Measurement of heparin, direct oral anti-coagulants and other non-coumarin anti-coagulants and their effects on haemostasis assays: A British Society for Haematology Guideline

Abstract

This guideline was compiled according to the BSH process at [https://b-s-h.org.uk/media/16732/bsh-guidance-development-process-dec-5-18.pdf]. The Grading of Recommendations Assessment, Development and Evaluation (GRADE) nomenclature was used to evaluate levels of evidence and assess the strength of recommendations. The GRADE criteria can be found at http://www.gradeworkinggroup.org. Literature search criteria can be found in Appendix A.

Review of the manuscript was performed by the British Society for Haematology (BSH) Haemostasis and Thrombosis Task Force and the BSH Guidelines Committee. It was also placed on the members section of the BSH website for comment.

This guideline aims to update healthcare professionals working in the UK on the measurement of anti-coagulants (other than coumarins) currently licensed for use in the UK, and their effects on laboratory assays (Table 1). It provides recommendations based on the body of literature produced since the previous guidance published in 2014.¹ Direct factor (F)XIa- and direct FXIIa-inhibiting anti-coagulants are at various stages of development but not yet licensed, so are not discussed in this guideline.² The recent guidelines from the International Society of Thrombosis and Haemostasis Scientific Standardization Committee (ISTH/SSC) on the nomenclature to be used when describing non-vitamin K anti-coagulation³ are followed.

Some anti-coagulants, such as unfractionated heparin (UFH) and argatroban, have been in clinical use for decades. Laboratory monitoring to guide dose adjustments has been with the widely available activated partial thromboplastin time (APTT). However, the COVID-19 pandemic highlighted the wide variability in the sensitivity of different APTT reagents in patients with acute illness, emphasising the need for accessible and cost-effective anti-FIIa and anti-FXa assays for routine monitoring.

The introduction of specific anti-FIIa or anti-FXa-based assays has provided a means to quantitate plasma drug concentrations of the newer fixed-dose FIIa inhibitors (FIIaI) and FXa inhibitors (FXaI). When used according to licence, monitoring these direct oral anti-coagulants (DOACs) is not required, but measuring drug concentration can add value in some circumstances (Table 2). As neither drug concentration nor dose-adjustment based on the measured concentration has yet been shown to affect efficacy or safety, it is incorrect to refer to a therapeutic range. In this manuscript, the term ‘expected range’ is used to acknowledge this limitation.

There are many reports in the literature about the effects of anti-coagulants on measurable parameters of haemostasis. Lack of awareness of these effects, which are variable depending on anti-coagulant, timing of sampling and reagents, can cause confusion and delay diagnosis and care. Table 3 gives a broad overview of the types of impact on laboratory assays that may be seen.

All assays described must be used in accordance with requirements of ISO15189.

Activated charcoal products (ACP) (tablets or filters) that adsorb some anti-coagulant activities from plasma samples have been suggested as a way of undertaking haemostatic tests while continuing anti-coagulant therapy.^4-6 The process appears to remove not only the effects of rivaroxaban, apixaban edoxaban, dabigatran, argatroban, protamine, aprotinin and polymyxin but also leaves heparin-like and coumarin anti-coagulant activity intact.^{7, 8} Many haemostatic parameters have been reported to be unaffected by ACP, including those associated with haemophilia, thrombophilia (including lupus anti-coagulant assays) and thrombin generation assays (TGA), making the process attractive for many diagnostic algorithms.^9-11 ACP do not remove the effects of heparin-based or coumarin anti-coagulation, and caution is required if there are high levels of FXaI present, as these may be incompletely removed.^{11, 12} If laboratories wish to use this approach, the impact on local assays and reagents needs to be assessed in-house and documented in accordance with the requirements of ISO15189. In general, delaying testing until off anti-coagulation is the preferred option whenever possible.

Heparin is a naturally occurring glycosaminoglycan polymer that has a physiological anti-coagulant function. It exists naturally as polymers of varying sizes (20 000–50 000 Daltons), and all pharmaceutical-grade UFH is derived from porcine or bovine intestinal mucosa. Fractionation of primary polymers produces smaller molecules of varying sizes referred to as low-molecular-weight heparin (LMWH). UFH produces its anti-coagulant effect mainly through inactivation of FIIa and FXa (as well as FIXa, FXIa and FXIIa to a lesser extent) through an anti-thrombin-dependent mechanism.¹³ UFH is a highly negatively charged molecule with a propensity for reversibly binding to proteins and surfaces. Pharmacokinetic limitations are caused by anti-thrombin-independent binding of heparin to plasma proteins released from platelets and endothelial cells, resulting in a variable anti-coagulant response leading not just to large interindividual variability but also to intraindividual variability influenced by the patient's inflammatory response.¹⁴ One effect of this is to release tissue factor pathway inhibitor (TFPI), inhibiting thrombin generation in vivo, which is likely to contribute to the anti-coagulant action.¹⁵ Therefore, the anti-coagulant effect is not directly proportional to the dose of UFH, which requires routine monitoring to optimise the balance between the required anti-thrombotic effect and excessive bleeding risk.

Tests suitable for monitoring UFH are APTT, activated clotting time (ACT) and heparin anti-FXa activity assay. None of these assess all the anti-thrombotic effects of UFH and all have limitations. Furthermore, evidence to support the widely used expected ranges by either APTT or heparin anti-FXa assay is weak.

When using tri-sodium citrate blood collection tubes for UFH monitoring, there should be only a small residual air space in the tube once blood is added, achieved mainly with a predetermined vacuum.^{16, 17} Citrated samples containing UFH destined for APTT must be centrifuged within 1 h of collection and analysed within 4 h to avoid leakage of platelet factor-4 (PF4), leading to neutralisation of heparin.^18-20 Dextran sulphate releases heparin from its complex with PF4 so when samples are analysed by anti-FXa assay using dextran sulphate-containing reagents, centrifugation can be delayed for up to 4 h since there is only minor or no loss of heparin anti-FXa activity and little clinically relevant impact on management decisions.^{20, 21} Samples collected into citrate–theophylline–adenosine–dipyridamole (CTAD) are stable for at least 4 h, even when used for an APTT.^21-23

One major limitation of the APTT for monitoring UFH is the lack of specificity. For example, a lupus anti-coagulant or deficiency of one or more clotting factors, such as FXII, may prolong the APTT. This may falsely raise the APTT into the target range despite suboptimal heparin levels. Conversely, the APTT may not be within the target range even if the heparin level is at the correct therapeutic concentration in the presence of markedly elevated levels of coagulation factors such as FVIII and fibrinogen. These are frequently elevated due to the acute-phase response that is common in patients requiring UFH.²⁴ Similarly, acquired anti-thrombin deficiency can be seen in critical care patients, sometimes contributing to a higher-than-expected UFH dosage requirement.²⁵ Furthermore, FVIII may be increased in patients with acute thromboembolic events independently of the acute-phase response, and during pregnancy.²⁶ This makes the APTT less sensitive to UFH, leading to the incorrect assumption that heparin levels are inadequate.²⁷ The target APTT range for UFH for venous thromboembolism (VTE) was established in a prospective study in 1972 which included only 234 patients (approximately 2/3 VTE and 1/3 with arterial events).²⁸ Although there was a low risk of recurrent thrombosis when using an APTT ratio of 1.5–2.5 times a control APTT, the evidence to support this as a target range was weak. The APTT method utilised is no longer in use and is not traceable to current methods. The target range is therefore not immediately applicable now; since APTT reagents vary markedly in their responsiveness to UFH,^{29, 30} the reasons for using UFH for anti-coagulation have changed and the instrument used for analysis contributes to additional variability.³¹ The target APTT range²⁸ was later shown to correspond to 0.2–0.4 iu/mL heparin when measured using protamine titration assay or heparin anti-FXa of 0.3–0.7 iu/mL.³² Laboratory studies without assessment of clinical outcomes have shown that establishing a target range for any particular APTT method by reference to protamine titration assay^{29, 30} or heparin anti-FXa^{24, 27, 30} compensates for the variable response of APTT methods to heparin using samples from 30 to 60 patients. The limitations of the APTT for UFH monitoring are many, and even use of an APTT target range calibrated against heparin anti-FXa failed to improve interlaboratory consensus as to status of therapy (subtherapeutic, therapeutic or supratherapeutic) when three different APTT reagents were used in 44 UFH patients compared to uncalibrated APTT ranges.³³

Any patient whose APTT is being considered for UFH monitoring should have a baseline APTT performed prior to commencement of UFH therapy. If the pre-treatment APTT is prolonged or shortened, then the APTT is unsuitable for monitoring UFH therapy for that patient. In these cases, a heparin anti-FXa assay is a better option to monitor drug response.³⁴

UFH anti-FXa of 0.3–0.7 iu/mL is generally accepted as the UFH therapeutic range for treatment of VTE and other indications requiring a treatment dose, as opposed to prophylactic doses of anti-coagulation. This was derived from a single, small, randomised trial in which VTE patients requiring larger-than-average UFH doses (>35 000 units/day) were randomised to monitoring with APTT or UFH anti-FXa using therapeutic ranges corresponding to 0.2–0.4 U/mL by protamine titration.²⁷ The UFH anti-FXa range was 0.35–0.67 iu/mL (later rounded to 0.3–0.7 iu/mL) using an assay without dextran sulphate (Stachrom Heparin, Diagnostic Stago, France). The APTT group received higher mean daily doses of UFH than the UFH anti-FXa monitored group. Recurrent VTE within the first 12 weeks of therapy occurred in 3/65 and 4/66 in the heparin anti-FXa and APTT groups respectively. There were four bleeding events in the APTT group and one in the group monitored by UFH anti-FXa.

Calibration of the UFH anti-FXa assay should be with a UFH calibrator traceable to international standards. Commercial companies have developed combined (UFH/LMWH) calibrators to produce a single-calibrated anti-FXa assay. There is limited current peer-reviewed literature regarding comparability to separate curves, and these should be locally validated against separate curves if adopted.³⁵

Dextran sulphate is added to some reagents used for heparin anti-FXa assays to disrupt binding of UFH to several plasma proteins, which may occur in vivo or in vitro. Reagents containing dextran sulphate may give higher results than those without it, at least when some of the heparin is bound to proteins in the sample.^{35, 36} This can lead to overestimation of heparin anti-FXa activity in cardiac surgery patients after heparin reversal by protamine.³⁷ On the other hand, inclusion of dextran sulphate in anti-FXa reagents protects against underestimation of the heparin available in vivo because of in vitro binding of heparin to PF4. A recent study using samples constructed by spiking UFH into normal plasma suggested using dextran-free heparin anti-FXa assays provided blood collection is performed carefully and the first tube of blood is discarded to limit the amount of PF4 produced artificially.³⁸ Nevertheless, there is currently no consensus on whether heparin anti-FXa assays with or without dextran sulphate should be selected for monitoring of UFH. Although the presence or absence of dextran sulphate impacts results of heparin anti-FXa assays as discussed, in a UK NEQAS survey, the interlaboratory coefficient of variation (CV) for users of different heparin anti-FXa assays (some with and some without dextran sulphate) was approximately 10% for UFH samples in the therapeutic range compared to 15%–25% for APTT results determined with multiple reagents, making the heparin anti-FXa assay a more attractive option.³⁶ Furthermore, use of heparin anti-FXa assays to monitor UFH achieved a faster time to therapeutic range and fewer dose adjustments per 24-h period compared to use of APTT in two studies.^{39, 40} A retrospective cohort study of nearly 20 000 patients concluded that cases monitored by heparin anti-FXa were less likely to have a transfusion than hospitalised patients monitored by APTT after controlling for age, gender, other risk factors and invasive procedures.⁴¹ Despite these theoretical and reported advantages of a heparin anti-FXa assay over APTT, a meta-analysis of 10 studies with 6677 patients found that use of APTT compared to heparin anti-FXa was not associated with increased risk of bleeding (RR 1.03, 95% CI 0.8–1.22) or an increased risk of thrombotic events (RR 0.99, 95% CI 0.76–1.30).⁴² There were no differences in mortality in individual studies analysed, although the data were not considered suitable for pooled analysis. Recently, it has been reported that the heparin anti-Xa assay was the preferred method for monitoring UFH used to treat a pulmonary embolus (PE) in a study of 192 patients, identifying low incidence of recurrence or PE-associated mortality.⁴³ Overall, there are fewer disadvantages related to a heparin anti-FXa assay compared to APTT for monitoring UFH, and it is a better reflection of a patient's response to UFH. However, there is no strong evidence that clinical outcomes improve if heparin anti-FXa is used instead of APTT for monitoring UFH.

Lack of availability of the heparin anti-FXa assay over a 24-h period at some centres still limits the transition away from monitoring UFH with the APTT. Data provided by UK NEQAS (personal communication, December 2023) showed 540 UK sites registered for heparin monitoring using the APTT, with 125 (23%) registered for anti-FXa measurement of UFH. However, the heparin anti-FXa assay should not need to be performed as frequently as the APTT as it is less prone to interference: in stable patients on UFH, once-daily testing should be sufficient.

Since the last guideline in 2014, edoxaban has been added to rivaroxaban and apixaban as a licensed direct FXaI in the UK. In a similar timeframe, betrixaban was developed but was not granted a licence in the UK, and has been subsequently discontinued.¹⁰⁶ Given the wider therapeutic index of the FXaI, monitoring of their effects and dose titration is not routinely advised. However, due to uncertainty surrounding certain patient cohorts not represented in the trials, previous BSH guidelines made recommendations for testing in specific settings.

As part of the ENGAGE atrial fibrillation (AF)-TIMI trial comparing edoxaban to warfarin in AF patients, trough plasma concentration was measured 1 month into the trial in a subset of patients and the outcomes followed. The work conducted was based on probability analyses and suggested a relationship between trough edoxaban concentration and stroke/systemic embolism and bleeding events.¹⁰⁷

A pharmacokinetic analysis of 2392 patients enrolled in the Averroes study measured trough apixaban concentration after 3 months of treatment. The mean trough concentration for those patients receiving 2.5 mg bd was reportedly 99 ng/mL (IQR 60–146 ng/mL), while the 5 mg bd group was 125 ng/mL (IQR 64–202 ng/mL). The trial was not sufficiently powered to detect an association between apixaban concentration and outcomes due to low event rates. However, post hoc analysis suggests that patients in the lowest decile of levels had a significantly greater risk of stroke than those with higher levels.¹⁰⁸ An association between bleeding and apixaban level was also reported.

The availability of the anti-FXa assay is useful in some clinical situations (Table 2). There are limited real-world studies evaluating FIIaI and FXaI concentrations and outcomes, and these do not provide definitive evidence for the role of monitoring. The START laboratory registry comprised data on 565 consecutive patients with AF. Analysis of this showed a link between lower FIIaI and FXaI trough concentrations and thromboembolic events. Only 10 patients (1.8%) had a thromboembolic event, and these were associated with high CHA₂DS₂-VASc scores.¹⁰⁹ A study in Japan recruited consecutive patients with acute ischaemic stroke or transient ischaemic attack (TIA), started on rivaroxaban or apixaban for AF between 2012 and 2017.¹¹⁰ Peak and trough levels were measured after 48 h of treatment and patients were followed for a median of 360 days. Patients with bleeding events on rivaroxaban (13/156) had higher anti-FXa peak levels than those without, although levels were still within the expected range. Those with bleeding on apixaban (11/156) had higher trough and peak levels than those without bleeding. In another study of 212 patients on FIIaI and FXaI, levels were measured in the 83% with bleeding or thrombosis. Of these, 72% had concentrations in the expected range. Higher concentrations were seen in older patients, those with impaired renal function or lower body mass index.¹¹¹ The study concluded that although there was no clear benefit from FIIaI and FXaI measurements, this was useful in certain circumstances. The International Council for Standardisation in Haematology (ICSH) published guidance on the laboratory assessment of FIIaI and FXaI which included a table of expected peak and trough concentrations in AF and VTE patients,⁷⁰ reproduced in Table 4.

Variable prolongation of the PT and APTT can be seen with the different FXaI, thought to be associated with the composition of the reagents (activators and phospholipids) and analyser combination (Table 3). In general, FXaI have a greater impact on PT-based assays than on APTT assays; some PT reagents are insensitive to apixaban even at levels above the expected range. In the UK, many routine laboratories have similar reagent sources and although the relationship between PT and FXaI activity has been demonstrated as being mainly linear, monitoring using the PT and APTT is not recommended as values can still be within normal reference ranges even when FXaI can be detected by other techniques.^112-114

When FXaI anti-FXa activity is measured, it should be calibrated using a chromogenic method with drug-specific calibrators. The lower limit of quantitation (LLoQ) for measurement of FXaI by anti-FXa assay varies by reagent, analyser and drug, and should be verified locally. These standard assays are adequate for covering therapeutic fixed-dose ranges for all the FXaI and can be used as a guide when considering patient eligibility for andexanet reversal if the LLoQ is below 50 ng/mL. A lower LLoQ can be achieved if low-range protocols and low-range calibrator/control sets are used.¹¹⁵

Some reports have suggested caution when transitioning between FXaI and other anti-coagulants. For a period, there may be a cumulative effect. Under these circumstances, it has been recommended to test more frequently,¹¹⁶ although interpretation should be cautious, especially if transitioning between drugs with anti-Xa activity, as no assays can be used to distinguish between them.

Four-factor prothrombin complex concentrate has been the main agent in the UK for the reversal of major bleeding associated with FXaI despite not being licensed for this purpose. Laboratory manifestations of its use are likely to be correction of PT, APTT and global haemostasis assays (associated with its composition of factors II, VII, IX and X), although FXaI anti-FXa activity may not be affected.

The human recombinant FXa agent, andexanet alfa, is an option in the UK for reversal of apixaban and rivaroxaban (but not edoxaban) in life-threatening and uncontrolled gastro-intestinal bleeding.¹¹⁷ The Department of Health and Social Care published a Prevention of Future Deaths report in 2020 regarding the lack of an effective antidote for the reversal of edoxaban when bleeding occurs.¹¹⁸ Reports have shown that the FXaI-calibrated anti-FXa assays can be used as part of the screening process to identify eligible patients. However, FXaI activity can persist below the LLoQ of these assays, and results may not be available rapidly enough in an emergency. Furthermore, dissociated antidote may allow for FXaI anti-FXa activity to be overestimated, possibly leading to perceived failure of the reversal process.^{119, 120} Alternative small-molecule antidotes are in development, some of which are based on modified molecules to compete in a similar manner to andexanet. Others have a wider anti-FIIa and/or anti-FXa effect providing a global antidote.^121-124

FXaI can have an impact on specific factor assays whether using one-stage PT, one-stage APTT or chromogenic assays to varying extents. This can be ameliorated by performing assays at higher dilutions if non-parallelism is seen.^{138, 139} Thrombophilia testing is also affected (Table 3) with prolonged clot-based protein C (PC) and protein S assays potentially overestimating levels and artefactually indicating activated PC resistance. Prolongation of the DRVVT assays used for lupus anti-coagulant testing is likely even with low levels of FXaI. Anti-thrombin will be overestimated in assays based on FXa-inhibition.¹⁴⁰ Thrombin-induced platelet aggregation studies are also affected in the presence of all FXaI.^{141, 142} However, fibrinolysis appears more complex with conflicting reports as to the impact of FXaI (compared to the confirmed impact of FIIaI).^{143, 144}

Laboratory clinicians and scientists should have a clear understanding of the effects of anti-coagulants on their haemostasis tests so that they may properly advise clinicians who request tests and adjust doses of anti-coagulants. This will improve patient care and avoid unnecessary tests or interventions. For instance, normal coagulation tests in a patient taking a DOAC should not be interpreted as indicating a lack of anti-coagulant effect and prolongations may not need further investigation if correctly attributed. Adverse events include interactions with other drugs. For example, inhibitors of P-glycoprotein or the cytochrome P450 pathway may increase the concentration of DOACs, while inducers can have the opposite effect.

Some of the anti-coagulants discussed have relatively narrow therapeutic indices. As they are often used in acutely unwell patients, the potential for harm through under- or overdosing is high. Critical laboratory results are those that are life-threatening and require immediate action. The ICSH includes tests used for monitoring anti-coagulation among these.¹⁴⁵ Most adverse events associated with anti-coagulation are potentially preventable medication errors.¹⁴⁶ Inadequate monitoring or failure to act on a laboratory result is frequent cause of errors, including fatality, when using UFH. Complex dosing protocols and difficulties in interpreting results have been identified as causes of errors in investigations by NHS patient safety organisations (direct communication to BSH). When monitoring an anti-coagulant, it is essential that the correct anti-coagulant is identified in the request to the laboratory. This enables the correct standard to be used and the result reported in a way that makes it clear that it is specific to that drug. Unfortunately, requests received in the laboratory sometimes do not identify the drug correctly or give any indication that the patient is on an anti-coagulant. Electronic requesting systems reduce the chance of this occurring by making specification of the anti-coagulant a compulsory field. Laboratories vary in how over-anticoagulated samples are reported, causing difficulties for frontline healthcare staff moving between NHS organisations. As this is a consequence of varying reagent sensitivity, it is most effectively mitigated by improving local training of clinical and laboratory staff.

All authors contributed equally to the writing of this guideline.

Support for this manuscript was supplied by the British Society for Haematology.

The BSH paid the expenses incurred during the writing of this guidance.

All authors have made a declaration of interest to the BSH and Task Force Chairs which may be viewed on request.

While the advice and information in this guidance are believed to be true and accurate at the time of going to press, the authors, the BSH or the publishers accept any legal responsibility for the content of this guidance.

Members of the writing group will inform the writing group Chair if any new evidence becomes available that would alter the strength of the recommendations made in this document or render it obsolete. The document will be reviewed regularly by the relevant Task Force and the literature search will be re-run every 3 years to search systematically for any new evidence that may have been missed. The document will be archived and removed from the BSH current guidelines website if it becomes obsolete. If new recommendations are made, an addendum will be published on the BSH guidelines website (www.b-s-h.org.uk/guidelines).

Obtained for Table 4.