Summary
Direct oral anticoagulants (DOAs) – inhibitors of thrombin or factor-Xa – are expected to replace vitamin K antagonists in most of their indications. Patients receiving long-term treatment with DOAs are likely to be exposed to elective or emergency surgery or invasive procedures. Owing to the present lack of experience in such conditions, we cannot make recommendations, but only propose perioperative management for optimal safety regarding the risk of bleeding and thrombosis. DOAs may increase surgical bleeding, they have no validated antagonists, they cannot be monitored by simple standardized laboratory assays and their pharmacokinetics vary significantly between patients. Although DOAs differ in many respects, the proposals in the perioperative setting need not be specific to each. For procedures with low haemorrhagic risk, a therapeutic window of 48 hours (last administration 24 hours before surgery, restart 24 hours after) is proposed. For procedures with medium or high haemorrhagic risk, we suggest stopping DOAs 5 days before surgery to ensure complete elimination in all patients. Treatment should be resumed only when the risk of bleeding has been controlled. In patients at high thrombotic risk (e.g. those in atrial fibrillation with a history of stroke), bridging with heparin (low molecular-weight heparin, or unfractionated heparin, if the former is contraindicated) is proposed. In an emergency, the procedure should be postponed for as long as possible (minimum 1–2 half-lives) and non-specific antihaemorrhagic agents, such as recombinant human activated factor VIIa or prothrombin complex concentrates should not be given for prophylactic reversal due to their uncertain benefit-risk.
Résumé
Les anticoagulants oraux directs (AOD), anti-IIa ou anti-Xa, sont destinés à remplacer les antagonistes de la vitamine K pour une majorité de leurs indications. Les patients traités par ces médicaments pourront bénéficier d’une chirurgie ou d’un acte invasif programmé ou urgent. L’absence d’expérience de cette situation ne permet pas d’émettre des recommandations, mais seulement des propositions pour la meilleure gestion possible vis-à-vis du double risque hémorragique et thrombotique. Les AOD sont à risque hémorragique en cas d’acte invasif, n’ont aucun agent de réversion validé, ne sont pas facilement mesurés par les tests de laboratoire et leur variabilité individuelle est importante. Bien qu’il existe des différences entre les AOD, elles n’imposent pas une prise en charge periopératoire particulière à chacun. Pour les actes programmés à risque hémorragique faible, il est proposé de réaliser une fenêtre thérapeutique de 48 h encadrant l’acte. Pour ceux à risque hémorragique modéré ou élevé, il est proposé d’interrompre le traitement à j5 pour assurer l’élimination complète du médicament chez tous les patients et de ne le reprendre que lorsque le risque hémorragique est contrôlé. Un relais par une héparine est justifié, si le risque thrombotique lié à l’indication initiale du traitement par l’AOD est élevé (patients en fibrillation atriale ayant un antécédent cardio-embolique). Dans les situations non programmées, l’acte invasif doit être retardé au maximum (1 à 2 demi-vies d’élimination du médicament si possible) et les agents hémostatiques non spécifiques (facteur VIIa recombinant, concentrés de complexe prothrombinique) ne doivent pas être utilisés à titre prophylactique.
Background
Direct oral anticoagulants (DOAs), inhibitors of thrombin or factor-Xa (and potentially inhibitors of factor IXa), are expected to replace vitamin K antagonists (VKAs) in the majority of their current indications, such as the long-term treatment and prevention of venous thromboembolism (VTE) and the prevention of thromboembolic events in patients with atrial fibrillation (AF). The target population for these different indications, once they are registered, could represent 1% of the population. These patients will be treated with a dose designated in the following text as the “curative dose”, summarized in Table 1 . This dose corresponds to a level of treatment defined by an international normalized ratio (INR) in the range 2–3 using the VKA treatment that served as the comparator in clinical trials.
| Target | Dose (mg) | T max (hours) | Elimination route and metabolism | Elimination t 1/2 (hours) * | Dialysable | |
|---|---|---|---|---|---|---|
| Dabigatran (Pradaxa ® ) | Thrombin inhibitor | 110/150 bid | 2 | Faecal 20%; renal 80% | 14–17 | Yes |
| Rivaroxaban (Xarelto ® ) | Factor-Xa inhibitor | 20 od | 2–4 | Faecal 65%; renal 33% | 7–13 | No |
| Apixaban (Eliquis ® ) | Factor-Xa inhibitor | 5/2.5 bid | 3–4 | Faecal 75%; renal 25% | 8–15 | No |
| Edoxaban | Factor-Xa inhibitor | 30/60 od | 1–2 | Mainly renal | 8–10 | ? |
* The intervals shown are not applicable to the entire patient population (see third section of Rationale).
A significant proportion of patients treated with these drugs will inevitably be exposed to various critical situations: accidental or voluntary overdose; spontaneous bleeding; trauma; or the need for scheduled or emergency surgery or invasive procedures. As part of a process of “risk management” and in anticipation of marketing authorization, which may be granted in the months or years to come for the various DOAs for the indications listed above, consultations were conducted within the Working Group on Perioperative Haemostasis (Groupe d’intérêt en hémostase périopératoire [GIHP]) and the French Study Group on Thrombosis and Haemostasis (Groupe d’études sur l’hémostase et la thrombose [GEHT]) concerning the perioperative management of patients on these new treatments. These proposals do not concern patients receiving DOAs for the prevention of VTE following major orthopaedic surgery, nor those who will receive them for other indications in the prevention of VTE currently covered by low molecular-weight heparins (LMWHs), which may represent a simple alternative that is relatively easy to control in at-risk situations. The group declined to issue proposals for the management of other critical situations (in particular, acute bleeding) pending the results of ongoing preclinical and clinical studies. The method used to develop these proposals was as follows. A first version based on the recommendations of the French Health Authority (Haute Autorité de santé [HAS]) for the perioperative management of patients treated with VKAs ( www.has-sante.fr ) was adapted for DOAs using the published pharmacological data collected mainly during the phase I and II studies of these drugs. Some of these are included in the “Rationale” section below. This text was then submitted for several rounds of critical analysis by the members of GIHP and the pharmacology group of the GEHT, until a consensus was reached.
Rationale
Treatment with a curative dose of a DOA is associated with a risk of bleeding DOAs were developed to be used without laboratory controls and without the need for dose adjustment, which could give them an important advantage in terms of convenience over VKAs. However, the incidence of spontaneous bleeding described as “major” during the trials was still significant compared with VKAs and remains, even at low doses, higher than that of placebo . Although the therapeutic range is wider and the stability of the anticoagulant effect better than that of VKAs, the risk of bleeding at the recommended dose should be taken into consideration in the event of surgery or invasive procedures.
Direct oral anticoagulants (DOAs) have no effective antidote
Unlike VKAs, which can be antagonized by vitamin K and by prothrombin complex concentrates (PCCs) that provide the vitamin K dependent factors II, VII, IX and X, the new DOAs have no specific antidote. An antagonist that acts as a decoy with regard to factor-Xa (recombinant non-carboxylated factor-Xa without clotting activity) is being tested ex vivo . This strategy is not applicable to inhibitors of thrombin, as thrombin is not γ-carboxylated, but other approaches can be imagined. For example, a humanized monoclonal antibody specific to dabigatran was recently proposed as a specific antagonist . Even if their clinical efficacy can be demonstrated, these agents will not be marketed for several years. Thus, the currently available non-specific reversing agents potentially include: non-activated (Kanokad ® , Octaplex ® ) or activated (FEIBA ® ) PCCs and activated recombinant human factor VIIa (r-fVIIa, NovoSeven ® ). These were proposed on the basis of an extrapolation of experimental and clinical data on the reversal of fondaparinux or idraparinux (parenteral antithrombin-dependent inhibitors of factor-Xa) or melagatran (the first direct oral thrombin inhibitor, now withdrawn from the market). However, these data are fragmentary. The effectiveness of these drugs is controversial and their safety in terms of thrombotic risk in the target population has not been established. The arterial risk of r-fVIIa and, potentially, that of the other antagonists that have not been studied as thoroughly, is a concern among vascular and often elderly patients. A venous risk has not been shown in clinical trials, but these excluded patients with a history of VTE, for whom a DOA could be prescribed. Finally, the modalities of these treatments in this indication (dose, frequency of administration, etc.) are not known. In practice, the safety of a scheduled surgical or invasive procedure is based primarily on the natural clearance of the anticoagulant after preoperative suspension of treatment. PCCs, r-FVIIa and possibly, in the case of failure of these, plasma purification techniques (dialysis or haemofiltration for DOAs such as dabigatran that bind only weakly to plasma proteins and plasmapheresis for other DOAs) could, in exceptional circumstances, provide an emergency solution when faced with refractory severe bleeding.
The interindividual pharmacokinetic variability of direct oral anticoagulants (DOAs) is large
The four drugs that are most advanced in their clinical development and regulatory proceedings (dabigatran, rivaroxaban, apixaban and edoxaban) differ in their rate and extent of digestive adsorption and in the mechanisms and rate of elimination, predominantly renal or hepatobiliary. The values reported in Table 1 for patients are similar to those reported for healthy volunteers . As the blood concentration of the drug decreases exponentially beyond T max , it is expected that at the suspension of treatment, about 10% of the value of C max (theoretically 12.5%) is reached after 3 half-lives. It is interesting to recall that in a surgical context, the risk of bleeding appears to be more closely related to the value of C min than to that of C max . However, the absorption, distribution, metabolism and elimination of these drugs are governed by many variables: liver and/or renal function, sex, weight, age, genetic polymorphisms of enzyme or efflux systems, drug-drug interference, etc. It follows that for a given patient, the residual concentration at a given time after halting treatment cannot be accurately calculated from the average values of the pharmacokinetics in the target population.
Thus, at steady state (i.e. 3 to 5 days on average after starting treatment), the C max and C min of dabigatran at a dose of 150 mg twice daily had coefficients of variation of about 70–80% in the BISTRO I study and a high dispersion of C min in the PETRO study . Similarly, the C min and C max following 20 mg rivaroxaban once daily in phase II trials on hip replacement surgery or 30/60 mg edoxaban in patients with AF , were subject to significant variability. Finally, in daily clinical practice, for the doses approved in orthopaedic surgery, the equilibrium C max measured beyond the fifth day is assigned coefficients of variation of the order of 81% and 67% for dabigatran and rivaroxaban, respectively .
Direct oral anticoagulants (DOAs) cannot be monitored by a simple standardized laboratory test, accessible to all non-specialized laboratories
At pharmacological doses, these drugs have an effect on routine coagulation assays (prothrombin time [PT], activated partial thromboplastin time [aPTT]). The sensitivity of these assays varies with the reagents used . This is particularly the case for PT, whether its results are expressed as a PT ratio or an INR. The INR performed in routine practice is standardized to measure the effect of VKAs, but this is not so for DOAs . Furthermore, for a given concentration, the results for PT and aPTT vary between individuals as a function of factors that have not yet been identified (G. Freyburger, personal communication). It is important to remember that the therapeutic ranges and safety thresholds of INR validated for VKAs do not apply to the new anticoagulants. For example, an INR of 1.5 corresponds to a low risk of surgical bleeding for a patient who has stopped VKA treatment. The same INR value corresponds, in vitro, to plasma overloaded with 0.2 μg/mL of rivaroxaban , a concentration in the range of the C max observed after oral administration of 20 mg. To avoid any confusion, the expression of results as an INR, even standardized, should not be done. To interpret the results of a coagulation test prescribed to monitor a patient’s haemostasis (e.g. in the case of postoperative bleeding), the interference of DOAs with routine coagulation assays should be known. This has been reported in the course of preventive treatment with dabigatran and rivaroxaban in major orthopaedic surgery , but is not published for patients treated with a curative dose. Based on phase II trials, INR values varying widely around 1.5 (expression as an INR unsuitable for reasons indicated above) and aPTT patient/control ratios between 1.5 and 2 are expected at C max at equilibrium. This interference should disappear after 2 half-lives, but as indicated below, the pharmacokinetic variability is large and the interference may persist beyond 24 hours. Finally, aPTT prolongation is characterized by a very pronounced plateau effect at high DOA concentrations in vitro, making this test unsuitable for the detection of overdoses.
Dedicated assays derived from thrombin clotting time or ecarin clotting time or factor-Xa inhibitor chromogenic testing will soon be available to measure DOA plasma concentrations using appropriate calibration. It is not clear, however, whether they will be available in the near future to non-specialized laboratories, which is a prerequisite if they are to be used in contexts of relative urgency.
The security objective that must be reached for DOAs in the context of surgery or an invasive procedure has not yet been defined.
The interpretation and use of laboratory test results, whatever they may be, implies that safety thresholds exist. As yet, no preclinical or clinical data allow us to define a minimum plasma concentration below which the risk of bleeding during surgery is not different from that of a non-treated subject.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
