This subanalysis from the EMIT-AF/VTE programme analysed dental procedures performed on patients receiving edoxaban in routine clinical practice. Of interest was whether edoxaban therapy was interrupted for dental procedures, the timing of interruption if it occurred, and the resulting clinical outcomes. The edoxaban management strategy for dental cleanings was generally to continue treatment throughout the procedure, whereas more invasive procedures were more frequently associated with edoxaban interruption. Most interruptions were short and consisted of skipping 1 or 2 doses before the procedure (i.e., preprocedural interruption). Overall, the safety results from this subanalysis demonstrate that dental procedures had a low risk of bleeding and that most bleeding was minor. The sole major bleeding event was recorded in the 29-day postprocedural observation window for a dental cleaning but was definitely not related, as it occurred during an ocular procedure performed 28 days after the cleaning.
The bleeding event rates that occurred within 7 days of dental procedures in this subanalysis were consistent with those observed in other observational studies using the typical 7-day postprocedure follow-up window [14,15,16]. However, clinical events in EMIT-AF/VTE were recorded up to 29 days postprocedure, which allowed for greater sensitivity to detect ischaemic/thrombotic events. Importantly, no thrombotic events were observed in this subanalysis, suggesting that the edoxaban management strategies adopted by dental professionals did not increase the risk of stroke for dental procedures, regardless of how or whether edoxaban was interrupted.
The DENTST study investigated the outcomes of 86 patients receiving NOACs (all NOACs except edoxaban; 145 total teeth extracted) and 21 patients receiving warfarin (50 total teeth extracted) [14]. In both treatment arms, no major bleeding events were reported during the 7-day follow-up period [14]. The rate of minor plus CRNM bleeding in the NOAC treatment group (36%) was similar to that observed in the warfarin control cohort (43%) [14]. The authors concluded that there was no need to adjust NOAC dosing prior to dental extraction [14].
Another prospective observational study with a postprocedural observation window of 7 days assessed outcomes of dental extractions for 119 patients receiving NOACs (n = 128 extractions), including edoxaban, and 248 patients receiving warfarin (n = 262 extractions) [15]. Postprocedural bleeding rates were low, with a trend towards fewer events with NOACs (4/128 [3.1%]) vs warfarin (23/262 [8.8%]); thus, the authors concluded it was not necessary to interrupt NOAC therapy for tooth extractions [15]. A retrospective cohort study with an observation window 30 min preprocedure to 7 days postprocedure reported similar results, demonstrating no differences in bleeding risk between NOACs and VKAs, despite a different definition of bleeding compared with other studies and the current subanalysis of the EMIT-AF/VTE programme [16]. Patients (N = 541) underwent 634 tooth extraction procedures (1196 teeth extracted) [16]. Seventy-two extractions (n = 41 procedures) involved NOACs (all 4 NOACs were represented in the study), n = 100 (50 procedures) involved VKAs, and n = 1024 extractions (543 procedures) involved no anticoagulants [16]. The incidences of postextraction bleeding (incidence per tooth) were 10.4% for NOACs, 12.0% for VKAs, and 0.9% for no anticoagulants [16]. Bleeding risk was not significantly different between NOAC and VKA groups (P = 0.49) [16].
Bajkin et al. reported postoperative bleeding in 3.3–5.7% of patients on any anticoagulant, which is consistent with the rate of bleeding observed with edoxaban use during dental procedures in this subanalysis of EMIT-AF/VTE [6]. Few examples of bleeding were observed during implant procedures, and these cases were well controlled with haemostatic measures [6]. The authors concluded that the accumulating data on management strategies strongly favour continuing anticoagulation before, during, and immediately after invasive dental procedures [6]. Collectively, the evidence in this review and other studies suggests similar or reduced bleeding risk during dental procedures with NOACs vs VKAs [6, 7, 14,15,16]. Notably, these studies generally did not describe management strategies aside from continuous anticoagulation and were not able to distinguish whether interruption patterns varied by anticoagulant [6, 7, 14,15,16]. While most studies recommended continuing anticoagulation through dental procedures, clinical practice regarding edoxaban management varied by procedure type, and some form of interruption was common, underscoring the need for studies reporting interruption patterns and outcomes of dental procedures for individual NOACs [6, 7, 14,15,16].
The European Heart Rhythm Association (EHRA) 2021 practical guide considers minor-risk dental procedures, such as dental extractions (1–3 teeth), paradental surgery, implant positioning, and subgingival scaling/cleaning, at minor risk of bleeding and does not recommend anticoagulant interruption [8]. Other more complex dental procedures categorised as low risk require an assessment of risk based on individual patient clinical characteristics and may or may not warrant interruption [8]. In the present study, edoxaban was typically interrupted during single and multiple tooth extractions. This is consistent with a 2019 survey of dental practitioners, which found that 62% would interrupt NOAC therapy for single tooth extractions and 94% would continue VKA therapy [17].
The EHRA practical guide does not distinguish between dental implants and crown placements, although it would be expected that implants would be associated with a longer edoxaban interruption compared with less-invasive crown placements [8]. The granularity of the EMIT-AF/VTE data allowed for the distinction of these two procedures, and no bleeding events were recorded for any dental implants or crown placements, regardless of whether or not edoxaban was interrupted. Furthermore, the mean and median duration of edoxaban interruption was longer for dental implants than for crown placements, suggesting that within the broad EHRA recommendations for minor-risk procedures, there is variation in the edoxaban management strategies adopted by dental professionals in real-world practice for specific dental procedures.
The strengths of the current subanalysis of the EMIT-AF/VTE programme include its large number of patients receiving edoxaban, the use of the ISTH bleeding definition, and the 29-day postprocedural observation window. The study was limited by the fact that it was a retrospective analysis of a prospective registry; thus, periprocedural edoxaban management was not controlled by a study protocol, and no information regarding bleeding management was available. The observed low risk of bleeding is consistent with the EHRA designation of dental procedures as minor or low-risk interventions but complicates the identification of rare, high-risk patients that may become apparent in larger studies [8]. In all, this study addresses a gap in the published data, which generally group all NOACs, by providing edoxaban-specific results for anticoagulant interruption surrounding dental procedures in routine clinical practice.