have been influenced by different baseline characteristics for patients who received GP IIb/IIIa inhibitors compared with those who did not.39

Only a few studies have been conducted with sufficient statistical power to assess correlations between target anti-Xa levels and ischaemic outcomes. A recent evaluation of 30-day survival in 803 consecutive patients with unstable angina/non-ST-segment elevation MI undergoing PCI found that the mean anti-Xa activity was significantly lower in patients who died than in those who survived (0.74 ± 0.05 vs. 0.92 ± 0.01 IU/mL; P = 0.0002). The 30-day mortality rate was significantly associated with suboptimal anticoagulation (anti-Xa levels < 0.5 IU/mL) and resulted in a three-fold increase in mortality compared with the patients with anti-Xa levels in the target range of 0.5–1.2 IU/mL (P = 0.004).25 This threshold for efficacy could not be evaluated in the present analysis of elective PCI patients as few patients were below 0.5 IU/mL, and most of those that were, fell between 0.4 and 0.5 IU/mL on only one of the two measurements.

The importance of ACT measurement and target levels is clearly evident in PCI management guidelines. The ACT target ranges used in this sub-analysis accurately reflect the therapeutic ranges recommended in current guidelines for the use of UFH in patients undergoing PCI.4,22 Because of marked variability in UFH bioavailability, ACT-guided dosing is advocated,4 especially for prolonged procedures when additional boluses may be required. Nevertheless, the evidence from this trial, and other studies, indicates that desirable anticoagulation with UFH, even with ACT-guided dosing, is difficult and target levels were only achieved in 20% of our patients undergoing elective PCI.

In the present analysis, an ACT of 325 s may offer an optimal risk:benefit ratio, although lower ACTs may be targeted with the concomitant use of GP IIb/IIIa inhibitors. Although the total number of ischaemic events did not significantly differ between patients receiving enoxaparin compared with UFH [enoxaparin: 150/2298 (6.5%) vs. UFH 72/1230 (5.8%), P = 0.43], our analysis shows that ACTs < 325 s are associated with more ischaemic events, particularly when GP IIb/IIIa inhibitors are not used. The multivariable logistic regression analysis demonstrated that above 325 s, there is no longer an association between ACT and the risk of ischaemic events (P = 0.86), but ACT levels > 325 s correlate with bleeding complications (with similar trends with or without GP IIb/IIIa inhibitors), suggesting no advantage of having ACTs higher than 325 s in the modern era of PCI, where clopidogrel is often used. Within STEEPLE, drug-eluting stents were used in 57% of patients, GP IIb/IIIa inhibitors in 40%, clopidogrel pre-treatment in 46%, and closure devices in 39%.

Our ACT data confirm and update previous information gained from meta-analyses24,40 and randomized studies.23 Brener et al.24 showed that higher weight-adjusted doses of UFH did not improve protection against ischaemic events. Similarly, our multivariable logistic regression analysis showed a flat curve across a wide range of values, suggesting little impact of ACT on ischaemic outcome. This study is underpowered to draw firm conclusions regarding the comparative incidence of ischaemic events in patients receiving enoxaparin and UFH. However, our data suggest that the total number of ischaemic events did not significantly differ between patients receiving enoxaparin compared with UFH. This statement is supported by the results of a recent meta-analysis by Dumaine et al.,41 which was sufficiently powered to make this comparison.

Our prospectively defined analysis on more than 3500 patients and evidence from other recent studies show that increasing ACTs do not confer a benefit and may even be detrimental. This underlines the uncertainty of targeting to specific ranges of ACT and questions the validity of the current guidelines on UFH management during modern PCI.

Although randomization in the STEEPLE trial was stratified for use of GP IIb/IIIa inhibitors, dosing of GP IIb/IIIa inhibitors was left to the individual physicians and some overdosing with GP IIb/IIIa inhibitors may have occurred in patients with renal impairment. While renal impairment is not a problem for single intravenous injections of enoxaparin, concerns have been raised regarding increased risk of bleeding in PCI patients with renal impairment who receive GPIIb/IIIa inhibitors.42 Renal insufficiency was not an exclusion criterion for the STEEPLE trial; therefore, changes in the bleeding risk associated with use of GP IIb/IIIa inhibitors (40% inhibitor use reported in this study) may have been influenced by patients with renal impairment.

Other limitations of this analysis of the STEEPLE trial may have been the omission of American College of Cardiology lesion classification from the multivariable analysis for ischaemic events.

Conclusions

The results of this analysis, along with the main findings from the STEEPLE trial, indicate that target anticoagulation levels were achieved more readily in patients receiving enoxaparin. An anti-Xa level of up to 0.9 IU/mL has a good safety and efficacy profile; poor achievement of target ACT with UFH makes assessing the optimal range difficult.

Funding

The STEEPLE trial was funded by sanofi-aventis.

Acknowledgements

We would like to thank Sandrine Brette for statistical support in the preparation of this manuscript.

Conflict of interest: This study was funded by sanofi-aventis.

Dr Montalescot reports receiving: grant support, consulting fees, and lecture fees from sanofi-aventis, Eli Lilly and Bristol-Myers Squibb; consulting fees and lecture fees from Merck Sharp & Dohme; consulting fees from Procter & Gamble, AstraZeneca, and Schering-Plough; and lecture fees from GlaxoSmithKline and Nycomed. Dr Cohen reports receiving: grant support from Aventis Pharmaceuticals; consulting fees from sanofi-aventis and AstraZeneca; and lecture fees from sanofi-aventis, Merck, and Schering. Dr Salette is an employee of sanofi-aventis. Dr Desmet reports having received consulting fees from Nycomed and The Medicines Company. Dr Macaya has no conflicts of interest to declare. Dr Aylward reports receiving grant support from sanofi-aventis, Procter & Gamble, Alexion, The Medicines Company, Schering-Plough and Eli Lilly, as well as consulting fees and lecture fees from sanofi-aventis and Bristol-Myers Squibb. Dr Steg reports receiving: grant support from sanofi-aventis; consulting fees from sanofi-aventis, Takeda, AstraZeneca, Bristol-Myers Squibb, Merck Sharp & Dohme, GlaxoSmithKline, Pfizer and Servier; and speakers bureau from Boehringer Ingelheim, Bristol-Myers Squibb, GlaxoSmithKline, Merck Sharp & Dohme, Novartis, Nycomed, sanofi-aventis, Sankyo, Servier and ZLB Behring. Dr White reports receiving: grant support from Alexion, sanofi-aventis, Eli Lilly, Merck Sharp & Dohme, The Medicines Company, Neuren Pharmaceuticals, GlaxoSmithKline, Pfizer, Roche, Fournier Laboratories, Johnson & Johnson, Procter & Gamble, Schering-Plough and Janssen-Cilag; consulting fees from Medicure, The Medicines Company, Neuren Pharmaceuticals and GlaxoSmithKline; and honorarium from sanofi-aventis and The Medicines Company. Dr Gallo reports having received grant support, consulting fees and lecture fees from sanofi-aventis, lecture fees from Abbott Interventional, Oryx Pharmaceuticals and Biovail Pharmaceuticals, and consulting fees from Biovail Pharmaceuticals. Dr Steinhubl reports receiving consulting fees from sanofi-aventis, The Medicines Company, Daiichi Sankyo, Eli Lilly, Cardax Pharmaceuticals and AstraZeneca.

Footnotes

  • This paper was guest edited by Freek W.A. Verheugt, Department of Cardiology, Heartcenter, University Medical Center Nijmegen, PO Box 9101, Nijmegen 6500 HB, The Netherlands.

References

View Abstract