Anticoagulation-associated intracranial haemorrhage in the era of reversal agents

For patients on oral anticoagulation, there is an annual intracranial haemorrhage rate of 0.3–0.6%. The majority of these bleeds (46–86%) are intracerebral, followed by subdural (13–45%), with the remainder being subarachnoid – as the aetiology is typically different, subarachnoid haemorrhage is not covered in this discussion. NOACs have been demonstrated to be at least as effective as warfarin, but with a lower risk of bleeding – as a result, most current guidelines recommend NOACs over warfarin, yet a significant number of people continue to be prescribed the vitamin K antagonist (VKA).

Intracranial haemorrhage is typically caused by disease of cerebral vessels. Around 15% result from large vessels (aneurysms, arterio-venous malformations), while the majority result from small vessel disease typically as a result of the deposition of extracellular material such as lipids or beta-amyloid in vessel walls – this is unlikely to have any relation to anticoagulation itself. Haematoma expansion is the second component to the disease; the other factors that drive expansion are vessel-tissue pressure gradient and shear forces, which may cause rupture of adjacent diseased vessels. Rates of haematoma expansion are higher in patients taking VKAs than those who are not on any anticoagulants, although the data for NOACs are not currently clear. The recommendation in the case of intracranial haemorrhage is that reversal should be undertaken as soon as possible.

Coagulation testing to identify suitability for reversal is relatively straightforward in VKAs; INR is based on prothrombin time, and forms a consistent test for level of anticoagulation. Testing for level of anticoagulation in NOAC-related bleeding is much more difficult. Dabigatran anticoagulation can be tested via measuring aPTT (adjusted prothrombin time) – this is the most sensitive widely available metric, and a normal result excludes clinically-relevant dabigatran concentration. More subtle and useful measures are available; diluted thrombin time and ecarin clot time or ecarin chromogenic assay can give a much better picture of the current status, but are not widely available. The same is true for the calibrated anti-factor Xa assays, which can be useful for measuring the action of rivaroxaban, apixaban or edoxaban.

The cornerstones of reversal of VKAs are vitamin K and prothrombin complex concentrate (PCC), with enough evidence now to suggest that fresh frozen plasma should not be used unless PCC is unavailable. Vitamin K should be given intravenously over the course of at least 20 minutes to avoid anaphylactic reaction, although full reversal often takes up to 12 hours. In terms of PCC, four-factor is preferred over three-factor products, while recombinant factor VIIa should not be used for reversal of VKAs. Reversal of dabigatran is best achieved via idarucizumab initially – a specific inhibitory antibody which binds to dabigatran with high affinity. After 24 hours, readministration of dabigatran provides effective anticoagulation. Andexanet alfa binds rivaroxaban, apixaban and edoxaban, and has undergone early trials although approval is pending and it is unlikely to be available until 2018. Ciraparantag (PER977) is in a similar situation, although it has yet to start phase III trials. Ciraparantag binds both dabigatran and the factor Xa inhibitors and has been shown to shorten clotting time in a dose-dependent way. PCC is non-specific in its action with regards to reversal of NOAC drugs and has demonstrated benefit both in laboratory metrics and in punch-biopsy related bleeding times. Although it lacks hard evidence, its approved status means it is the treatment of choice for reversal of NOACs, excluding dabigatran. Activated PCC has limited evidence at present, and no clear reason for treatment choice over PCC, unless that has already failed – there is a theoretical risk of increasing thrombotic complications. There is minimal evidence for the use of recombinant factor VIIa in NOAC reversal, and it should not be used unless other options are unavailable or have failed.

Once it is determined that there is a requirement for reversal of anticoagulation, this should be done as soon as is possible – particularly with NOACs, as requirement for reversal may be harder to accurately quantify. Renal and hepatic function should be considered, as derangement may lead to more potent action.

Unanswered questions in the current literature include whether reversal actually improves clinical outcomes: one study suggested that there may be limited benefit only, while two others suggested outcomes were improved. The INCH trial demonstrated that more rapid INR normalisation with PCC led to less early haematoma expansion – although it was not appropriately powered to detect these differences definitively. Being able to run placebo-controlled trials in an ethically appropriate manner is likely to be impossible, while enrolment of significant numbers of patients is also difficult in this field. The authors mention that more information is particularly needed with regards to the safety and efficacy of idarucizumab, ciraparantag and andexanet alfa in the intracranial haemorrhage domain.

This is a useful review article which summarises recommendations for oral anticoagulation reversal in intracranial haemorrhage based on current evidence.