Management of Anticoagulated Patients in Dentoalveolar Surgery: A Retrospective Study Comparing Bridging Versus Unpaused Vitamin K Antagonist Medication.

Objectives: The aim of this study was to investigate the occurrence of postoperative bleeding following dentoalveolar surgery in patients with either continued vitamin K antagonist medication or perioperative bridging using heparin. Study design: A retrospective study was performed analyzing patients who underwent tooth extraction between 2012 and 2017. Patients were retrospectively allocated into two comparative groups: un-paused vitamin K antagonist medication versus bridging using heparin. A healthy, non-anticoagulated cohort with equivalent surgery served as a control group. Main outcome measures were: the occurrence and frequency of postoperative bleeding, the number of removed teeth, the surgical technique of tooth removal (extraction/ osteotomy/ combined extraction and osteotomy) and the prothrombin time. Results: In total, 475 patients were included in the study with 170 patients in the group of un-paused vitamin K antagonist medication VG, 135 patients in the Bridging group BG and 170 patients in the control group CG. Postoperative bleeding was signicant: CG vs.VG p=0.004; CG vs. BG p<0.001, BG vs.VG p<0.001. A signicant correlation of number of the extracted teeth in the BG (p=0.014) and no signicance in VG (p=0.298) and CG (p=0.210) and in the BG vs. VG and CG with p<0.001 in terms of surgical intervention extraction. No difference observed in terms of prothrombin time. Conclusion: Bridging increases the risk for bleeding compared to un-paused vitamin K antagonist medication. The perioperative management of anticoagulated patients requires a well- coordinated interdisciplinary teamwork to minimize or at best avoid both: postoperative bleeding and thromboembolic incidences. words: out of 170 patients (12.9%) in the control group, in 44 out of 170 patients (25.9%) in the vitamin K antagonist group and in 65 out of 135 patients (48.1%) in the Bridging group. Comparing the groups statistically, signicant differences were found for the control group versus the Bridging group (p < 0.001) and the control group versus the vitamin K antagonist group (p = 0.004).


Introduction
Due to the increase in life expectancy and the rising number of patients with cardiovascular diseases, the number of anticoagulated patients continues to increase worldwide 1 . Although the therapeutic anticoagulation management of the underlying diseases (e.g., apoplexy, atrial brillation, coronary heart disease, peripheral arterial occlusive disease, heart valve replacement or thrombosis) is primarily an internal medicine issue 2 , blood-thinning medication plays also a major role in oral surgery. Dentoalveolar surgical procedures such as tooth extractions, tooth osteotomies or root tip resections are part of the everyday dental practice. For anticoagulated patients, they require a close interdisciplinary coordination between cardiologists, general practitioners and oral surgeons 3 . Inconsiderate perioperative disruption of the anticoagulative medication bears the risk of potentially lethal thromboembolic events for the patients.
In their review, Wahl et al. reported 22 embolic events after discontinuation or reduction of anticoagulation, 6 of which ended fatally, whereas no fatal consequences were observed after postoperative bleeding with existing anticoagulation 4 . Although fortunately not being life-threatening in the vast majority of cases, postoperative bleeding following oral surgery causes severe discomfort for the patients, who might be hospitalized and may face follow-up operations for hemostasis. Anticoagulation therapy can be carried out using various classes of drugs (e.g. vitamin K antagonists, heparin, direct oral anticoagulants), which all intervene differently in the coagulation cascade and have their assets and drawbacks. Despite the availability of substances, which are easier to handle (e.g. direct oral anticoagulants), the coumarin derivative and vitamin K antagonist Phenprocomoun (Marcumar®), is still widely used for the prevention of thromboembolic events in atrial brillation or following heart valve replacement or pulmonary embolism. There are widely discussed approaches to the management of vitamin K antagonists prior to oral surgery: suspension for several days 2, 5 , temporary Bridging with heparin 6, 7 , reducing the dosage without Bridging 8,9 or unchanged dosages and hemostasis by local hemostatic measures 2, 5 10, 11 . Heparin offers the advantage of a good controllability due to its short halflife 12 . However, it is not recommended to interrupt the heparin therapy for surgical interventions 13 . In any case, preoperative coagulation lab testing is essential to evaluate the patients' level of anticoagulation 14 .
When investigating bleeding risks in a cohort of 1884 patients who received a surgical intervention with an adjusted INR of > 2.0 and were either bridged with low-molecular-weight heparin or a placebo, Douketis et al. found the risk of bleeding was 1.3% in the placebo group and 3.2% in the experimental group 15 .
In general, available literature on this topic is heterogeneous. Whilst there is a relative consensus pro Bridging regarding major surgery, especially for small-to-medium-sized surgical interventions including oral surgery the recommendations differ considerably even in o cial guidelines 16,17 . It was therefore the aim of this study to analyze bleeding complications in a cohort of anticoagulated patients having oral surgery. A special focus was laid on the comparison of Bridging versus un-paused vitamin K antagonist medication.

Materials And Methods
A monocentric retrospective patient cohort analysis was performed in a German university dental clinic, including all patients with a permanent vitamin K antagonist medication, who had oral surgery done between 2012 and 2017 in the clinic or were transferred to the clinic for treatment following oral surgery.
As a rst step, patient search was conducted by screening the digital clinic documentation system (MCC®, Meierhofer AG, Munich, Germany) and the digital patient le (Soarian Clinicals®, Cerner Health Services, Erlangen, Germany) using the following key words: tooth extraktion, tooth osteotomy, surgical intervention, tooth, bleeding event, Marcumar®, Bridging, heparin, anticoagulation and thromboembolic event. As a next step, further selection of patients was carried out, by including only patients, who had an oral surgical intervention (tooth extraction, tooth extraction and osteotomy or osteotomies). Both, in-and outpatients were considered. Furthermore, all patients with hemorrhagic diatheses or blood-thinning medication other than vitamin K antagonists (e.g. direct oral anticoagulants or platelet aggregation inhibitors) were excluded.
Depending on whether the vitamin K antagonist medication was temporarily paused and substituted by heparin perioperatively (=Bridging) or continued without interruption, we retrospectively allocated the selected patients into two groups: a Bridging group named BG and a vitamin K antagonist group named VG. Additionally, a control group of healthy patients without any anticoagulants, who had equivalent oral surgery, was added as a control group (named CG).
For each patient the following data was acquired from the digital patient le: Patients with incomplete documentation of the above listed information were not considered for the study. The primary outcome of the study was the frequency of postoperative bleeding in each group. Secondary outcomes were the type of surgery (tooth extraction with or without osteotomy), the number of extracted teeth and the INR. Ethical approval was obtained from the local medical faculty ethics committee (registration No.192_19Bc).

Statistical analysis
Statistical analysis was performed using the statistical programming language R V3.6.1 (R Core Team (2019). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria). The non-parametric Kruskal-Wallis and Mann-Whitney U tests were used, as well as the Chi-square test, Exact-Fisher's exact test and Cochran-Mantel-Haenszel test. The level of signi cance was p<0.05 in all tests performed.

General patient data
As a result of the data analysis, a total of 475 patients were included in the study, distributed to the three groups as follows: Bridging group (BG: n = 135), vitamin K antagonist group (VG: n = 170) and control group (CG: n = 170). The overall mean age was 71.76 years. Mean age in the groups was as follows: Furthermore, bleeding occurred signi cantly more often in the Bridging group than in the vitamin K antagonist group (p < 0.001), as shown in Fig. 1. The average number of postoperative bleeding events was 0.15 in the control group, 0.74 in the Bridging group and 0.29 in the vitamin K antagonist group, the number of postoperative bleeding events in the groups is presented in Table 1. There were signi cantly less bleeding events in the control group compared to the groups of anticoagulated patients (CG vs. BG: p < 0.001; CG vs. VG: p = 0.002). The comparison of the Bridging group and the vitamin K group revealed a higher number of bleeding events in the Bridging group (p < 0.001). Table 1 Showing the number of postoperative bleeding events within the three groups. B0= no postoperative bleeding, B1=postoperative bleeding, B2,B3,B4= two, three and four postoperative bleeding events.
Control group with patients n=170, bridging group n=135, vitamin k antagonist group n=170. bleeding. The INR differed signi cantly between the groups of Bridging and vitamin K antagonist medication (p < 0.001), but it did not signi cantly differ between the groups of bleeding/non-bleeding patients (p > 0.05).

Discussion
Anticoagulated patients continue to pose a challenge in everyday clinical practice 18 . This is particularly true for surgical procedures, including oral surgery. On the one hand, discontinuation or Bridging of anticoagulation can lead to thromboembolic events with a potentially lethal outcome 4,19 . On the other hand, intra-and postoperative bleeding can be burdensome for the patient and may complicate surgery and wound healing. Nevertheless it can be controlled su ciently by local hemostatic measures in the majority of cases 20 . As a result, the question whether to perform perioperative Bridging or to continue vitamin K antagonist medication in oral surgery, is becoming an increasingly contentious issue. There seems to exist a vague consensus pro Bridging when it comes to major surgical procedures such as extensive oncological or reconstructive operations, but for small to moderate surgical procedures, opinions and study results about the perioperative anticoagulation management differ widely. Clemm et al. investigated bleeding complications of anticoagulated patients in dental implant surgery (implant insertion and augmentative procedures). Comparing different anticoagulative schemes, they found a bleeding risk of 12.5% in a Bridging group (low-molecular-weight heparin), 6.7% in the vitamin K antagonist group, 1.4% in a platelet aggregation inhibitor group and 0.6% in a control group 21 11.3%, which is quite similar to our result. In contrast, the bridging group, with an incidence of 0%, did not record a single event. However, the bridging group consisted of only 6 patients, and the vitamin-k-inhibitor group included 80 patients 23 .
Postoperative bleeding event, were also recorded in correlation with the surgical intervention (single tooth extraction, serial tooth extraction or osteotomy). Single tooth extraction within the vitamin K antagonist group resulted in a rate of postoperative bleeding events of 10.5%, a rate of serial extraction of 16.7% and a rate of osteotomy of 10%. In the control group with 603 procedures, they found 0% postoperative bleeding events in single tooth and serial extractions and 1.3% in osteotomies. These results correspond to those of another study where 214 patients, who underwent tooth extraction of one to ve teeth per procedure without a signi cant correlation 24 . In our patient population, the occurrence of postoperative bleeding signi cantly correlated with the number of teeth removed within the bridging group and in terms of the surgical intervention extraction but not within osteotomy or within the other groups (VG and CG). This does not correspond with the ndings of another author. Bleeding did not correlate with the extension of the surgical procedure 2 . One reason for the increased postoperative bleeding in correlation with the number of teeth removed in the BG within this study may be that external patients were also included. Currently bridging of vitamin k is still common practice beyond the university hospitals for tooth removal. Therefore, it was not possible to differentiate between the other in uencing factors (such as wound management or invasiveness during tooth extraction) that might be crucial in terms of bleeding, especially in the BG but also CG and VG. As these data could not be collected due to the retrospective design of the study. However, also the continuation of vitamin K antagonists still poses a challenge. This is because even in this case, inconsiderable secondary bleeding may occur, although it can be easily treated by local hemostatic measures 2,4,5,22 . Most medical specialists recommend adjusting or reducing the INR value without permanently leaving the therapeutic area 4 . The risk of a lethal thromboembolic event, which is 0.2% in the literature and should not be disregarded 4 .
There are shortcomings of this study that need to be discussed. First, the retrospective study design led to discrepancies between the groups in terms of group size and composition. The extent and type of the surgical procedure varied between the groups and since operations were performed by different surgeons, the surgical techniques varied to a certain extent.
Within the limitations of the current study, it can be concluded that postoperative bleeding events occur signi cantly more frequently in bridged patients than in patients with un-paused vitamin K antagonist medication. It therefore appears reasonable to continue vitamin K antagonist medication perioperatively for the investigated class of small-to-medium sized oral surgery cases. A close interdisciplinary collaboration between oral surgeons and other medicine specialists is essential to minimize perioperative risks for the patients.

Declarations
Ethical approval    Showing the postoperative bleeding events within the groups (CG, BG and VG) in correlation to the surgical intervention (extraction; extraction and osteotomy; osteotomy). With no signi cances for osteotomy and extraction/osteotomy but within extraction in the BG to the VG and CG with p<0.001.