Design and Implementation of an Anti–Factor Xa Heparin Monitoring Protocol

Tanya Williams-Norwood, Clinical Pharmacy Service; VA Northeast Ohio Healthcare System, Cleveland, Ohio.

Tanya Williams-Norwood

Clinical Pharmacy Service; VA Northeast Ohio Healthcare System, Cleveland, Ohio.

Megan Caswell

Internal Medicine, Cleveland Clinic Foundation, Cleveland, Ohio.

Barbara Milner, RN, MSN, CNS, Clinical Nurse Specialist and Joseph C. Vescera, PharmD, MBA, Associate Chief of Inpatient Pharmacy

Kelly Prymicz

Chelko Consulting Group, Westlake, Ohio.

Amy G. Ciszak, BSN, RN, Clinical Applications Coordinator, Informatics and Analytics , Carol Ingle, RPh, CACP, Clinical Pharmacy Manager of the Anticoagulation Clinic , and Christopher Lacey, PharmD, BCPS, Associate Chief

Evi X. Stavrou, Staff Physician and Medical Director of Anticoagulation Clinic, VA Northeast Ohio Healthcare System; Oscar D. Ratnoff Professor in Medicine and Hematology; Assistant Professor in the Department of Medicine at Case Western Reserve University School of Medicine, 10701 East Blvd, Cleveland, OH, 44106.

Evi X. Stavrou

Staff Physician and Medical Director of Anticoagulation Clinic, VA Northeast Ohio Healthcare System; Oscar D. Ratnoff Professor in Medicine and Hematology; Assistant Professor in the Department of Medicine at Case Western Reserve University School of Medicine, 10701 East Blvd, Cleveland, OH, 44106.

Abstract

Background:

The VA Northeast Ohio Healthcare System introduced a new nurse-driven anti–factor Xa (anti-Xa) protocol for monitoring unfractionated heparin to replace the previous activated partial thromboplastin time protocol.

Objective:

To design, implement, and evaluate the efficacy of the anti-Xa monitoring protocol.

Methods:

An interdisciplinary team of providers collaborated to develop and implement a nurse-driven, facility-wide anti–factor Xa protocol for monitoring unfractionated heparin therapy. The effectiveness of this protocol was evaluated by retrospective analysis.

Results:

We reviewed 100 medical records for compliance with the new anti-Xa monitoring protocol. We then evaluated 178 patients whose anticoagulation was monitored with the anti-Xa assay to determine the time to therapeutic range. We found that 80% of patients receiving the anti- Xa protocol achieved therapeutic anticoagulation within 24 hours, as compared with 54% of patients receiving the activated partial thromboplastin time protocol (P < .001). Protocol conversion also yielded a decrease in blood draws, dose adjustments, and potential calculation errors.

Conclusions:

Monitoring intravenous heparin therapy with the anti-Xa assay rather than activated partial thromboplastin time resulted in a shorter time to therapeutic anticoagulation, longer maintenance of therapeutic levels, and fewer laboratory tests and heparin dosage changes. We believe the current practice of monitoring heparin treatment with activated partial thromboplastin time assays should be reexamined.

Intravenous (IV) unfractionated heparin (UFH) remains the mainstay of treatment for thromboembolic conditions such as acute coronary syndrome, ischemic stroke, deep vein thrombosis, and pulmonary embolism. Because of its narrow therapeutic window, timely monitoring of heparin anticoagulation and prompt dose adjustments are necessary. 1 For decades, the anticoagulant effect of heparin has usually been monitored by the activated partial thromboplastin time (aPTT), a global coagulation assay that primarily reflects the function of the intrinsic and common path-ways of the coagulation cascade. However, the performance of aPTT is affected by pre-analytic, analytic, and biologic variables. In acutely ill hospitalized patients, a plethora of conditions, including acute sterile inflammation seen with surgery, acute coronary syndrome, stroke, infection, and liver disease, can interfere with aPTT results. 2–4 Given that several biologic factors can influence aPTT independent of the effects of UFH, institutions have started to transition to monitoring heparin with anti–factor Xa (anti-Xa) levels rather than with aPTT. The anti-Xa assay is a functional assay that measures the inhibition of factor Xa by UFH. 5–8 Any residual factor Xa that is unbound by the antithrombin-heparin complex is free to interact with a chromogenic substrate added to the plasma sample. 9 This interaction between unbound factor Xa and the chromogenic substrate produces color that can be measured optically. The intensity of color produced is inversely proportional to the amount of heparin present. Values are plotted against a standard curve and results are expressed as units per milliliter of anti-Xa activity, representing the concentration of heparin in the sample. In contrast to the aPTT assay, the anti-Xa assay is unaffected by the presence of lupus anticoagulant, liver disease, acute inflammatory states, or consumptive coagulopathy and currently is perceived as the best method to measure the functional activity of heparin. The purpose of this article is to report the design, implementation, and efficacy evaluation of a nurse-driven anti-Xa monitoring protocol.

Literature Review Comparing aPTT and Anti-Xa Assays

The issue of monitoring UFH remains an underrepresented area of research compared with studies reporting UFH efficacy. Van Roessel et al 4 reported a cohort study of 58 critically ill patients (totaling 171 blood samples) who were treated with therapeutic doses of heparin. Because the heparin anti-Xa assay is not commonly influenced by acute inflammatory conditions, the goal of their study was to determine the accuracy of the aPTT test by comparing it with the heparin anti-Xa assay. The researchers reviewed aPTT and anti-Xa results of 108 data pairs. The data showed that aPTT was disproportionately high (above the preset therapeutic range) in 33 data pairs and disproportionately low in 30 data pairs. According to the receiver operating characteristic curves, the sensitivity of aPTT for detecting underdosing and overdosing was low, at 0.63 and 0.37, respectively. The authors concluded that aPTT testing is not the most accurate method to monitor heparin anticoagulation in critically ill patients. 4

Price et al 11 reviewed 2321 paired aPTT and anti-Xa assay results from 539 patients. Samples for all paired tests were obtained simultaneously. The authors determined the frequency of discordant aPTT and anti-Xa values, demographics, coagulation status, indications for UFH, and clinical outcomes in the study population. Approximately 42% of data pairs had discordant results, with supratherapeutic aPTT results relative to anti-Xa values. Patients with a minimum of 2 consecutive aPTT values higher than anti-Xa values had a higher incidence of bleeding at 21 days (9% vs 3%, P = .03) and higher 30-day mortality (14% vs 5%, P = .02) than did patients with consistently concordant results. 11

Frugé and Lee 12 conducted an observational study at a 500-bed private community hospital to compare patients receiving the institution’s aPTT protocol (n = 79) with a group of patients receiving an anti-Xa monitoring protocol (n = 42). At 24 hours, 74% of patients in the anti-Xa group had reached the goal anticoagulation range, whereas 63% of patients in the aPTT group had reached the therapeutic range (P = .2). The study also found that approximately 57% of patients in the anti-Xa group and 27% in the aPTT group reached therapeutic anticoagulation within 6 hours (P = .001). The researchers reported a mean of 1.00 dosage adjustments per patient in the anti-Xa group and 1.71 dosage adjustments per patient in the aPTT group within the first 24 hours (P = .003). A decrease in costs was due to significantly less variability in anti-Xa measurements, quicker attainment of therapeutic levels, and fewer dose adjustments with the anti-Xa protocol than with the aPTT protocol. 12

These studies indicate that the anti-Xa assay is superior to aPTT for monitoring UFH because it measures the actual enzymatic activity of factor Xa in the clotting cascade with less interference from analytic and biologic variables. 10,11 Published studies also support the premise that monitoring intravenous UFH with anti-Xa levels achieves a faster time to therapeutic range and warrants less laboratory testing than does monitoring with aPTT testing.

Care Setting

The VA Northeast Ohio Healthcare System is a 660-bed facility with 237 acute care beds. The system includes national and regional referral services for cardiac surgery, vascular surgery, spinal cord injury, invasive cardiology, behavioral health addictions, diabetes mellitus, and psychiatric services. The inpatient wards (3 intensive care units, 1 progressive care unit, and 5 medical-surgical units) are in close proximity to one another, facilitating transfers because nursing staff members often work interchangeably between sister units. On average, the facility treats over 2500 patients with heparin (subcutaneous, intramuscular, IV bolus, and IV infusion) each year. From June 1, 2017, to June 1, 2018, the facility treated 2665 patients with IV heparin; 26% of these patients (695) were treated with IV heparin infusions.

Project Timeline

From 2008 to 2016, a nurse-driven aPTT heparin protocol was the standard method of heparin monitoring. With this protocol, a licensed provider ordered the heparin infusion and the bedside nurse administered heparin after calculating the initial bolus dose and infusion rate. To monitor the patient’s response to heparin therapy, aPTT values were obtained at standard times (5 am , 1 pm , and 9 pm ). The nurse adjusted the heparin infusion according to the aPTT results. After a patient had 2 consecutive therapeutic results, an aPTT value was obtained once daily throughout the treatment duration.

In 2010, the effectiveness of the nurse-driven aPTT protocol was evaluated via a retrospective medical record review comparing 50 patients who received empirical dosing and 50 patients who received the nurse-driven aPTT heparin protocol. This study revealed that 16% of patients in the empirical dosing group never achieved therapeutic anticoagulation, as compared with 8% of patients in the nurse-driven aPTT protocol group (P < .05). 13 However, despite this favorable difference, alternative assays were explored to further improve the quality and safety of anticoagulation management.

In November 2014, the medical center was designated a reference center for implantation of left ventricular assist devices as destination therapy for end-stage heart failure. In 2015 the hematologist introduced the anti-Xa assay as a more precise method to manage heparin therapy for this patient population. With use of the anti-Xa assay, patients with left ventricular assist devices consistently achieved a therapeutic range of anticoagulation within 24 hours. Because of these documented favorable results, transitioning inpatient heparin monitoring to a nurse-driven anti-Xa protocol was considered. Staff and equipment needs were therefore analyzed over the next several months.

The proposal was presented to the Medical Executive Committee for approval. Once approved, an interdisciplinary team consisting of a staff hematologist, pharmacists, clinical nurse specialists, nurse informatics staff, bedside nurses, and phlebotomists began to meet weekly in January of 2016 to develop a protocol and discuss project implementation plans. Once the proposal was approved, the team continued to meet weekly to initiate the steps necessary to implement the practice change.

The laboratory department purchased new anti-Xa processing equipment and trained staff members on its use. A clinical nurse specialist, in collaboration with the pharmacy and nursing informatics departments, developed and led an education simulation program. Over 200 staff nurses and a select number of providers were recruited to provide input on how to best implement the project. In anticipation of implementation, logistical procedures were initiated and all IV pumps were updated with the new protocol ( Figure 1 ).

Timeline of development and implementation of the nurse-directed anti-Xa protocol. anti-Xa indicates anti–factor Xa; aPTT, activated partial thromboplastin time; CNS, clinical nurse specialist; LVAD, left ventricular assist device.

Protocol Implementation

In August 2016, the facility implemented the new nurse-driven heparin protocol with anti-Xa monitoring. This protocol is divided into a low-intensity protocol and a high-intensity protocol according to actual body weight, not to exceed 125 kg. Indications, loading dose, and maintenance dose for each protocol are shown in Table 1 . The high-intensity protocol has specific provisions for use of a lower loading dose for patients with atrial fibrillation in the immediate perioperative period. 10 Heparin monitoring with the anti-Xa assay is conducted 4 hours after initiation and after any changes in infusion rate until the therapeutic range is reached. The therapeutic anti-Xa range is defined as 2 consecutive results within 0.3 to 0.79 IU/mL. Once the patient’s heparin level is in the therapeutic range, an anti-Xa assay is obtained once daily.

Table 1:

Low-Intensity and High-Intensity Heparin Protocols a

Methods

Study Procedures and Oversight

This quality improvement project was reviewed and approved by the institutional review board. Because of the potential for significant patient harm during optimization of IV heparin use, the project was pursued with careful attention to patient-specific risks and benefits. To ensure that any identified problem was addressed quickly, an interdisciplinary team of providers, clinical nurse specialists, bedside nurses, pharmacists, nurse informaticists, and phlebotomists monitored and supervised this project.

Objectives

The primary objective was to develop a safe and effective anticoagulation protocol that decreased the potential for calculation errors, allowed timely intervention on the basis of anti-Xa results, and reduced resource use secondary to a decreased need for laboratory monitoring, repeat bolus doses, and infusion rate changes. Primary end points included protocol compliance, protocol efficiency (assessed by comparing the performance of anti-Xa monitoring with that of the aPTT-based protocol), and safety measurements.

Statistical Analysis

Assuming an observed area under the curve of 0.75 and an α of .05, we calculated that a group of at least 50 patients would provide greater than 90% power for a 2-sided test of the null hypothesis (area under the curve = 0.5). This power calculation was performed with the pROC package in R version 3.5.1 (R Foundation) and informed our goals for the number of medical records to be reviewed for protocol compliance. Data obtained in this study are reported as means and 95% CIs unless stated otherwise. We used the Mann-Whitney U test to compare outcomes between the 2 protocol groups and Pearson analysis to determine correlation between anti-Xa and aPTT assays. All statistical comparisons were 2-tailed; P values less than .05 were considered significant.

Implementation

With the previous aPTT protocol, nurses used an independent double-check system to calculate the initial bolus dose and rate, initial infusion dose and rate, and subsequent dose adjustments. This process required the nurse to validate all calculations with a second registered nurse who used a different calculation method (eg, traditional calculator, heparin calculation spreadsheet). Errors and near misses were rare; however, if the initial dose was incorrectly calculated, errors could potentially persist with subsequent rate changes.

To negate the need for nursing calculations, the anti-Xa protocol requires that the initial bolus dose/rate and infusion dose/rate be ordered by the provider and verified by a pharmacist. Once verified, the order is released into the electronic medical record (EMR) with directions to nursing staff indicating the exact bolus volume (in milliliters) and the infusion rate. For example, with the previous aPTT protocol, if a 75-kg patient had an order to receive therapy according to the high-intensity protocol (bag concentration: 25 000 U heparin in 250 mL of 0.45% sodium chloride), the nurse would perform the following calculations. (1) Bolus: 75 kg × 80 U/kg = 6000 U. Conversion to milliliters: 6000 U × (250 mL / 25 000 U) = 60 mL. (2) Infusion rate: 75 kg × 18 U/kg/h = 1350 U/h. Conversion to milliliters per hour: 1350 U/h × (250 mL / 25 000 U) = 13.5 mL/h. With the current anti-Xa protocol, these calculations are already completed by the provider and validated by pharmacy staff, and each order to the registered nurse simply indicates a bolus of 60 mL and an infusion rate of 13.5 mL/h. Because dose adjustments are managed according to weight ranges, dosing charts are used for all infusion adjustments, which again eliminates the need for dose calculations.

A main objective was to standardize laboratory processing times to obtain anti-Xa assay results in a timely manner. Because the start times for heparin infusions can vary, the provider order allowed the bedside nurse to time all laboratory orders. This process was new, and it eliminated the need for the previous set schedule of blood draws (5 am , 1 pm , and 9 pm ) that was unyielding to issues such as various initiation times, infusion hold times, or unavailability of patients who were off the ward for procedures.

Once nurses initiated or adjusted the heparin infusion, they managed a “clock” by documenting in the EMR the date and time of the next blood draw. After the date and time were entered in the EMR, the laboratory was notified by computer alert when the next heparin anti-Xa assay was due. To ensure that laboratory tests were obtained in a timely manner, dedicated phlebotomists were assigned solely to collect samples from patients undergoing heparin protocols. Once collected, samples were promptly returned to the laboratory for processing and the results were entered into the EMR. The protocol allowed for 60 minutes of processing time, defined as the time from blood draw to the reporting of results in the EMR.

To assess protocol compliance, we reviewed 100 patient records. We identified all patients with standard UFH IV infusion protocol orders by querying the pharmacy information system. We selected 100 consecutive patients from this population. Data collected included general demographic information such as admission date and age; weight at the start of the infusion; baseline platelet count; baseline hemoglobin level; and indication for heparin anticoagulation ( Table 2 ). We collected data for up to 4 days or until the heparin infusion was stopped or interrupted. Patients who received the anti-Xa protocol for less than 24 hours and patients with left ventricular assist devices were excluded from the analysis.

Table 2:

Baseline Demographics of Male Patients Receiving Heparin Anticoagulation Monitored With the Anti–Factor Xa Assay (N = 100)

Results

All patients who began the high-intensity protocol (n = 70) were appropriately allocated at 100% compliance. In the low-intensity group (n = 30), 24 patients were appropriately allocated; the remaining 6 patients were treated for atrial fibrillation but had undergone recent vascular surgery, so the low-intensity protocol was chosen to decrease their bleeding risk ( Table 3 ).

Table 3:

Staff Adherence to High-Intensity and Low-Intensity Heparin Protocols

a Six patients with atrial fibrillation were placed on the low-intensity protocol because of high bleeding risk.

The correct initial heparin dose was ordered 96% of the time. Dose adjustments across both protocols were correct in 92% of patients, and only 12 patients had a critical anti-Xa value. Of the 12 patients with a critical value, the protocol for heparin dose adjustment was appropriately followed in 11 (92%). The mean bolus of heparin (100 U/mL) was 40 mL (range, 0–40 mL) in the low-intensity population and 43.7 mL (range, 0–100 mL) in the high-intensity population. The mean initial infusion rate for patients in the low-intensity group was 9.73 mL/h (range, 6–14 mL/h), with a mean therapeutic infusion rate of 12.36 mL/h (range, 8–20 mL/h). The mean initial infusion rate in the high-intensity population was 15.4 mL/h (range, 8–20 mL/h), with a mean therapeutic infusion rate of 14.9 mL/h (range, 7.1–23 mL/h). The mean laboratory processing time ranged from 20 to 45 minutes.

Once protocol compliance was established, we investigated the efficacy of the anti-Xa monitoring protocol. Patients with left ventricular assist devices, UFH treatment for less than 24 hours, or treatment interrupted for more than 10 hours were excluded from analysis. We identified 178 patients who met the criteria for analysis from August 2016 through March 2017. A pharmacist and the staff hematologist independently reviewed all data sets. The data indicated that 80% (n = 143) of patients receiving the nurse-driven anti-Xa protocol achieved a therapeutic range of anticoagulation within 24 hours. Forty-eight of these 178 patients continued receiving IV UFH for more than 24 hours. Patients who had achieved therapeutic anticoagulation at 24 hours continued to have therapeutic levels at 36 hours ( Figure 2 ). A mean of 3 blood draws on day 1, 2 blood draws on day 2, 1 blood draw on day 3, and 1 blood draw on day 4 were required to achieve therapeutic anticoagulation according to anti-Xa testing. This result correlates with the finding that most patients reached the therapeutic anticoagulation range within 24 hours, thus requiring fewer blood draws.

Percentage of patients reaching therapeutic range of anticoagulation monitored with the newly implemented anti–factor Xa (anti-Xa) assay and and the activated partial thromboplastin time (aPTT) assay (historic controls). Data represent means (SDs). Note that patients who achieved therapeutic levels of anticoagulation at 24 hours with anti-Xa monitoring remained at therapeutic levels at 36 hours. aPTT indicates activated partial thromboplastin time.

To compare the newly implemented anti-Xa protocol with the previous aPTT monitoring protocol, we conducted a retrospective medical record review of 604 patients whose heparin anticoagulation was monitored with the aPTT-based protocol. This comparison showed that a significantly lower percentage of patients in the aPTT-monitored group than in the anti-Xa group achieved therapeutic anticoagulation at 24 hours (54% vs 80%, respectively; P < .001) (Figure 2 ).

Although not part of the original study design, several aPTTs (n = 94) were obtained at the same time as anti-Xa levels. We reviewed these values to determine the degree of correlation between therapeutic aPTT values and anti-Xa levels. As shown in Figure 3 , only 46% of the aPTT results were concordant with anti-Xa levels (43 of 94 paired tests). Although the correlation between aPTT and anti-Xa was found to be positive, the strength of the association was weak. The data show poor correlation between the aPTT and anti-Xa assays (r = 0.1088).

Correlation between activated partial thromboplastin time (aPTT) and anti–factor Xa (anti-Xa) assay results. Analysis was performed with the Pearson correlation. The boxed area represents therapeutic values for both the anti-Xa and aPTT assays. Values directly above and below the boxed area represent situations in which the anti-Xa level would have been in the therapeutic range but the corresponding aPTT was outside the therapeutic range. Values directly to the right and left of the box are situations in which the aPTT was in the therapeutic range but the corresponding anti-Xa levels were outside the range. Although the correlation between aPTT and anti-Xa values was found to be positive, the strength of the association was weak. These data show that there is poor correlation among the aPTT and anti-Xa assays.

We evaluated all study patients for safety outcomes including incidence of major and minor bleeding and development of heparin-induced thrombocytopenia. Major bleeding was defined as a drop in hemoglobin level of greater than 2 g/dL, a need for transfusion of 2 or more units of packed red blood cells, bleeding into the retroperitoneum, intracranial hemorrhage, or hemarthrosis. 14 Minor bleeding was defined as all documented bleeding events not meeting criteria for major bleeding. 14 Heparin-induced thrombocytopenia was defined as a drop in platelet count with documented positivity for platelet factor 4 or serotonin release assay. Monitoring of blood counts and review of narrative notes to determine bleeding or drop in platelet count began at the initiation of heparin therapy and continued throughout the duration of treatment. No major bleeding events or heparin-induced thrombocytopenia occurred in all cases reviewed. One minor bleeding event was documented in a patient receiving the low-intensity heparin protocol.

Discussion

Converting to a new system of heparin monitoring requires a concerted effort on a number of levels. At our institution, the leadership team’s willingness to reallocate funds to purchase equipment and reassign laboratory personnel for this effort, all to improve patient care, was vital. Teams of nurses and physicians worked closely together to compile the revised protocol and enlisted the contributions of pharmacy and laboratory department leaders. Other key factors before converting were the creation of appropriate software templates in the EMR and the hospital-wide education of nurses and providers for proper execution of the heparin protocols. Since our new protocol was implemented over 24 months ago, staff members have expressed widespread satisfaction because of fewer blood draws, fewer dosage adjustments, and decreased time to therapeutic range. More than 24 months after implementation, the nurse-driven heparin anti-Xa protocol has been successful at the medical center. Consistent with prior studies, more patients achieved therapeutic heparin anticoagulation within 24 hours, and this premise has been further sub-stantiated by nurse feedback. Nurses representing the intensive care units, progressive care unit, and medical-surgical areas unanimously indicated that (1) the IV pumps were easy to program, (2) the anti-Xa protocol was easier to use than the aPTT protocol for titrating the infusion, (3) nurses valued having the pharmacy calculate and verify the initial bolus and infusion rates, and (4) nurses preferred the anti-Xa protocol over the aPTT protocol.

A common concern regarding the transition to anti-Xa assay monitoring is cost. However, studies show that the increased cost is negated by increased sensitivity and decreased need for monitoring over time. A study by Rosborough 15 aimed to determine the excess cost incurred by monitoring heparin therapy with anti-Xa activity instead of aPTT. The study was a prospective, randomized, unmasked, cohort, single-center study in a 625-bed, private teaching hospital. Two hundred sixty-eight patients were included in the study and were treated with heparin on the basis of ideal weight, with a bolus of 75 U/kg followed by an initial infusion of 20 U/kg. Patients were randomized to be monitored by either the anti-Xa assay or aPTT. The researchers found that the group of patients being monitored by the anti-Xa assay had fewer monitoring tests and dose changes over 24 hours than did those in the aPTT group, a difference that neutralized much of the increased cost of the anti-Xa assay. 15 We can deduce that after accounting for labor costs, anti-Xa testing is at least a cost-neutral monitoring method.

Conclusions

Monitoring IV UFH therapy with the anti-Xa assay results in more expeditious achievement of therapeutic anticoagulation, longer maintenance of therapeutic levels, and fewer laboratory tests and heparin dosage changes compared with testing with the aPTT assay. We have reported our implementation process to give practitioners a clinical skill set that will guide them through the process of switching monitoring methods for UFH. We strongly believe that the conversion process is feasible across organizations of all sizes. Our results support the emerging notion that the anti-Xa assay is superior to aPTT testing for monitoring UFH. On the basis of the data, we believe the current paradigm of monitoring UFH with aPTT-based assays needs to be reexamined. However, we acknowledge that a prospective, randomized trial would be ideal to thoroughly address clinical questions such as length of stay and safety end points.

ACKNOWLEDGMENTS

We would like to thank the leadership team at VA Northeast Ohio Healthcare System for their unconditional support in implementing this hospital-wide protocol.