Anticoagulation Use prior to Common Dental Procedures: A Systematic Review

Currently, the number of patients on oral anticoagulation is increasing. There is a paucity of data regarding maintaining oral anticoagulation (especially novel oral anticoagulants) around the time of specific dental procedures. A dentist has three options: either to stop anticoagulation, to continue it, or to bridge with heparin. A systematic review of 10 clinical trials was conducted to address this issue. It was found that continuing anticoagulation during dental procedures did not increase the risk of bleeding in most trials. Although none of the studies reported a thromboembolic event after interruption of anticoagulation, the follow-up periods were short and inconsistent, and the heightened thromboembolic risk when stopping anticoagulation is well known in the literature. Heparin bridging was associated with an increased bleeding incidence. We recommend maintaining oral anticoagulation with vitamin K antagonists and novel oral anticoagulants for the vast majority of dental procedures along with the use of local hemostatic agents.


Introduction
e use of anticoagulation is increasing in the population, and it is almost a daily occurrence to have a patient presenting for a dental procedure on vitamin K antagonists (VKAs) or novel oral anticoagulants (NOACs). Before considering stopping oral anticoagulation periprocedurally, the physician must balance between the risk of thromboembolism and bleeding associated with that procedure [1].
In the case of a surgical procedure, three possibilities are available: first to maintain warfarin, second to interrupt it, and third to withhold it and to do heparin bridging before the procedure. Stopping warfarin before a procedure can be detrimental to the patient's health, increasing thromboembolism and mortality rates [2,3]. romboembolic events were seen in 0.7% to 1.1% in patients who stopped anticoagulation before an invasive procedure [1,4]. A survey showed that most German dentists tend to stop VKAs before dental procedures [5]. Also, dentists registered in Michigan had nonuniform approaches towards patients on warfarin [6].
Concerning NOACs, a four-year cross-sectional study showed no significant bleeding when continuing anticoagulation with dental procedures, regardless of the invasiveness of the procedure [7]. e analysis of the RE-LY trial revealed that no significant differences in bleeding and thromboembolic complications exist between dabigatran and warfarin [8]. Although dabigatran has no antidote, it has a short half-life. us, a quick reversal of anticoagulation is possible if needed [8]. In an analysis of the EINSTEIN studies, rivaroxaban, another NOAC, has caused less major hemorrhagic events than AVK/bridging therapy when treating deep venous thrombosis and pulmonary embolism [9].
e American College of Chest Physicians Evidence-Based Clinical Practice Guidelines 9th edition recommends "either to maintain VKAs along with an oral prohemostatic agent or to interrupt them a couple of days before minor dental procedures." A need for bridging was not mentioned [10]. e European Society of Cardiology in 2009 [11], along with the American Academy of Oral Medicine in 2016 [12], recommends, for the majority of outpatient dental procedures, continuing VKAs if the international normalized ratio (INR) is in the therapeutic range. Because there is not enough data available regarding NOACs, the American Dental Association suggests continuing anticoagulation for the vast majority of dental procedures unless the patient is at a very high risk of bleeding, when a physician referral might be appropriate before the procedure [13].
While maintaining anticoagulation with VKAs during dental interventions, the postoperative bleeding risk might be reduced by adopting local hemostatic measures. Many agents were found to be effective: tranexamic acid mouthwash [14,15] for 2 days [14], oxidized cellulose and sutures [16], gelatin sponge [17,18], fibrin adhesives [19], HemCon Dental Dressing [20][21][22], platelet-rich plasma gel [23], and Histoacryl glue [24]. However, some obstacles exist that limit the use of those agents, for example, the high cost of fibrin glue [15,16] and the complex technique of tranexamic acid usage [25]. On the other hand, a Serbian study showed that local pressure is sufficient for adequate hemostasis in most cases of teeth extraction if INR is less than or equal to 3 [26]. It is noteworthy that suturing is not always necessary and should be reserved for instances where local hemostasis fails or when there is extensive tissue damage [17].
Although the data on VKAs are quite extensive and knowing that the bleeding risk in patients on NOACs might be higher, we are attempting a review of the literature of both VKAs and NOACs in the setting of a dental procedure. Rather than dividing the dental procedures largely into mild, moderate, and high risk of bleeding, we will attempt the evaluation of the risk of bleeding periprocedurally with specific dental procedures.

Methods
We have performed a systematic review of the literature on PubMed regarding anticoagulation during dental procedures. e keywords used were as follows: anticoagulation, anti-coagulation, Vitamin K, bridging, dental, dentist, tooth, teeth, and oral. e range of the studies is from 1996 to 2016, with most of the studies being after 2000.
From each study, we collected the following data: the number of patients, age, indications for anticoagulant treatment, exclusion criteria, the regimen of anticoagulation, bridging used, the procedure done, local hemostatic agents used, preoperative INR, target INR before undergoing the procedure, thromboembolic outcome, maximum follow-up period, and bleeding characteristics.
Concerning the latter, every study had a unique tool to assess and quantify bleed. We reported the outcomes accordingly.
is review aims to suggest recommendations for every specific dental procedure when it comes to continuing or interrupting VKAs and NOACs.
For every procedure, we determined the risk of bleeding and the recommendations regarding VKAs and NOACs. We attempted to base our recommendations on the results of well-established randomized controlled trials (RCTs) and controlled clinical trials (CCTs). When data are lacking, we reported an expert's opinion. e dental procedures assessed were as follows: surgical teeth extraction, implant surgery, excision of cystic formations, biopsies, alveoloplasty, frenectomy, periodontal surgeries, and microsurgical endodontics (apicectomy).

Participant Characteristics.
e total number of participants was 1331; at least 457 of them had their anticoagulation uninterrupted during the procedure. Most studies consisted of two groups: the first had oral anticoagulation continued during the dental procedure, the other had it stopped a few days before, with or without bridging with heparin. Warfarin was the main oral anticoagulant used, although some studies had other VKAs and only one studied NOACs. e bulk of the studies practiced local hemostatic measures after the surgeries. e primary procedure studied throughout was dental extractions, with or without a raise of a mucoperiosteal flap.
e indications for anticoagulant treatment were multiple, and the follow-up period extended from 1 day to 1 month. Most studies had their target INR within the therapeutic range in the anticoagulant group and therefore their preoperative INR falling within that range. Patients at risk of bleeding were predominantly excluded, like those with liver disease, renal disease, and coagulation abnormalities and those on drugs that increase that risk (Table 1).

Study Outcomes.
Every study had its protocol to assess bleeding outcome. A statistically significant difference in bleeding among groups was only observed in 2 studies: the first showing increased bleeding when bridging with LMWH [32] the second showing an increase in mild bleeding in VKA group when compared with no anticoagulation [33]. Only 4 patients across the 10 studies were reported to need hospitalization due to bleeding. e number of teeth extracted was associated with an increased risk of bleeding in one study [32]. is relationship was not seen in two other trials [27,29]. ere was no association between INR levels and postoperative bleeding [17]. A thromboembolic event was not observed in any of the studies, even in patients who interrupted their anticoagulation.
All studies recommended oral anticoagulation to be continued if INR is in the therapeutic range or inferior to 3. When maintaining oral anticoagulation, some studies found local hemostasis helpful. Bridging with LMWH [32] or giving heparin with reduced VKA dose [30] was found to increase the risk of bleeding (Table 2).

Recommendations.
Most evidence exists for surgical teeth extraction (5 RCTs and 4 CCTs). Concerning the rest of the procedures, the studies are mostly CCTs. For periodontal surgeries and endodontic microsurgeries, no controlled trials are available yet. After being certain that the patient is not overly anticoagulated and the drugs are adjusted based on creatinine level, we do recommend continuing anticoagulation in the vast majority of patients along with the use of local hemostatic agents. Although thromboembolic events were not seen in the trials studied (probably due to the short follow-up periods), it is well established that interrupting anticoagulation increases thromboembolic risk; therefore, this should be avoided as much as possible (Table 3).

Discussion
Bleeding during dental procedures occurs mostly in patients that are overly anticoagulated. A simple procedure can turn into a nightmare if the patient is on an AVK and his INR is above 4, or if he is on a NOAC with renal dysfunction.
When an anticoagulated patient presents for a dental procedure, the dentist has three main options: to continue the same dose of oral anticoagulation with local hemostatic agents, to diminish the dose, or interrupt it altogether a few days before [33]. Our systematic review has revealed that the first option is the best in most procedures, with none of the 10 studies recommending the remaining two options since no statistically significant difference in postoperative bleeding existed between most groups continuing and interrupting oral anticoagulation. Other studies have also come to the same conclusion: if INR is reasonable and local hemostatic measures adopted, there is no adverse outcome for continuing oral anticoagulation in dental procedures [38][39][40][41][42]. We recommend that VKAs must be continued in all surgical procedures if INR is in the therapeutic range. As for NOACs, they must also be maintained in most procedures. Local hemostatic agents are mostly needed in both cases.
LMWH bridging has been deemed not necessary in dental procedures [17], or even harmful by increasing bleeding risk [32,43] without altering the INR level. It has been found that heparin and reduced acenocoumarol [30] increase bleeding risk as well after dental procedures. Also, trying to replace heparin bridging with oral vitamin K one day before the procedure was unsuccessful as vitamin K did not adequately correct INR [44].
Special measures were taken in most studies to diminish bleeding risk [17,27,28,[33][34][35], like reducing soft tissue and bone injuries and minimizing the need to raise a mucoperiosteal flap during the procedures. However, it must be noted that whether a mucoperiosteal flap raise was needed or not in dental extractions [17,27,28,32,34,35], the outcome remained in favor of maintaining oral anticoagulation. Also, in implant surgery, bleeding risk was not associated with the invasiveness of the surgery [33].
ere was no association between the number of teeth extracted and postoperative bleeding [27,29,30], except in one study [32]. In this particular study, the sample was relatively small, and the patients were their own control, unlike the other studies. As a matter of fact, bleeding mainly occurs where local inflammation is severe [18].
In contrast to previous studies [2,3] and in line with others [4,25], a short interruption of oral anticoagulation did not seem to increase the risk of thromboembolic events in the 10 trials. However, the follow-up period, extending from one day to one month, was relatively small, and the thromboembolic risk could not be fully assessed based on the trials.
In brief, there is an immense need for cooperation between physicians and dental surgeons [17,45]. Although they both admit lacking full knowledge concerning oral anticoagulation in dental surgeries, dentists and physicians tend to mutually criticize [46]. Multiple measures are proposed for better cooperation, like having common classes in schools and establishing guidelines together. If a physician referral is necessary prior to a dental procedure, the dental surgeon should inform the physician that major bleeding is less likely in most procedures and that most guidelines recommend the continuation of anticoagulation, since physicians tend to overestimate the risk of bleeding.
is review has many limitations. e methods of assessing bleed were not uniform across the studies, which make an accurate comparison of bleeding outcome challenging. All the studies had VKAs as their oral anticoagulants except one CCT which included NOACs. Moreover, there is a lack of RCTs dealing with procedures other than teeth extraction, which creates a gap in the literature for the     Author and year of publication Methods of assessing bleed      [17,27,28,30,32,34,35] or <3 [29,31] Local hemostatic agents were judged essential in most studies [17, 28- Expert opinion: continue anticoagulation with caution with local hemostatic measures * e corresponding RCTs or CCTs deal only with the unique procedure cited above. * * e corresponding RCTs or CCTs deal with multiple procedures, among which one has been cited. RCT � randomized controlled trial; CCT � controlled clinical trial; VKAs � vitamin K antagonists; NOACs � novel oral anticoagulants; OAT �oral anticoagulation therapy. variable.
ere is a need for RCTs for specific patient populations, as patients with atrial fibrillation, for example, may be more predisposed to have a thromboembolic event [47].

Conclusion
For the vast majority of dental procedures, VKAs and NOACs must be maintained. Local hemostatic agents should be applied, and special attention should be given to INR level and renal function. Stopping and reinitiating oral anticoagulation can be troublesome for both the physician and the patient with an increased risk of thromboembolic events, and the best approach is multidisciplinary.