The management of patients receiving long-term oral anticoagulation who require dental extraction is based on an assessment of risks: the risk of procedure-related bleeding if anticoagulants are continued measured against the thromboembolic risk if anticoagulants are stopped. Before the introduction of the International Normalized Ratio (INR) and standardization of the anticoagulant range for most indications, studies suggested that there was increased bleeding if coumarins were continued in patients undergoing dental procedures . More recent studies assessing bleeding among anticoagulated patients as compared to patients who stopped warfarin or as compared to non-anticoagulated patients demonstrated no increase in bleeding among the anticoagulated patients [2–5]. There are currently no guidelines detailing anticoagulant management or optimal INR level in patients receiving warfarin who undergo dental extraction. Consensus guidelines from the American College of Chest Physicians provide a Grade 2B recommendation that anticoagulant therapy be continued and tranexamic acid or ε-aminocaproic acid mouthwash be used if there is concern regarding local bleeding . Despite this recommendation, anticoagulants are routinely discontinued; in a 1996 survey, 56% of physicians indicated that they discontinued anticoagulants routinely in patients undergoing a single dental extraction .
Using a cross-sectional survey design, we compared how oral and maxillofacial surgeons (OMS) and hematologists manage warfarin in patients requiring dental extraction, to assess perceived risks of thrombosis and bleeding between these two groups of clinicians and to examine the rationale for warfarin discontinuation despite the literature recommendations. The survey was mailed in November 2005 to practicing OMS and hematologists in Ontario, Canada. Survey recipients were identified using provincial membership registries (Royal College of Dental Surgeons of Ontario and the College of Physicians and Surgeons of Ontario). Membership in these associations is mandatory for all actively practicing OMS and hematologists. The survey consisted of two parts: (i) management of warfarin (how warfarin was discontinued and reinitiated); and (ii) assessment of perceived thrombotic and bleeding risks [using short descriptions of hypothetical patients with low, moderate and high thrombotic and bleeding risks, we asked respondents to rate the risk of thrombosis or bleeding using a visual analog scale from 1 to 10 (lowest to highest risk)]. The survey was pretested by a hematologist and general dentist to ensure face validity and accuracy. Data were collected on paper surveys and entered into an excel database. Because the respondents may not have answered all of the questions in the survey, proportions were calculated on the basis of the total number of respondents for any particular question. To compare the perceived thrombotic and bleeding risks, the mean score for each scenario was calculated for OMS and hematologists, and compared by unpaired t-tests using spss, version 12.0 (SPSS 15.0 for Windows; SPSS Inc., Chicago, IL, USA).
We mailed 291 surveys to 168 OMS and 123 hematologists in November 2005. In total, 137 (47%) surveys were returned; 82 (48.8%) eligible OMS and 55 (44.7%) eligible hematologists responded. The responding OMS were mostly community-based (75.6%) with a mean of 24 years (range: 1–48 years) in clinical practise. The hematologists were mainly from academic or combined academic and community practise (76.4%) with a mean of 13 years (range: 1–38 years) in clinical practise.
The warfarin management for the two groups is presented in Table 1. Warfarin is discontinued at least 50% of the time by 70% of hematologists for a mean of 4.4 days prior to dental extraction aiming for an INR no higher than 2.5, with 72% using bridging anticoagulation prior to the procedure. Warfarin is discontinued at least 50% of the time by 37% of the OMS for a mean of 3.2 days prior to the procedure aiming for an INR no higher than 3.0, with 41% using bridging anticoagulation. The mean visual analog scores for the thrombotic and bleeding scenarios are shown in the latter half of Table 1. In all thrombotic scenarios, with the exception of the highest thrombotic risk, OMS consistently rated thrombotic risk higher than did hematologists (P < 0.001). In contrast, the OMS rated bleeding risk lower than did hematologists, but this was only significant for patients at the highest bleeding risk. In scenarios combining both thrombotic and bleeding risks, warfarin discontinuation was approximately 50% higher among hematologists than among OMS; in a scenario combining high thrombotic risk and low bleeding risk, warfarin was still discontinued by approximately 19% and 35% of OMS and hematologists, respectively (data not shown).
|Frequency of discontinuing warfarin, n (%)|
|Never||15 (18.5)||1 (1.9)||< 0.0001|
|< 25%||22 (27.2)||6 (11.3)|
|25–50%||14 (17.3)||9 (17.0)|
|50–75%||9 (11.1)||7 (13.2)|
|> 75%||18 (22.2)||15 (28.3)|
|Always||3 (3.7)||15 (28.3)|
|Recommended time off warfarin (days)||3.2||4.4||< 0.0001|
|Highest acceptable INR||3.0||2.5||< 0.0001|
|Reinitiation of warfarin carried out within 24 h, n (%)||50 (66.7)||48 (90.6)||0.026|
|Recommend antifibrinolytic mouthwash, n (%)||31 (37.8)||33 (62.3)||0.010|
|Use bridging anticoagulation, n (%)||30 (40.5)||38 (71.7)||< 0.0001|
|Risk scenarios (visual analog score)*|
|Chronic atrial fibrillation without risk factors||3.42||2.48||0.006|
|Recurrent idiopathic pulmonary embolism||7.14||5.81||0.002|
|Mechanical mitral valve, stroke and hypertension||7.21||7.56||0.432|
|Single tooth extraction||3.16||3.07||0.824|
|Multiple (full mouth) extraction||6.90||7.22||0.447|
|Multiple impacted teeth||7.12||8.04||0.024|
In this cross-sectional survey of practicing OMS and hematologists in Ontario, Canada, we found that, despite the available data in the literature and consensus guidelines, warfarin is routinely discontinued by 70% of hematologists and 37% of OMS for dental extraction and surgical dental procedures. OMS perceive thrombotic risks to be higher than hematologists’ assessments of risk; this is reflected in their practise of minimizing duration off warfarin therapy and their willingness to perform extraction at a higher INR, using local measures to minimize bleeding. Hematologists frequently discontinue warfarin at approximately 50% higher frequencies than OMS, and minimize thrombotic risk by using bridging anticoagulant therapy, which refers to the use of short-acting anticoagulants (usually low molecular weight heparin) during the time that patients are not therapeutically anticoagulated with warfarin.
Several randomized and non-randomized studies support continuation of warfarin for dental extraction. Serious bleeding or bleeding that cannot be controlled using local measures is rare among patients who continue anticoagulation, estimated to comprise 0.4–2% of patients undergoing dental extraction [8,9]. Use of local measures, including sutures, cautery, laser coagulation, hemostatic agents (oxycellulose, absorbable gelatin, collagen, and fibrin glue), and antifibrinolytic mouthwashes [10–14], has been shown to decrease bleeding among anticoagulated patients. However, these studies were relatively underpowered to detect differences in bleeding due to their small size, and many relied on patient self-reporting as opposed to blinded, objective bleeding assessments. Despite these limitations, these data suggest that anticoagulated patients with INR values of approximately 2.0–3.0 at the time of dental extraction rarely experience serious bleeding and can be managed using local measures if necessary.
Consequently, the frequent practise of discontinuing warfarin may place patients at unnecessary risk of thromboembolic complications. Although the risk may be reduced by using bridging anticoagulation, this practise is associated with financial expense and inconvenience. The discrepancy in practise between OMS and hematologists in our survey may be related to a number of factors. OMS consider patient-specific thrombotic risk to be greater, and patient-specific bleeding to be less, than hematologists, and are more likely to use strategies that minimize thrombotic risk. Conversely, hematologists consider patient-specific bleeding risks to be higher than do OMS, which is reflected in their use of strategies that minimize bleeding. The hematologists’ behavior may also reflect a referral bias, as hematologists may be more likely to discontinue warfarin because patients have been specifically referred for warfarin discontinuation and there may be an expectation on the patient and referring physician’s behalf that this will occur. In a recent survey of US dental faculty responsible for teaching dental students about anticoagulation management, 46% of the faculty indicated they teach students to discuss warfarin discontinuation with physicians and patients for single simple extraction, and this increased to 64% for multiple extractions . Other factors may also contribute to this discrepancy in practise, including hematologists’ lack of awareness of this literature, as most studies are published in dental journals, a limited understanding of the nature of dental extractions, and extrapolation of their familiarity with using bridging anticoagulation for other surgical procedures to patients undergoing dental extraction.
In summary, a significant proportion of surveyed hematologists and OMS inappropriately discontinue warfarin for dental extraction, which may place patients at increased risk of thromboembolic complications. This may be due to increased perceived risks of thrombosis among OMS as compared to hematologists, and increased bleeding risks among hematologists as compared to OMS. Hematologists minimize thrombotic risks by using bridging therapy, which is probably unnecessary, is costly, and is inconvenient for patients. Increasing education and awareness among hematologies and OMS that dental extraction does not require routine anticoagulant discontinuation is a priority that may minimize patient risk.