key: cord-0780069-gs99lzgy
authors: Jakowenko, Nicholas; Nguyen, Steffany; Ruegger, Melanie; Dinh, Ashley; Salazar, Eric; Donahue, Kevin R.
title: Apixaban and rivaroxaban anti-Xa level utilization and associated bleeding events within an academic health system
date: 2020-09-06
journal: Thromb Res
DOI: 10.1016/j.thromres.2020.09.002
sha: 68edb38e367d4befd04d88bd8b4661ff03ef8d06
doc_id: 780069
cord_uid: gs99lzgy

BACKGROUND: Oral factor Xa inhibitors (FXaI) can be administered in fixed doses without the need for routine laboratory monitoring. Anti-Xa assays can estimate anticoagulant effect for specific FXaI's. The aim of this study was to characterize anti-Xa levels in patients taking apixaban or rivaroxaban with major bleeding events. METHODS: Apixaban and rivaroxaban anti-Xa assays ordered within our hospital system from May 2016 to September 2019 were evaluated. The primary outcome was major bleeding events defined by International Society of Thrombosis and Haemostasis criteria. Median anti-Xa levels for each FXaI were calculated for those with and without major bleeding, as well as those who did and did not receive reversal agents. RESULTS: A total of 606 anti-Xa levels were analyzed. There were 146 major bleeding events documented, with the most common site being intracranial (63%). Median anti-Xa levels in patients with and without major bleeding were similar, whereas those on apixaban therapy who received reversal agents typically had higher anti-Xa levels (73 ng/mL vs. 153 ng/mL, p = 0.0019). Factors significantly associated with increased odds of bleeding were an age > 80 years, inappropriately high dosing regimens, and modest anti-Xa levels (100–300 ng/mL) for rivaroxaban specifically. CONCLUSIONS: Older age and inappropriately high dosing regimens were associated with major bleeding in patients taking apixaban and rivaroxaban. Further investigation into the utility of anti-Xa levels for FXaI is warranted.

Oral factor-Xa inhibitors (FXaI) have become widely used for both venous thromboembolism (VTE) treatment and prophylaxis, as well as stroke prevention in non-valvular atrial fibrillation.

Several large-scale clinical trials have shown that edoxaban, apixaban, and rivaroxaban are as or more effective than warfarin with a more favorable bleeding profile. [1] [2] [3] [4] [5] Attributes owing to their widespread use include minimal drug-drug interactions (DDI), tolerability, and lack of routine laboratory monitoring requirements. 6 Liquid chromatography (LC)-mass spectrometry (MS) can accurately quantify FXaI plasma concentrations and is considered the gold standard of assessing FXaI plasma concentrations, 7 however not all laboratories have 24-hour access to perform this type of analysis. Commercially available anti-Xa chromogenic assays may be utilized for FXaI levels with appropriate drug calibrators, 8, 9 and have demonstrated good correlation between plasma FXaI concentration and anti-factor Xa activity across a range of plasma concentrations. 10 Some researchers have attempted to establish "on therapy" anti-Xa level ranges for apixaban and rivaroxaban in a pragmatic setting, 11 whereas other pharmacokinetic studies conducted in healthy volunteers suggest expected plasma levels based on FXaI type, dose, and time since last dose. 12, 13 Certain clinical scenarios such as breakthrough thrombosis, bleeding events, and hepatic and renal insufficiency may warrant assessment of physiologically relevant FXaI concentrations, [14] [15] [16] [17] [18] and there is some evidence correlating plasma concentrations of certain direct oral anticoagulants (DOAC) such as dabigatran and edoxaban with bleeding events. 19, 20 However, there is a paucity of data describing the relationship between FXaI anti-Xa levels and clinical outcomes in patients taking apixaban and rivaroxaban in a real-world setting. The objective of this study was to evaluate anti-Xa levels for apixaban and rivaroxaban and their association with major bleeding events.

This retrospective, observational cohort study was approved by the Houston Methodist Research Institute Institutional Review Board (IRB). All anti-Xa levels for apixaban and rivaroxaban ordered at Houston Methodist, a seven-hospital system, from May 2016 to September 2019 were evaluated regardless of whether there was a documented inpatient dose of apixaban or rivaroxaban. Of note, all anti-Xa level tests for FXaI were performed at one laboratory in the Texas Medical Center Hospital. Levels were excluded if any of the following factors were identified: unfractionated heparin (UFH) or low molecular weight heparin (LMWH) product administration within 24 hours prior to an anti-Xa level being drawn; the level was deemed to be ordered in error (i.e., incorrect assay ordered based on drug exposure or the patient was not taking a FXaI); or critical information was missing from the electronic health record (EHR). The EHR was also queried for patient demographic information, baseline laboratory values, and all FXaI administrations. Reasons for anti-Xa level ordering and clinical decisions regarding FXaI were chart reviewed by two investigators (NJ and AD). If patients had multiple levels obtained during the same encounter, only initial levels were included in final analyses.

The primary endpoint of this study was the prevalence of major bleeding events as defined by the International Society on Thrombosis and Haemostasis (ISTH) which included bleeding that was fatal, located in a critical area or organ (i.e., intracranial, intraspinal, retroperitoneal, intraocular, pericardial, or intramuscular), caused a drop in hemoglobin of ≥ 2 g/dL, or led to the transfusion of two or more units of blood. 21 Major bleeding events were reviewed and adjudicated by two of the authors (NJ and AD) was confirmed either by presence of pertinent laboratory or imaging findings, as well as provider documentation for each patient. Median anti-Xa levels were determined both in patients with and without a major bleeding event. Anti-Xa J o u r n a l P r e -p r o o f levels obtained at the time of, or within 24 hours after a bleeding event, were considered to be in the context of a major bleed and categorized as such. Secondary endpoints consisted of reversal agent use (four-factor prothrombin complex concentrate (4F-PCC) and andexanet alfa), reasons for obtaining an anti-Xa level, changes in FXaI therapy at hospital discharge, and all anti-Xa levels following a documented inpatient dose. Pre-specified subgroup analyses were planned in order to compare patients with and without a major bleeding event on the basis of age, body mass index (BMI), baseline creatinine clearance (estimated with the Cockcroft-Gault equation), presence of AKI as defined by the Kidney Disease Improving Global Outcomes (KDIGO) criteria 22 , use of renal replacement therapy (RRT), dose appropriateness based on approved package labeling 23, 24 , and concomitant use of antiplatelet drugs or inhibitors of FXaI metabolism.

Apixaban and rivaroxaban anti-Xa levels were determined using a calibrated STA® -Liquid Anti-Xa Assay that is calibrated with commercial apixaban and rivaroxaban calibrators, and are available 24 hours a day, seven days a week at our institution. In this assay, exogenous factor Xa is added to a patient's plasma sample which can then complex with any present FXaI. A substrate that is specific for factor-Xa is then added to the sample, and the factor-Xa-substrate complex then cleaves a chromophore to produce color, with the amount of color produced being inversely proportional to FXaI presence. Expected peak levels (2-4 hours post-dose) are provided in the result details as well. The lower limit of the reportable range of our assays for both apixaban and rivaroxaban are < 20 ng/mL, and the upper limit 451 ng/mL and 472 ng/mL for apixaban and rivaroxaban, respectively. Expected peak levels for apixaban are 16-108 ng/mL and 102-155 ng/mL following a 2.5 mg and 5 mg dose, respectively. Likewise, typically observed peak levels for rivaroxaban are 90-190 ng/mL and 180-340 ng/mL following a 10 mg J o u r n a l P r e -p r o o f and 20 mg dose, respectively. These ranges have been adopted from previously conducted pharmacokinetic studies. 13, 25 Statistical analysis Continuous data were reported as mean and standard deviation or median and interquartile range (IQR) as appropriate. Categorical data were compared with the chi-squared or Fisher's exact test. Differences between two groups were assessed with a student's t-test or Mann-Whitney U test depending on the data distribution, with a p-value of ≤ 0.05 indicating a statistically significant difference at the two-sided level. Binary logistic regression was performed in order to ascertain any factors that predisposed patients to increased odds of experiencing a major bleeding event. A backwards stepwise selection method was implemented utilizing the following variables: age > 80 years, weight < 60 kg, baseline CrCl < 50 mL/min, AKI within 48 hours of level, strong CYP3A4 or P-gp inhibitor use, antiplatelet use, inappropriately high dosing regimen, and anti-Xa level stratified into tertiles (<100, 100-300, and >300 ng/mL). Variables with a p-value of < 0.2 were included into the final models. Three models were constructed, one combining both apixaban and rivaroxaban, and two for each individual FXaI. Descriptive and comparative analyses were performed with GraphPad Prism version 6.0 (GraphPad Software, San Diego, CA), and binary logistic regression was carried out using STATA version 14.2 (StataCorp, College Station, TX).

During the study period, a total of 979 anti-Xa levels for apixaban and rivaroxaban were Patients who received 4F-PCC or andexanet alfa had statistically significant higher median anti-Xa levels than those who did not receive a reversal agent for apixaban (73 ng/mL vs. 153 ng/mL, p=0.0019), but not for rivaroxaban (76 ng/mL vs. 131 ng/mL, p=0.21). There were 12 patients who received 4F-PCC (n=8) and andexanet alfa (n=4) outside the context of major bleeding. Three patients who were given 4F-PCC also received fresh frozen plasma (FFP), and six patients received only FFP. In addition, two patients received aminocaproic acid for a major bleed.

Reasons for obtaining level and changes to therapy 

There were 143 anti-Xa levels following an inpatient dose of apixaban and 15 anti-Xa levels following a dose of rivaroxaban (figure 4). There were 9 patients taking apixaban who still had detectable (>20 ng/mL or greater than the lower limit of assay quantification) anti-Xa levels beyond 30 hours (not depicted), all of whom had AKI or were on hemodialysis, but none had a major bleeding event. Median (10-90 th percentile) anti-Xa levels observed 2-4 hours after a dose of apixaban were as follows: 2.5 mg, 120 ng/mL (76-238 ng/mL); 5 mg, 200 ng/mL (78->451 ng/mL). One rivaroxaban level was drawn within 2-4 hours after a 15 mg dose (448 ng/mL), and four levels following a 20 mg dose, ranging from 51 ng/mL to >472 ng/mL. None of the patients with detectable anti-Xa levels for rivaroxaban beyond four hours had a major bleeding event, although three patients had AKI.

Several studies have commented on reasons for obtaining anti-Xa levels for apixaban and rivaroxaban in clinical practice. 9, 17, 26, 27 To our knowledge, this is one of the largest cohorts of patients with anti-Xa levels for both apixaban and rivaroxaban available in the context of major bleeding at the time of a bleeding event. Several other reports have attempted to describe the association between anti-Xa levels and clinical outcomes in a real-world setting. One study examined 411 C-peak anti-Xa levels for patients taking apixaban, rivaroxaban, and dabigatran for atrial fibrillation. 28 In a multivariate regression analysis, higher C-peak plasma concentrations were significantly associated with development of bleeding events (OR=2.7, 95% CI=1.3-5.4). 28 Likewise, Japanese patients taking rivaroxaban for atrial fibrillation who experienced a major (n=3) or non-major (n=19) bleed had higher peak anti-Xa levels (2.40 IU/mL vs. 1.85 IU/mL, p=0.001) and lower baseline creatinine clearance (46.2 mL/min vs. 58.8 mL/min, p=0.024). 29 Within our cohort, subgroup analyses and logistic regression models showed that advanced age was associated with major bleeding, which is in alignment with findings from other analyses of ROCKET-AF and ARISTOTLE, 30,31 but did not demonstrate that lower baseline renal function, AKI, or antiplatelet use were associated with increased odds of major bleeding. Given that the timing of last dose was not always known, this may influence the sensitivity of these analyses specific to renal function. Moreover, increasing anti-Xa levels in our study were not necessarily associated with an increased risk of odds for major bleeding, with the exception of modest anti-Xa levels (100-300 ng/mL) for rivaroxaban. This finding also mirrors the anti-Xa rivaroxaban levels seen in patients with acute major bleeding events who received andexanet alfa in the ANNEXA-4 trial. 32 In addition, we identified that patients with inappropriately high-dose regimens of FXaI had increased odds of experiencing major bleeding.

This finding is not surprising, and remains in line with evidence-based recommendations to dose adjust based on patient-specific factors for certain indications when clinically indicated. 33 

In the ANEXXA-4 trial, the median anti-Xa levels for patients taking apixaban and rivaroxaban were 149.7 ng/mL and 211.8 ng/mL, respectively. 32 Our study found similar median FXaI levels, many of which may be considered as abnormally elevated given that our ranges are based on levels seen 2-4 hours after a dose and many patients presented with an unknown last dose of FXaI. Interestingly, the investigators of ANNEXA-4 did not demonstrate a correlation between anti-Xa activity reduction and hemostatic efficacy with receiver operator characteristic (ROC) curve analyses. This is an important consideration, and although anti-Xa activity reduction may not be a strong predictor of clinical response to andexanet alfa, assessment of the extent of residual anticoagulation in patients treated with FXaI's may help guide the use of other reversal strategies. Furthermore, anti-Xa levels may be useful in guiding decisions for reversal of anticoagulation in emergent settings, especially in the absence of a complete history from a patient or when the last dose of drug is unknown and they have risk factors for drug accumulation. 9, 34 Given that andexanet alfa is not widely accessible across all institutions, it is vital to consider the role of other reversal strategies and agents including fresh frozen plasma, blood products, and 4F-PCC. [35] [36] [37] Some comparative data between andexanet alfa and 4F-PCC are emerging, 36 The most common reasons for obtaining an anti-Xa level in our study were similar with reports from other centers [14] [15] [16] 41 and included new bleeding, prior to a surgical procedure, and prior to heparin or fibrinolytic therapy. One study included 102 patients taking either apixaban or rivaroxaban examined anti-Xa levels prior to cardiac ablation. The investigators found detectable plasma DOAC concentration (>30 ng/mL) measured 24 hours after the last intake in 51.3% of patients. 42 In more severely ill patients, residual anticoagulation resulting from decreased drug clearance may have important implications, especially when transitioning patients to parenteral anticoagulants. 43 Anti-Xa levels were frequently utilized to assess for residual anticoagulation or potential drug accumulation in our study, and we similarly observed some patients with detectable anti-Xa levels beyond 24 hours after a dose, most often with concomitant renal impairment. The addition of strong inhibitors of FXaI metabolism may warrant the checking of anti-Xa levels to assess drug exposure. This was evidenced in a study including 12 hospitalized patients being treated for novel coronavirus infection COVID-19, where lopinavir/ritonavir administration resulted in an average 6.14-fold increase in trough plasma concentrations of apixaban and rivaroxaban. 44 Our analysis did not show an association of potential drug interactions and adverse events, however overall exposure for unique drugs was low and impact of different dosing regimens was not assessed. Further investigation is warranted into exploring the impact of drug interactions on adverse events on a larger scale.

Currently no guidelines endorse changes in therapy in response to anti-Xa levels for apixaban and rivaroxaban, however there were instances in our cohort where changes to FXaI therapy were in part due to an anti-Xa level in addition to the overall clinical context. Our results for all anti-Xa levels following a documented dose of FXaI underscores the importance that peak levels have a large inter-patient variability, which may be partly due to differences in anti-Xa assays and reagents used across different laboratories. Despite the fact that the expected level ranges currently utilized at our institution were derived from pharmacokinetic studies in healthy J o u r n a l P r e -p r o o f volunteers, growing evidence illustrates that in a real-world setting patient exhibit a wide range of levels after a given dose. For example, in a study of Korean patients, median peak levels for a 5 mg and 2.5 mg dose of apixaban were 202 ng/mL and 151 ng/mL, respectively. 18 The peak anti-Xa levels seen in our patients differed from those studies from which our expected level ranges were derived, highlighting the importance that hospitals may opt to develop and validate institution-specific ranges if anti-Xa levels are utilized on a consistent basis.

This study has several important limitations to consider, a major limitation being that anti-Xa testing was done in the clinical context of the patient, and not by randomization. The decision to assess anti-Xa levels for specific FXaI was therefore patient-and provider-specific, and it is possible that routine anti-Xa testing in unselected patients may yield different results. Second, the inherent limitations of a retrospective cohort study cannot be ignored, and although outcomes and changes to therapy were carefully chart reviewed in an objective manner, there may have been other contributing factors that were not captured. Third, certain risk factors that may predispose patients to develop bleeding (i.e., history of prior bleeding, concomitant comorbidities, and bleeding risk calculator tools such as HAS-BLED scores) were not included in our baseline demographic information. Therefore, it is uncertain how these risk factors may influence the odds of bleeding irrespective of corresponding anti-Xa levels within our patient population. Additionally, the majority of anti-Xa levels obtained in this study would be considered peak or random levels based on the time that they were drawn. Therefore, we cannot make any observations regarding trough levels which have been described in other studies and may display less inter-patient variability. Lastly, the reported upper limit of our institution's anti-Xa assay for apixaban and rivaroxaban is 451 ng/mL and 472 ng/mL, respectively. It is not typical practice at our institution to further dilute samples that exceed this upper limit, therefore the full extent of FXaI exposure may be underestimated in certain patients.

Journal Pre-proof

In conclusion, this study captured FXaI-specific anti-Xa levels in patients taking apixaban and rivaroxaban who experienced a major bleeding event. Both patients with and without major bleeding events exhibited levels that indicated circulating drug, however higher levels were not necessarily associated with increased odds of major bleeding within our cohort. We did identify advanced age, inappropriately high dosing regimens, and modest peak anti-Xa rivaroxaban levels as being associated with major bleeding. Management of selected patients with major bleeding events included the use of reversal agents such as 4F-PCC and andexanet alfa, and those who received a reversal agent tended to have higher anti-Xa levels than those who did not. The most common reason for obtaining anti-Xa levels seen in our cohort included confirmed or suspected bleeding, heparin or fibrinolysis clearance, and clearance for a procedure. While our study may highlight the utility of anti-Xa levels for apixaban and rivaroxaban in certain clinical situations including confirmed or suspected bleeding, clearance for heparin or fibrinolytic therapy, and prior to a surgical procedure, changes to FXaI therapy should always be made in the context of each patient and take into account other factors. J o u r n a l P r e -p r o o f A p ix a b a n 5 m g A p ix a b a n 2 .5 m g A 5 m g e x p e c te d r a n g e 2 .5 m g e x p e c te d r a n g e R iv a ro x a b a n 1 5 m g R iv a ro x a b a n 2 0 m g B 1 0 m g e x p e c te d r a n g e 2 0 m g e x p e c te d r a n g e

The peak anti-Xa levels in the above figure are following a documented dose of apixaban or rivaroxaban. The expected peak levels boxes refer to the expected anti-Xa levels seen 2-4 hours after a given dose of FXaI and are adopted from previously conducted studies.

 Quantification and assessment of anticoagulation can be accomplished with factor Xa inhibitor (FXaI) specific anti-Xa levels  Patients on oral FXaI who are elderly and have inappropriately high dosing regimens may be at an elevated risk for major bleeding  It remains to be determined if elevated anti-Xa levels for apixaban and rivaroxaban may confer a direct increased risk for bleeding in the absence of other factors  Anti-Xa levels may be useful in the inpatient setting for confirmed or suspected bleeding, clearance for parenteral anticoagulation, and prior to an emergent procedure J o u r n a l P r e -p r o o f Journal Pre-proof 

Oral apixaban for the treatment of acute venous thromboembolism

Oral rivaroxaban for symptomatic venous thromboembolism

Edoxaban versus warfarin in patients with atrial fibrillation

Apixaban versus warfarin in patients with atrial fibrillation

Rivaroxaban versus warfarin in nonvalvular atrial fibrillation

DOACs vs vitamin K antagonists: A comparison of phase III clinical trials and a prescriber support tool. Open Access Maced

Determination of dabigatran, rivaroxaban and apixaban by ultra-performance liquid chromatography -tandem mass spectrometry (UPLC-MS/MS) and coagulation assays for therapy monitoring of novel direct oral anticoagulants

Laboratory measurement of the non-vitamin K antagonist oral anticoagulants: Selecting the optimal assay based on drug, assay availability, and clinical indication

Anti-xa oral anticoagulant plasma concentration assay in real life: Rivaroxaban and apixaban quantification in emergency with LMWH calibrator

Laboratory assessment of the anticoagulant activity of direct oral anticoagulants: A systematic review

Evaluation of global laboratory methods and establishing on-therapy ranges for monitoring apixaban and rivaroxaban: Experience at a single institution

Practical guidance on the use of laboratory testing in the management of bleeding in patients receiving direct oral anticoagulants. Vasc Health Risk Manag

Safety, pharmacokinetics and pharmacodynamics of multiple oral doses of apixaban, a factor xa inhibitor, in healthy subjects

Assessment of direct oral anticoagulant assay use in clinical practice

Laboratory measurement of the direct oral anticoagulants: Indications and impact on management in clinical practice

Direct oral anticoagulant drug level testing in clinical practice: A single institution experience

Real-world monitoring of direct oral anticoagulants in clinic and hospitalization settings

Laboratory measurement of apixaban using anti-factor xa assays in acute ischemic stroke patients with non-valvular atrial fibrillation

The effect of dabigatran plasma concentrations and patient characteristics on the frequency of ischemic stroke and major bleeding in atrial fibrillation patients: The RE-LY trial (randomized evaluation of long-term anticoagulation therapy)

Association between edoxaban dose, concentration, anti-factor xa activity, and outcomes: An analysis of data from the randomised, double-blind ENGAGE AF-TIMI 48 trial

Subcommittee on Control of Anticoagulation of the Scientific and Standardization Committee of the International Society on Thrombosis and Haemostasis. Definition of major bleeding in clinical investigations of antihemostatic medicinal products in non-surgical patients

KDIGO clinical practice guidelines for acute kidney injury

Bristol-Myers Squibb Company

Population pharmacokinetics and pharmacodynamics of once-and twice-daily rivaroxaban for the prevention of venous thromboembolism in patients undergoing total hip replacement

Perioperative management of patients on direct oral anticoagulants

The danger of relying on the APTT and PT in patients on DOAC therapy, a potential patient safety issue

Drug levels and bleeding complications in atrial fibrillation patients treated with direct oral anticoagulants

Monitoring of anti-xa activity and factors related to bleeding events: A study in japanese patients with nonvalvular atrial fibrillation receiving rivaroxaban

Factors associated with major bleeding events: Insights from the ROCKET AF trial (rivaroxaban once-daily oral direct factor xa inhibition compared with vitamin K antagonism for prevention of stroke and embolism trial in atrial fibrillation)

Major bleeding in patients with atrial fibrillation receiving apixaban or warfarin: The ARISTOTLE trial (apixaban for reduction in stroke and other thromboembolic events in atrial fibrillation): Predictors, characteristics, and clinical outcomes

Full study report of andexanet alfa for bleeding associated with factor xa inhibitors

Appropriateness of DOAC prescribing before and during hospital admission and analysis of determinants for inappropriate prescribing

ACC expert consensus decision pathway on management of bleeding in patients on oral anticoagulants: A report of the american college of cardiology task force on expert consensus decision pathways

Neurocritical Care Society (NCS) Pharmacy Study Group. Factor xa inhibitor-related intracranial hemorrhage (FiX-ICH): Results from a multicenter

Evaluation of andexanet alfa and four-factor prothrombin complex concentrate (4F-PCC) for reversal of rivaroxaban-and apixabanassociated intracranial hemorrhages

Safety, efficacy, and cost of fourfactor prothrombin complex concentrate (4F-PCC) in patients with factor xa inhibitor-related bleeding: A retrospective study

Intravenous thrombolysis in patients with stroke taking rivaroxaban using drug specific plasma levels: Experience with a standard operation procedure in clinical practice

Residual anticoagulation activity in atrial fibrillation patients with temporary interrupted direct oral anticoagulants: Comparisons across 4 drugs

Direct oral to parenteral anticoagulants: strategies for inpatient transition

Direct oral anticoagulant plasma levels' striking increase in severe COVID-19 respiratory syndrome patients treated with antiviral agents: The cremona experience

The authors would like to thank Diane Dreucean and Yung Tran for their contributions.

This research did not receive any specific grant from funding agencies in the public, commercial, nor not-for-profit sectors.

J o u r n a l P r e -p r o o f J o u r n a l P r e -p r o o f