Author + information
- David J. Moliterno, MD, FACC*,* (, )
- James B. Hermiller, MD, FACC†,
- Dean J. Kereiakes, MD, FACC‡,
- Eric Yow, MS§,
- Robert J. Applegate, MD, FACC∥,
- Gregory A. Braden, MD, FACC¶,
- Eric J. Dippel, MD, FACC#,
- Mark I. Furman, MD, FACC**,
- Cindy L. Grines, MD, FACC††,
- Neal S. Kleiman, MD, FACC‡‡,
- Glenn N. Levine, MD, FACC§§,
- Tift Mann III, MD, FACC∥∥,
- Ravi N. Nair, MD, FACC¶¶,
- Ronald A. Stine, MD, FACC,##,
- Steven J. Yacubov, MD, FACC*** and
- James E. Tcheng, MD, FACC§
- ↵*Reprint requests and correspondence:
Dr. David J. Moliterno, Department of Cardiovascular Medicine, The Cleveland Clinic Foundation, 9500 Euclid Avenue, F-25, Cleveland, Ohio 44195.
Objectives The aim of this study was to discern a target range of anticoagulation for enoxaparin during percutaneous coronary intervention (PCI) as measured by the Rapidpoint ENOX (Pharmanetics Inc., Morrisville, North Carolina), a new point-of-care test.
Background In the U.S., enoxaparin has been used in only a small proportion of PCI procedures, partly because a rapid enoxaparin-specific assay was unavailable.
Methods We analyzed data from 445 enrolled patients receiving subcutaneous or intravenous enoxaparin in a prospective, multicenter study. Serial anticoagulation measurements and clinical outcomes were recorded.
Results The in-hospital composite occurrence of death, myocardial infarction, and urgent target vessel revascularization was 5.4%, and Thrombolysis In Myocardial Infarction (TIMI) major bleeding, minor bleeding, and any reported bleeding occurred in 0.2%, 1.3%, and 7.9% of patients, respectively. No significant association between procedural ENOX times and ischemic events was observed (p = 0.222), although the event rate was 4.0% among those with ENOX times between 250 to 450 s versus 7.2% for those outside this range (p = 0.134). Increasing ENOX time at sheath removal was correlated with any bleeding (p = 0.010) with a 1% increase for every ∼30-s rise.
Conclusions Ischemic events were infrequent, and the rate appeared lowest in the mid-range of ENOX times. Bleeding events increased with increasing ENOX times. These observations, combined with a suggested procedural anti-Xa level of 0.8 to 1.8 IU/ml, translate into a recommended ENOX time range of 250 to 450 s for PCI and <200 to 250 s for sheath removal.
Several large-scale clinical trials have demonstrated the superiority of low-molecular-weight heparin (LMWH) over unfractionated heparin (UFH) for medical treatment of acute coronary syndromes (ACS), resulting in the adoption of LMWH as a standard of care (1). The shift from UFH to fractionated heparin, such as enoxaparin, has been slower in some medical centers where a large proportion of ACS patients transition to percutaneous coronary intervention (PCI), partly because a rapid assay for monitoring enoxaparin has not been available. Systemic anticoagulation with UFH has been an integral part of PCI ever since the procedure's inception, and the activated clotting time (ACT) is routinely used to help balance the thrombotic and hemorrhagic risks associated with procedural anticoagulation.
A rapid point-of-care assay for enoxaparin (Rapidpoint ENOX, Pharmanetics Inc., Morrisville, North Carolina) has been recently developed and was approved by the U.S. Food and Drug Administration (FDA) (2). The purpose of this study was to measure ENOX times and anti-factor Xa activity levels, and to compare these measures with ischemic and hemorrhagic events observed in patients undergoing PCI. The primary objective was to discern a desirable range for procedural anticoagulation and separately to determine a threshold level below which hemostasis can be achieved for vascular access sheath removal.
Patients were eligible for enrollment if they were to undergo elective intent-to-stent PCI and had received at least two doses of subcutaneous (SC) enoxaparin (1.0 mg/kg) within the previous 24 h or had not received antithrombin therapy within 8 h immediately prior to PCI and were to be treated with intravenous (IV) enoxaparin (0.75 to 1.0 mg/kg) for the PCI procedure. Patients were excluded if they had ongoing ST-segment elevation myocardial infarction (MI), recent (<24 h) treatment with fibrinolytic therapy, a baseline ACT >140 s, a prothrombin time international normalized ratio ≥1.5, a history of heparin-induced thrombocytopenia, or a contraindication to aspirin, clopidogrel, or enoxaparin. The study enrolled a total of 673 patients. During enrollment, the FDA approved the ENOX test and required a change in technique. The assays of the final 445 patients (all samples collected into citrated tubes) were used to examine their relationship with clinical outcomes. The protocol was approved by each institution's review board, and all patients gave written informed consent.
Medications and treatment
Guidelines were provided for adjunctive pharmacologic therapies, although the actual selection of agents and dosing was according to institutional practice and standards. Likewise, PCI with any FDA-approved device was allowed provided the intent was to place a stent at all target lesions. Aspirin (325 mg) was to be given the morning of the procedure. Clopidogrel administered either as a 300-mg loading dose on the day of the procedure (before PCI) or 75 mg daily for four days before the procedure was encouraged. The use and type of glycoprotein (GP) IIb/IIIa inhibitor was left to the discretion of the interventionalist. For patients receiving SC enoxaparin and having received their last dose within 8 h of PCI, no further enoxaparin was recommended. For patients who had received their last dose between 8 and 12 h before PCI, a 0.3-mg/kg IV bolus of enoxaparin was suggested (3). These guidelines were the same regardless of whether a patient received a GP IIb/IIIa inhibitor. For patients not already receiving SC enoxaparin, a 0.75-mg/kg IV bolus was to be given immediately before PCI if receiving a GP IIb/IIIa inhibitor during the procedure. For patients not already receiving SC enoxaparin and not receiving a GP IIb/IIIa inhibitor during PCI, a 1.0-mg/kg IV bolus of enoxaparin was to be given.
At the conclusion of PCI, an arteriotomy closure device could be used according to local practice. Otherwise sheaths were to be removed 4 h after the last bolus administration of IV enoxaparin. For patients receiving only SC enoxaparin, sheaths were to be removed 6 to 8 h following the last SC dose. For patients who received SC heparin with an additional pre-procedural dose (0.3 mg/kg) given intravenously, the sheath was recommended to be removed 4 to 6 h after the IV booster dose.
For patients receiving SC enoxaparin without an IV supplement, a sample was collected immediately prior to PCI. For patients receiving SC enoxaparin and an IV supplement, a sample was collected before and after the IV bolus. For patients treated only with IV bolus enoxaparin, a sample was collected 5 to 10 min after the bolus. Whole blood samples were obtained from the femoral arterial sheath (after first discarding 3 ml of blood) and placed into a 3.2% sodium citrated tube (2.7 ml). Immediately upon collection, the samples were tested in duplicate with the ENOX test card. The remaining citrated blood was centrifuged, and the resulting plasma was aliquoted in cryovials and frozen at −70°C. The anti-Xa concentrations in the archived plasma samples were determined in a central coagulation core laboratory.
The Rapidpoint Coag system consists of a microprocessor-based analyzer and single-use assay-specific test cards (Fig. 1). The Rapidpoint photo-mechanically monitors fibrin clot formation in a flat capillary chamber on the surface of a test card. After the test card is inserted into the reader, it is warmed to 37°C, and the operator adds a single drop of citrated whole blood to the sample well. As blood is drawn into the reaction chamber, the reagents are re-hydrated, and paramagnetic iron particles are stimulated to move in the test chamber by an oscillating magnetic field. Within the blood sample, Factor Xa is rapidly produced by a specific Factor X activator, thereby initiating the clotting cascade. As a clot begins to form, fibrin strands attach themselves to the iron particles, impeding their movement. Particle movement is monitored by an infrared optical system, and a preset reduction in particle movement signals the test's end point. Enoxaparin from the patient's blood complexes with antithrombin to inhibit Factor Xa (and to a lesser extent Factor IIa) and proportionally lengthens the clotting time. The ENOX card is optimized for enoxaparin monitoring, but is affected by other heparins. The ENOX times are not affected by thienopyridines or GP IIb/IIIa inhibitors. Among 120 normal volunteers, ENOX times ranged from 106 to 160 s (mean ± 2 SD), and among 166 unanticoagulated patients citrated blood samples ranged from 70 to 180 s (mean ± 2 SD). The correlation between ENOX times and measured plasma anti-Xa levels is high (r = 0.88), with a value of 260 s approximating a plasma anti-Xa level of 1.0 IU/ml (2).
The frozen plasma samples were sent to a central coagulation core laboratory (Hemostasis Reference Laboratory, Henderson Research Centre, Hamilton, Ontario) for batched anti-Xa analysis. The plasma levels of enoxaparin were assessed by measurement of the anti-Xa activity using the amidolytic method with a chromogenic substrate (Stachrom Heparin assay [Diagnostica Stago, Asnieres-Sur-Seine, France] and the AMAX 190 instrument [Heinrich Amelung, Lemgo, Germany]) (4).
End points and statistics
Clinical and procedural data were prospectively entered onto case report forms specific for this trial. Samples for cardiac marker assays (creatine kinase and its MB fraction) were obtained at 8, 16, and 24 h post-PCI or until hospital discharge, whichever occurred first. Clinical outcomes including bleeding complications were assessed to seven days or hospital discharge, whichever occurred first. Periprocedural MI was defined as the development of new Q waves after PCI in two or more contiguous leads or a ≥3-fold elevation in the creatine kinase or its MB fraction in one sample. Major bleeding was defined according to the Thrombolysis In Myocardial Infarction (TIMI) criteria and included a hemoglobin drop ≥5 g/dl or intracranial hemorrhage. Minor bleeding was defined according to the TIMI criteria and included a blood loss between 3 and 5 g/dl with a known bleeding site or a loss of <3 g/dl associated with gross hematuria, hematemesis, or hemoptysis (5). Any bleeding, by study definition, included the above plus the occurrence of access site complications including hematoma >5 cm, significant re-bleeding after initially achieving hemostasis, or bleeding delaying hospital discharge.
Descriptive statistics included demographic, medication, laboratory, and procedural details, as well as ENOX times, anti-Xa activity levels, and clinical events. For continuous variables, mean (± SD) and median (25th, 75th percentile) values were determined. For categorical variables, percentages were calculated, and comparisons were performed with a chi-squared test. To explore outcome relationships, scatter plots with kernel regression curves (6)were created for the composite ischemic end point versus procedural ENOX times and for any bleeding event versus pre-sheath removal ENOX times. Logistic regression was used to model the relationship between probability of an outcome and the natural log of the ENOX time. The modeling of the dose-response relationships between the Rapidpoint citrated ENOX time and reference laboratory assessments of anti-Xa activity was performed using the standard Emaxcurve (7)estimated within the nonlinear mixed effect model framework. A 0.05 level was used to test statistical significance.
From February 2002 to September 2002, a total of 673 patients were enrolled at 15 U.S. medical centers (see online Appendix, which can be found at http://www.cardiosource.com/jacc.html). Patient demographics for the final 445 patients are listed in Table 1. The cohort was 69% male, an average age of 63 years, and 52% had a primary diagnosis of unstable angina or recent MI. All patients received enoxaparin, and 75% received a concomitant GP IIb/IIIa inhibitor. During the time of the study it was separately appreciated that precision with ENOX samples collected into citrated tubes was greater than that with non-citrated samples, and this sample type became FDA approved for the assay (2). Therefore, the final Evaluating Enoxaparin Clotting Times (ELECT) study analyses were performed using data from the last 445 patients, all of who had citrated samples collected. The baseline characteristics of these patients were similar to the overall cohort. The routes of enoxaparin administration and concomitant use of GP IIb/IIIa agents among patients are depicted in Figure 2. The two groups of patients receiving SC enoxaparin (i.e., with and without a GP IIb/IIIa inhibitor) were combined for analysis as only two patients received enoxaparin subcutaneously without a GP IIb/IIIa inhibitor. The three enoxaparin dosing strategies: 1) 1.0 mg/kg subcutaneously twice daily, 2) 0.75 mg/kg single IV bolus with a GP IIb/IIIa inhibitor, or 3) 1.0 mg/kg IV bolus without a GP IIb/IIIa inhibitor, corresponded to 8%, 68%, and 24% of the final study group, respectively. Procedure-related details are listed in Table 2.
ENOX times and anti-Xa levels
The Emaxrelationship between ENOX times and central laboratory anti-Xa levels was moderate (Fig. 3) (r = 0.75), and values for the various collection times are listed in Table 3. Among patients receiving SC enoxaparin for an ACS, the respective median ENOX times and anti-Xa levels were 208 s and 0.9 IU/ml upon arrival at the catheterization laboratory. After a 0.3-mg/kg enoxaparin supplemental bolus for patients whose last SC dose was between 8 to 12 h earlier, the median ENOX time and anti-Xa level increased to 257 s and 1.4 IU/ml, respectively. The procedural ENOX times for those receiving a 0.75 and 1.0 mg/kg single IV bolus of enoxaparin were 412 and 433 s, and these corresponded to anti-Xa levels of 1.5 and 1.9 IU/ml, respectively. At the time of sheath removal, the overall group's median ENOX and anti-Xa levels decreased to 239 s and 0.8 IU/ml; both results were approximately 40% to 50% lower among those whose sheath was removed several hours after the procedure as compared with those whose sheath was removed immediately after PCI and an arteriotomy closure device was used.
Overall ischemic and hemorrhagic event rates were low (Table 4) with an in-hospital composite of death, MI, and urgent target vessel revascularization (TVR) occurring in 5.4% of patients. This ischemic composite, as in many contemporary PCI-stent trials, was largely comprised (87%) of procedure-associated MI, as death and urgent TVR occurred in <1% of patients. TIMI major and minor bleeding episodes were observed in 0.2% and 1.3% of patients, respectively. “Any bleeding” occurred in 7.9% of patients. Transfusions were needed in only 1.3% of patients, and the bleeding index was 0.7 ± 1.2.
Relationship of ENOX times with adverse events
The mean procedural ENOX time for patients with and without an ischemic event was similar (461 and 429 s, respectively). A kernel regression curve plotting ENOX times versus ischemic events (Fig. 4) shows a nadir event rate at approximately 300 to 350 s; however, using a logistic regression model, no significant association between ischemic events and ENOX times was observed (p = 0.222). The average ENOX value prior to sheath removal was 303 s among those with any bleeding event versus 266 s among those without bleeding. The kernel regression curve plotting ENOX values versus any bleeding shows increasing bleeding with increasing ENOX values over most of the clotting time range (Fig. 5), and the positive slope effect for the regression model was significant (p = 0.010).
Among the 445 patients studied with citrated samples, 2 patients died. One patient was a 78-year-old woman undergoing right coronary artery PCI who was given IV enoxaparin and abciximab. The procedure required extensive stent placement for threatened vessel closure and was complicated by dissection into the ascending aorta. The procedural ENOX time was low (177 s), and it was speculated the IV catheter became infiltrated into the SC tissue before enoxaparin administration. The second patient was a 63-year-old man who had been treated for two days with IV eptifibatide and twice daily SC enoxaparin for a non–Q-wave MI. During PCI of a saphenous vein bypass graft, abrupt closure occurred. The patient had a moderate-sized infarction and subsequently expired. The blinded procedural ENOX time was 163 s, and it was discovered that the patient had inadvertently not received the morning dose of enoxaparin before being transferred to the catheterization laboratory.
During PCI, the most commonly used anti-thrombin agent remains UFH. This heterogeneous agent has many known limitations, and LMWH and direct thrombin inhibitors have shown equivalence or superiority over UFH in several PCI trials (8–10). Conversely, the volume of experience with UFH and its monitoring with the ACT is extensive. Cited concerns regarding use of enoxaparin as a primary anticoagulant for PCI or for bridging upstream enoxaparin with its use in the catheterization laboratory include the inability to rapidly monitor its activity (3). Because enoxaparin contains various chain lengths, it too produces a several-fold range of anticoagulation among patients, although to a lesser degree than UFH. Our study is the first to systematically assess a rapid, point-of-care enoxaparin monitor during PCI and to correlate these results with clinical outcome.
Our observation that periprocedural ischemic event rates are low (5.4%) and not clearly linked to the measured extent of LMWH anticoagulation in the GP IIb/IIIa era complement the findings of Chew et al. (11)who demonstrated a relatively flat dose-response relationship between ACT and ischemic events among 3,876 patients receiving UFH with GP IIb/IIIa inhibitors and of Tolleson et al. (12), who observed similar findings in the 2,064-patient Enhanced Suppression of the Platelet IIb/IIIa Receptor with Integrilin Therapy (ESPRIT) study. The suggestion of a U-shaped event curve has been reported by others relating level of anticoagulation to outcome events during ACS and PCI (11,13). Considering enoxaparin specifically, observational PCI studies, such as National Investigators Collaborating on Enoxaparin (NICE)-1 and NICE-4, and more recently studies by Choussat et al. (17)and Carnendran et al. (18), have tested progressively lower doses of IV enoxaparin (1.0, 0.75, and 0.5 mg/kg, respectively) over time (14–18). Although these studies differed regarding design, patient cohort, devices, and GP IIb/IIIa inhibitor use, it is noteworthy that both the 30-day ischemic event rates and major plus minor bleeding rates have steadily declined (Table 5). This pattern of trial results also suggests that the extent of anticoagulation may not be tightly linked to ischemic events among contemporary stent GP IIb/IIIa procedures or that other factors are relatively more important.
Bleeding from a “relative excess” of anticoagulation has been observed with virtually all anticoagulants, and in most cases the bleeding dose-response is positive and linear. In the present study with enoxaparin, the kernel smoothing plot shows a 1% increase in any bleeding for every ∼30 s increase in ENOX time (Fig. 5). Although these events were primarily nuisance bleeding, such as groin hematoma, they were predictable and frequent at higher ENOX times. As the ENOX time prior to sheath removal decreased to under 200 s, the composite bleeding event rate fell below 5%. Also unique to our study was the use of arteriotomy closure devices; prior enoxaparin-in-PCI studies have precluded their use. Although the use of closure devices was not randomized in the ELECT study, we observed a greater than doubling in the rate of any bleeding events with closure devices as compared with manual compression (12.2% vs. 5.7%). As expected, patients receiving a closure device had higher ENOX times at the time of sheath removal (Table 3).
Optimal ENOX time range
A key goal of this study was to gain insight regarding an optimal procedural ENOX time for performing PCI and a level for safe arterial sheath removal. Although no statistically significant association was found between ENOX times and ischemic events, we made several observations: 1) a nadir of ischemic events did appear to be present between 300 to 350 s; 2) both ischemic and bleeding events rates appeared to be unacceptably high both in our dataset and in prior observations (14,15,19)(Table 5) once the ENOX time exceeded 450 s or the anti-Xa level exceeded 1.8 IU/ml; and 3) the overwhelming majority of patients receiving either SC or IV enoxaparin had procedural ENOX values ≥250 s. Based on these conditions, we made a single post hoc analysis separating patients into two groups according to whether or not the procedural ENOX clotting time was between 250 to 450 s. The ischemic composite of death, MI, and urgent TVR occurred in 4.0% of patients within the suggested procedural ENOX range versus 7.2% (p = 0.134) among those whose values were <250 or >450 s.
The safety of arterial sheath removal is affected by several unrelated factors such as underlying peripheral vascular disease, sheath size, duration of sheath dwelling, and extent of anticoagulation. Because bleeding rates were directly correlated with ENOX times, it is intuitive that lower ENOX times are desirable prior to sheath removal, yet delayed sheath removal increases the risk of vascular complications. Therefore, we propose sheath removal once the ENOX time is within the 200 to 250 s range guided by the clinical scenario. This ENOX range corresponds to anti-Xa levels of ≤0.6 to 0.8 IU/ml and begins to approach 2 SDs beyond the upper limit of normal for unanticoagulated patients (180 s). For patients at increased bleeding risk (e.g., concomitant GP IIb/IIIa therapy, larger sheath size, presence of peripheral vascular disease) the ENOX time should be closer to 200 s; whereas those at lower risk can likely have safe sheath removal at values nearer 250 s.
There are several limitations to our study that are primarily related to the sample size and the relatively low rates of ischemic and bleeding events. Given the small number of adverse events, it was not possible to correlate major clinical outcomes with ENOX times. Although an apparent pattern emerged for ischemic events and a significant pattern for all bleeding events, further studies will be needed to test the optimal ENOX time regarding clinical outcome. The targeted anti-Xa range of 0.8 to 1.8 IU/ml must be considered somewhat arbitrary as no prospective study has been performed. This range was based upon prior and current observations and upon the American College of Chest Physicians' consensus guidelines to have anti-Xa levels ≥0.6 IU/ml for medical therapy (19,20). Finally, we did not enroll patients with acute MI, so that further study will be needed for both these and higher-risk patients.
In this enoxaparin-in-PCI study, ischemic events were infrequent although lower in the mid-range of ENOX times measured. Bleeding events were primarily limited to the access site and were more common with increasing ENOX times. These observations, combined with suggested procedural anti-Xa levels of 0.8 to 1.8 IU/ml, translate into a recommended ENOX range of 250 to 450 s for PCI and <200 to 250 s for safe arterial sheath removal. Ongoing and future studies will be able to assess the clinical relevance of ENOX times during partial enoxaparin reversal with protamine and to discern whether booster doses of enoxaparin for “sub-therapeutic” anticoagulation are helpful or needed.
Participating centers and coordinators: Cleveland Clinic: P. Welsh, R. Bartow, G. McConnell; Duke University: C. Martz, P. Gottlieb, E. Yow, V. Hasselblad, D. Joseph, K. Raffetto, S. Wu, M. Roe; Forsythe Hospital: S. Key, J. Stone; Genesis Heart Institute: V. Takes; Houston VAMC: N. Hinton-Davis; Lindner Center: M. Mueller; Baylor College of Medicine: K. Maresch; Riverside Methodist Hospital: P. Vieira; Norfolk General: N. Sullivan; St. Vincent's Hospital: L. Burkert; University of Massachusetts: S. Ball; University Hospital of Cleveland: L. Hickel; Wake Forest University: T. Young; Wake Heart Associates: F. Wood; William Beaumont Hospital: S. Didocha; Mayo Clinic: P. Berger.
For a list of the medical centers, investigators, and research coordinators who participated in ELECT, please see the September 17, 2003, issue of JACCat http://www.cardiosource.com/jacc.html.
☆ This study was funded by a research grant from Pharmanetics Inc., Morrisville, North Carolina.
- acute coronary syndrome
- activated clotting time
- Food and Drug Administration
- low-molecular-weight heparin
- myocardial infarction
- percutaneous coronary intervention
- Thrombolysis In Myocardial Infarction
- target vessel revascularization
- unfractionated heparin
- American College of Cardiology Foundation
- Braunwald E.,
- Antman E.M.,
- Beasley J.W.,
- et al.
- ↵ENOX Package Insert. Available at: http://enoxtest.com/pdfs/enoxpackageinsert.pdf. Accessed on June 30, 2003
- Rao A.K.,
- Pratt C.,
- Berke A.,
- et al.
- Wand M.P.,
- Jones M.C.
- ↵Hochhaus G, Derendorf H. Dose optimization based on pharmacokinetic-pharmacodynamic modeling. In: Derendorf H, Hochhaus G, editors. Handbook of Pharmacokinetic/Pharmacodynamic Correlation. Boca Raton, FL: CRC Press Inc., 1995:79–120
- Bhatt D.L.,
- Lee B.I.,
- Casterella P.J.,
- et al.
- Karsch K.R.,
- Preisack M.B.,
- Baildon R.,
- et al.
- Chew D.P.,
- Bhatt D.L.,
- Lincoff A.M.,
- et al.
- Tolleson T.R.,
- O'Shea J.C.,
- Bittl J.A.,
- et al.
- Granger C.B.,
- Hirsch J.,
- Califf R.M.,
- et al.
- Collet J.P.,
- Montalescot G.,
- Lison L.,
- et al.
- The Thrombolysis In Myocardial Infarction (TIMI) 11A Trial Investigators