Tutorial in oral antithrombotic therapy: Biology and dental implications

Objectives: Recent developments of new direct oral anticoagulants that target specific clotting factors necessitate understanding of coagulation biology. The objective of this tutorial is to offer dental professionals a review of coagulation mechanisms and the pharmacodynamics of the conventional and new oral anticoagulants. Also, we summarized the dental implications of the conventional and new anticoagulants. Method: We searched Medline using search terms “antithrombotic”, “antihemostasis” or “anticoagulation” and combined them with the search results of “dental”, “oral surgery” or “periodontal”. We restricted the results to “human” and “English”. Results: The early coagulation cascade, the new cell-based coagulation model, the pharmacokinetics and pharmacodynamics of conventional antithrombotics, and new oral anticoagulants were reviewed. The new direct factor Xa inhibitors and the direct thrombin inhibitor (s), called direct oral anticoagulants (DOAs) have rapid onset of action, fast elimination on cessation, and fewer drug-drug or drug-food interactions than warfarin. However, the lack of antidotes raises concerns that some dental procedures may trigger serious hemorrhagic events. Additionally, careful perioperative withdrawal and resumption protocols for the DOAs are reviewed, because DOAs’ blood levels are dependent on renal function. Also, various reversal strategies in the event of excessive bleedings are summarized. Perioperative management of dental patients taking new DOAs and conventional oral anticoagulants are also discussed. However, the perioperative strategies for DOAs are yet to be validated in randomized trials. Key words:Coagulation cascade, cell-based coagulation model, factor Xa inhibitors, direct thrombin inhibitors, prothrombin complex concentrates.


Introduction
The increasing elderly �o�ulation and long life-ex-�ectancy lead to a high �revalence of chronic illnesses including heart disease and stroke. (1) These diseases often require antithrombotic thera�y to �revent thromboembolic (TE) events. The indications for antithrombotic thera�y are to �revent TE events and stroke in: (I) Atrial fibrillation and other cardiac arrhythmias; (II) Venous thromboembolism (dee� vein thrombosis, �ulmonary embolism); (III) Acute coronary syndrome and myocardial infarction; (IV) Pulmonary hy�erten-IV) Pulmonary hy�erten-) Pulmonary hy�ertension; and (V) Cardiac valve disease and �rosthetic valve re�lacement. (2,3) Oral antithrombotic drugs can be divided into two categories: anti-�latelets and anticoagulants. Table 1 sum-marizes these categories. Acetylsalicylic acid (as�irin) is the most widely used anti�latelet agent and the most commonly �rescribed oral anticoagulant has been warfarin. Consequently, instructional articles automatically refer to oral anticoagulants as warfarin and its derivatives. (2,(4)(5)(6)(7) However, the coagulation conce�t has been modified into a new, cell-based hemostasis model and several new oral anticoagulants targeting specific clotting factors have been introduced in 2010 -2011. Only recently, two cursory reviews on these new direct oral anticoagulants (DOAs) have a��eared in the dental literature (8,9). The objectives of the �resent review are (1) to educate general dental �rofessionals about coagulation cascade and the �harmacology of new and old anticoagulants and (2) to suggest �eri-surgical manage-
Co-administration with NSAIDs increase bleeding.
Possible rebound effects on cessation. Di�yridamole is used in combination of as�irin, Di�yridamole is indicated only when other anti�latelets are not suitable. Vitamin K antagonists (VKA) ( To conduct this review, we searched PubMed with search terms "anti-�latelet", "antithrombotic", "anticoagulation", or "anti-hemostasis", �ublished between 1966-2012 and in a se�arate search, we used the search terms "dental" "oral surgery" or "�eriodontal" and merged two searches. We collected 113 dentistry-related references.
In the first section of this review, we reviewed the early coagulation cascade; in the second section, we introduced the new coagulation model; in the third section, we �resented the new direct oral anticoagulants; and in the fourth section, we discussed �erio�erative management strategy.

Concepts on early coagulation cascade
Hemostasis involves a multi�art �hysiological �rocess that limits blood loss at the site of an injury while maintaining normal blood flow elsewhere in the circulation. An early model of coagulation derived from in vitro ex�eriments and �resented in the mid-1960s (10,11) involved a series of biological ste�s via intrinsic and extrinsic �athways leading to a common �athway to activate factor X (f.X of intrinsic and extrinsic �athways is considered inde-�endent. (13) While this conce�t was a huge achievement in the understanding of molecular interactions regulating coagulation, it had several limitations in ex�laining the hemostatic �rocess in vivo. Thus, several new theories were �ro�osed: one that extrinsic and intrinsic �athways are actually working inter-de�endently has been well-received. (12) An exam�le of this collaboration of two �athways is that the f.VIIa/TF com�lex from the extrinsic �athway can activate f.IX in the intrinsic �athway as well as f.X in the common �athway. (14) These conce�ts generated a new �aradigm advocating that cellular com�onents are the key �layers in coagulation which is controlled by the ex�ression of a variety of cell features and �rotein rece�tors that localize the com�onents of coagulation on specific cell surfaces. Thus, a conce�t of coagulation evolved from a mechanism being controlled by kinetics of various �roteins to a new conce�t of cellular com�onents controlling the coagulation �rocess. (12) We now review the new, cell-based coagulation model.

Concepts on the new, cell-based coagulation model
The new model of coagulation is a well-coordinated series of cell activation, culminating in a blood clot. (12) Three stages �ro�osed in this new coagulation models are initiation, amplification (priming) and propagation. In the initiation stage, TF from fibroblasts (or other TF bearing cells such as monocytes or endothelial cells) binds the activated factor VII, (a circulating coagulation factor) and forms the TF/VIIa com�lex. TF/VIIa com�lex in turn activates factors IX and X. The activated factor X (Xa) forms a Xa / Va com�lex on the surface of fibroblasts and generates a sufficient amount of thrombin to induce �latelet activation. (15) Thus, coagulation now moves to a more efficient site, the negatively charged, �latelet surface. In the second stage, the amplification (or priming) �hase, the coagulation now has moved from tissue factor bearing fibroblasts to platelets that are much better suited for the coagulation with their negatively charged surfaces. Platelets adhere and accumulate cofactors on their surface and am�lify the coagulation. Platelets and coagulation factors are activated by thrombin, which in turn binds and cleaves �latelet �rotease-activated re-ce�tors (PAR1 and PAR4), triggering a com�lex signaling cascade. Furthermore, thrombin activates factors V, VIII, and XI also. (15) This is a crucial �oint many health care �rofessionals are concerned about, because multifaceted anticoagulation ability will com�licate the reversal �rocess if excessive hemorrhage occurs. (16) In the final propagation phase, the tenase (factor "ten" activator) and �rothrombinase (f.Xa / f.Va) com�lexes are assembled on the �latelet surface, and large-scale thrombin generation takes �lace. (12) Tenase can originate from either the extrinsic �athway or from the intrinsic �athway. Intrinsic tenase is made of f.IXa and f. VIIIa with Ca++ and extrinsic tenase consists of TF/f. VIIa with Ca++. Intrinsic tenase is 50 times more efficient than extrinsic tenase in f.Xa ex�ression. Additionally, �rothrombinase com�lex (f.Xa / f.Va) is �rotected from tissue factor inhibitors, for it does not contain TF and thus a �owerful �roduction of f.Xa and thrombin formation results. This leads to a large scale fibrin formation. (15) These �rogressive �hases are summarized in figure 2. Under normal conditions, tissue factors do not come into contact with circulating vascular clotting factors. However, in some instances such as �ercutaneous coronary intervention, hemodialysis, or cardiac valve re�lacement surgery, these two com�onents come into contact resulting in a clot and subsequent adverse outcomes, such as myocardial infarction or stroke. In various inflammatory states, TF expression can also be u�-regulated in monocytes and endothelial cells by bacterial antigens, inflammatory cytokines, (17) or tumor necrosis factor. (18) This ex�lains the hy�er-coagulative state accom�anying infection. Now that we have reviewed both old and new coagulation models, it will be easier to understand where the new antithrombotic drugs inhibit the coagulation mechanism.

New antithrombotic agents
3-1. New anti�latelet drugs Platelets �lay a central role in thromboembolic (TE) events and four �otential rece�tors have been recognized in �latelet activation. (19) The first pathway via the thromboxane A2 synthesis is through which the conventional anti�latelet agent, as-�irin, exerts its anti�latelet action. As�irin is a cyclooxygenase (COX)-1 inhibitor that irreversibly inhibits platelet activation. It is an efficient and inexpensive drug but cause mucosal irritation and may cause internal bleeding. Thus, in recent years, a grou� of new an-ti�latelet agents has been develo�ed. The second �athway includes thieno�yridines that include clo�idogrel (Plavix ® ), and prasugrel (Effient ® ) that irreversibly block adenosine di�hos�hate (ADP) activation of �latelets through the ADP rece�tor P2Y12. (20) A non-thieno�yridine P2Y12 rece�tor inhibitor, ticagrelor (Ticlid ® ) is a reversible anti-�latelet agent much more �otent and has faster onset of action than clo�idogrel. However, it accom�anies more adverse effects such as dys�nea, hematoma and excessive bleeding. (21) The third �athway is via von Willebrand factor (vWF) and �latelets' glyco�rotein (G�)-Ib rece�tor. When ru�ture of an atherosclerotic �laque ex�oses collagen and vWF, the A1 domain of vWF binds to the G�-Ib rece�tor of e465 �latelet. The fourth �athway for �latelet activation is via �rotease activator rece�tor-1 (PAR-1). If �latelet activation is blocked through any one of these �athways, �latelet membrane glyco�rotein IIa/IIIb cannot bind to fibrinogen and unable to generate a hemostatic plug.  ity: numerous drug-drug and drug-food interactions, the genetic differences in VKA metabolism via cytochrome P450, and wide inter-individual differences in �harmacokinetics as well as the narrow thera�eutic window which requires frequent testing. Therefore, �atients' compliance directly affects its clinical efficacy. Warfarin stays in the thera�eutic range a��roximately 60% of the time, with an average of ~12% of the time in the su�ra-and ~20% of the time in the sub-thera�eutic range. (23) For these reasons, the im�etus for develo�ment of safer anticoagulants has been active resulting in several new direct oral anticoagulants (DOAs). These are direct factor Xa inhibitors rivaroxaban (Xarelto ® ), a�ixaban (Eliquis ® ), edoxaban and direct thrombin inhibitor dabigatran (Pradaxa ® ). (Tables 2,3) summarizes the �harmacokinetics �harmacodynamics, including dietary �otentiators or inhibitors of new DOAs as well as warfarin. New DOAs have ra�id onset of action, fast elimination on cessation and have fewer drug-drug or drug-food in-teractions. This obviates the need for frequent testing as with warfarin. However, no reversal agents are available at �resent and minimal trauma such as subgingival �eriodontal scaling, gingivectomy, or sim�le exodontia have �otential to become serious hemorrhagic events. We will discuss the new direct oral anticoagulants in detail.

3-2-a. Direct factor Xa inhibitors (Xaban)
Rivaroxaban is one of the direct factor Xa (f.Xa) inhibitors a��roved by FDA for the �revention of thromboembolic (TE) events in patients with atrial fibrillation. (24) The other drug in this grou� is a�ixaban for which FDA just a��roved its use in Jan. 2013. These drugs inhibit both free f.Xa and bound f.Xa within the �rothrombinase com�lex. Rivaroxaban reaches its �eak �lasma level at ~3 hours after administration and ~65% is excreted via urine and ~35% via biliary and fecal routes. Therefore, in the elderly or in �ersons with renal im�airment, rivaroxaban would have a �rolonged half-life.
• CYP: cytochrome P-450. • *CYP2C9*3 allele may be associated with retarded elimination of (S)-warfarin and may increase the clinical effects. • Interaction is a partial list of only highly probable interaction. Many probable, possible interactions are reported. In reference † • SSRIs: selective serotonin reuptake inhibitors References for Table 2 INR is sufficiently precise to measure the clinically effective serum level of DOAs. (28) Not as extensive as warfarin, some drug-drug interactions have been demonstrated with dabigatran: with quinidine, ketoconazole, amiodarone, and vera�amil that can increase the dabigatran �lasma concentrations. (26) Co-administration of other anticoagulants or nonsteroidal anti-inflammatory drugs will also increase dabigatran level. On the other hand, rifam�icin and some �roton �um� inhibitors such as ome�razole will decrease dabigatran levels.

Clinical implications of the new anticoagulants
Currently, there is no �ublished research evidence utilizing new DOAs in the dental setting. However, we summarized �erio�erative strategies that a��eared in medical journals so that dental �rofessionals will have adequate basic knowledge when they consult �hysicians regarding their �atients on new DOAs. In addition, active research in the develo�ment of the antidotes for new DOAs makes it im�erative for dental �rofessionals to have a clear understanding of coagulation mechanisms. 4-1. General dental considerations Scheduling early in the week and early morning of the day a��ointment may be necessary to afford additional visit in case of excessive bleeding. Furthermore, additional su��lies to deal with the �ost-surgical bleeding should be considered for �atients taking anticoagulants. (4,29,30) Some suggested additional su��lies are absorbable hemostatic dressings such as oxidised cellulose (Surgicel ® ), collagen s�onge (Haemocollagen ® ) or resorbable gelatin s�onge (S�ongostan ® ), along with resorbable and non-resorbable sutures. Also suturing the extraction sites and �ressure a��lication with gauzes saturated with tranexamic acid has also been suggested to �revent �ost-surgical bleeding. (31) Careful �re-o�erative assessment of �atients' medical profile was a common recommendations and some suggested routine laboratory studies including: a com-�lete blood count (CBC), �rothrombin time (PT), activated �artial thrombo�lastin time (aPTT), INR in case of VKA, and bleeding time. Arranging an after-hours res�onse team should be considered for those general dental �ractitioners (GDPs) without hos�ital �rivileges to counter �ost-o�erative bleeding. Most bleeding risk studies com�ared those who sto��ed warfarin to those who stayed on warfarin during the dental �rocedures and found that bleeding was not significantly different. However, if the bleeding incidence was com�ared to the controls not taking warfarin, the indirectly calculated relative bleeding risk was about a 10-fold increase in those who are on warfarin (9-10%) than those who do not take anticoagulants (1%). (32) In addition, some delayed bleedings were re�orted after 24 or 48 hours. (33) Therefore, �ost-surgical bleeding in �atients taking  anticoagulants is much greater than what is ex�ected in the general dentist's office (GDPs) and setting up an effective �ost-surgical res�onse �lan will be necessary. As to the risk factors for �ost-dental surgical bleeding, we �ostulated that (a) the severity of systemic morbidity requiring a high degree of anticoagulation, (b) co-morbidity such as renal insufficiency and hepatic dysfunction, and (c) the surgical trauma involved in dental �rocedures may all influence the bleeding risk. However, there is no clear definition of significant bleeding in the dental setting or clear guidelines for dental �rocedures with high bleeding risk. We reviewed the definition of significant bleeding by several authors and found Lockhart and colleague's definition is the most reasonable. (34) They defined significant bleeding as: 1) Continues beyond 12 hours; (However,) in our o�in-Continues beyond 12 hours; (However,) in our o�inion, even bleeding that continues to 5-6 hours should be considered significant.
2) Causes the �atient to call, return to the dental �racti-Causes the �atient to call, return to the dental �ractitioner, or visit the emergency de�artment; 3) Results in the develo�ment of a large hematoma or ecchymosis within the oral or adjacent soft tissues; 4) Requires a blood transfusion or other blood �roducts. As for the determination of invasiveness of dental �rocedures, Muthukrishnan et al. surveyed general dental �ractitioners (GDPs) and found that 71% of the GDPs surveyed considered subgingival debridement as a high bleeding risk, 48% inferior alveolar block, and 32% subgingival restoration as high bleeding risks. (35) For exodontia, single and u� to 3 extractions were considered a minor bleeding risk. 4-2. Managing �atients on anticoagulants in dentalsurgical setting Recent American College of Chest Physicians' (ACCP) Evidence-Based Clinical Practice Guidelines reflects the controversy surrounding the withdrawal/continuation of vitamin K antagonists (VKAs) and modified the �ast recommendation: In dental �rocedures with low bleeding risk, VKAs may be continued with im-�lementation of local haemostatic measures or may be interru�ted for 2-3 days �rior to surgery. (28) However, most dental �rofessionals favor continuing VKAs when International Normalized Ratio (INR) is below 3.5. As Ball em�hasized (36) (1) consulting the �rescribing �hysician(s) to ascertain the �atient's medical status that required anticoagulation thera�y, (2) checking the latest INR and testing it again on the day of the dental �rocedure may be very im�ortant in �reventing thromboembolic (TE) events because VKA levels fluctuate with diet and �atient's com�liance. If INR > 3.5 with moderate bleeding risk, withdrawal of VKA with �hysician's a��roval for 2-3 days may be necessary. Also hemostatic measures described in section 4-1 should be em�loyed. In high bleeding risk �rocedures including �eriodontal surgery, surgical extraction, multi�le ex-tractions (> 3 teeth), osteo�lasty, and extensive head and neck surgery, a withdrawal of VKA in conjunction with subcutaneous or intravenous he�arin usage was recommended. (36) This protocol is illustrated in figure 4. The withdrawal of new DOAs �rotocol is more com�lex than VKA because no sim�le reversal agents are available. The �rotocol is governed by the half-life of the drug, invasiveness of the (dental) �rocedures, the �atients' co-morbidities, and TE risks. As the elimination half-life of dabigatran is 12-18 hours, if the bleeding risk is high, the last dose should be given 3 days �rior to the surgery (ski� for 2 days). (28) If moderate hemostasis is acce�table, the last dose should be given -2 days (ski��ing 1 day). In case of moderate renal im�airment, namely creatinine clearance is (CrCl) 30-50 mL/min, ski� 4 doses (ski� for 2 days) for low bleeding risk �rocedures and 6-8 doses (ski� 3-4 days) for high bleeding risk �rocedures. (28,37) For rivaroxaban with half-life of ~9-10 hours, the renal function is less critical in drug elimination. Thus, the last dose should be given on -2 days for the low bleeding risk and on -3 days for high bleeding risk �rocedures res�ectively. Another factor that should be considered is the individual �atients TE risk. If TE risk is high, bridging with he�arin(s) without overla��ing with DOAs is suggested. The �erio�erative protocols for DOAs are summarized in figure 5. For the anti�latelet agents, the �erio�erative �rotocol is similar to the other antithrombotics. If the bleeding risk of �rocedure is low, not sto��ing the anti�latelet is recommended. (37) Resum�tion of as�irin results in nearly immediate maximal drug effects. However clo�idogrel will take almost 7 days to reach the maximal effective level. Administering the initial loading dose may shorten this lag �eriod (37). 4-3. Reversal of direct oral anticoagulants Although new DOAs have the advantage of stable �harmacological action and less interaction with foods or other drugs, the most im�ortant disadvantage of these new drugs is the lack of specific reversal agents. In case of excessive bleeding, the knowledge of reversal strategy will em�ower dental �rofessionals. In case of warfarin-related major bleeding, vitamin K may be administered. However, it takes almost 24 hours to bring INR to normal range. Therefore, in urgent hemorrhage caused by VKA over-dose, nonspecific prohemostatic agents, such as �rothrombin com�lex concentrates (PCCs), are indicated. (38) The first line of defense in any bleeding is discontinuation of the causative medication and utilization of local measures such as �ressure with tranexamic acid-soaked gauzes, or other hemostatic �roducts. If DOA administration is recent, gastric lavage or activated charcoal can be effective. (38) If these measures fail, hemodialysis or administration of either 3-factor (II, IX, and X) or 4-factor (II, VII, IX, and X) PCC should be considered. e470 Also in the murine model, fresh frozen �lasma (FFP) was able to reduce hemorrhage. However, due to the inconvenience of thawing FFP, PCCs or recombinant factor VIIa (rFVIIa) is gaining �o�ularity over FFP. Ex vivo trial in humans also showed that activated four factor PCC reversed the effects of rivaroxaban and dabigatran. (39,40) In the U.S., unactivated 4-factor PCCs are not available but activated PCCs that contain similar concentrations of non-activated factors II, IX, and X and activated factor VII are available. PCC has been used in the dental setting successfully. (41) However, caution must be exercised in using PCC or rFVIIa because they carry thrombogenic risks.

Conclusion
Although the rate of thromboembolic events is low, (a�-�roximately 0.5%) in the dental setting (42,43), most dental �rofessionals favor continuing VKAs (warfarin) for dental surgery because the adverse outcomes of discontinuation are much more serious than bleeding risks. However, the emergence of new direct oral anticoagulants �oses some challenges due to the lack of sim�le reversal agent in the event of �ost-surgical bleeding. Future randomized trials utilizing new direct oral anticoagulants like the direct factor Xa inhibitors or direct thrombin inhibitors are warranted.