Dear editor,

We appreciate the pertinent comment from Dr Tripodi on the proposals of the Working Group on Perioperative Haemostasis, regarding management of major bleeding and emergency surgery in patients on long-term treatment with direct anticoagulants . Several clarifications need to be made.

As highlighted by Dr Tripodi, recommendations cannot be established because of the lack of data in the emergency situation. Therefore, several mundane proposals can be put forward, such as: ‘ideally, surgery must be delayed’; but how many times, and for what purpose? In our working group, we decided that experts must develop an opinion precisely for such cases where there is no clear guidance. Moreover, we addressed the proposals for emergency surgery (i.e. surgery that must be done within 48 hours) that, in instances such as peritonitis, cannot be delayed.

If specific dosages are available for an anticoagulant drug, we fixed a threshold of < 30 ng/mL for safe emergency surgery. We agree with Dr Tripodi that the precise threshold is unknown. This value was derived from (very cautious) pharmacokinetic data, alongside data in guidelines used in clinical trials for elective surgery in patients receiving these anticoagulants. The modelling of this dataset led us to select a threshold of 3 ng/mL, but we accept that this value remains theoretical. However, if we propose use of such a specific dosage to help clinical decision-making in the case of surgery, we have to indicate targets for concentrations of these drugs.

In France – but we cannot assume the same will be true elsewhere – we know that all hospitals in charge of emergencies do not dispose 24/24 of such dosages; therefore, we propose an alternative, weaker strategy, based on available haemostatic tests (prothrombin time [PT] and activated partial thromboplastin time [aPTT]), effectively with a high level of uncertainty. We proposed that both a PT and an aPTT (and not PT OR aPTT, as suggested by Dr Tripodi) can be taken as a threshold limit. We agree that this proposal has severe limitations, but again, either all hospitalized teams use specific tests (which is actually not possible in France, from a cost perspective), or we decide to manage emergencies without having performed any test at all. Regardless, as discussed in our publication, we agree completely with Dr Tripodi that all laboratories must analyse the accuracy of their PT and aPTT measurements for determining the concentration of dabigatran or rivaroxaban.