Abstract
Objectives: The management of patients on anticoagulation therapy is challenging. The objective of this study was to conduct a systematic review to establish the effectiveness of hemostatic interventions to prevent postoperative bleeding following dental extractions among patients taking warfarin.
Methods: A systematic review of the literature was conducted using PubMed, EMBASE and the Cochrane Central Register of Controlled Trials databases and applying relevant MeSH terms. Identified studies were screened independently by 2 reviewers using the following selection criteria: tooth extraction, patients taking warfarin as the only anticoagulant, randomized controlled trials and a hemostatic intervention.
Results: Six articles were included in the final review, all evaluating different interventions. Oral or local hemostatic agents were compared in 4 studies where patients continued taking warfarin before and after the procedure; in 3 studies, there were no differences between the agents in preventing postoperative bleeding and, in 1, Histoacryl glue was superior to a gelatin sponge. Two studies compared warfarin continuation with temporary discontinuation and found that continuation did not increase the risk of bleeding in patients who had an international normalized ratio (INR) within the therapeutic range.
Conclusions: Patients with an INR within the therapeutic range can safely continue taking the regular dose of warfarin before dental extractions. There is no evidence to support or reject the superiority of local hemostatic agents over warfarin discontinuation.
Current management of dental extractions in patients on anticoagulation therapy is still a challenging and controversial area for dental professionals. The perceived risk of bleeding among these patients is usually weighed against the risk of thromboembolic events, both of which could have severe consequences.1 Oral anticoagulants are usually prescribed for patients with previous thromboembolic events (e.g., stroke, myocardial infarction, pulmonary embolism), atrial fibrillation, prosthetic heart valves and peripheral vascular disease.2,3 The clinical management of these patients is generally complex because they are usually older, have multiple comorbid conditions and are taking multiple medications.
Warfarin, a vitamin K antagonist, remains the most commonly prescribed oral anticoagulant for the prevention and treatment of thromboembolic events.1,3 Warfarin has a narrow therapeutic index measured by the international normalized ratio (INR), i.e., 2.0–3.0 for most indications,4 and can cause major and fatal bleeding if not regularly monitored.5,6 In the past, it has been suggested that the decision to continue or withdraw warfarin among patients undergoing dental extractions be tailored to INR levels: continuing the use of warfarin with local hemostatic agents if INR is within therapeutic range1,7 and postponing warfarin to make dose adjustments if INR exceeds 3.58 or 4.0.1
A 2009 systematic review and meta-analysis of 5 randomized controlled trials (RCTs) concluded that continuing the regular dose of warfarin does not increase the risk of bleeding during minor dental procedures compared with altering or discontinuing the dose.9 However, the analysis did not evaluate the effect of other hemostatic agents used by patients in the study. In contrast, the latest guideline from the American College of Chest Physicians recommended, for patients undergoing minor dental procedures, “continuing VKAs [vitamin K antagonists] with coadministration of an oral prohemostatic agent or stopping VKAs 2 to 3 days before the procedure instead of alternative strategies,” a weak recommendation based on low- or very low-quality evidence (grade 2C).10
In dental patients, a variety of oral and local hemostatic agents, such as tranexamic acid oral rinse, cellulose and gelatin foams, applied locally, have been used as physical matrices to aid clotting initiation.2,11 Although the effectiveness of these agents has been evaluated in various clinical trials, no systematic review has focused specifically on patients taking warfarin and undergoing tooth extraction. Considering the evolving clinical recommendations and the existing gap in the literature, the objective of this study was to conduct such a review to establish the effectiveness of hemostatic interventions in preventing postoperative bleeding following dental extractions among patients taking warfarin.
Methods
We conducted a systematic review of the literature to determine which hemostatic interventions (considering both the use of hemostatic agents and warfarin discontinuation) are effective in preventing postoperative bleeding in patients taking warfarin, who are undergoing dental extractions. We followed the Preferred Reporting Items in Systematic Reviews and Meta-Analyses (PRISMA) guidelines.12
Study Selection Criteria
Participants: Patients taking warfarin and undergoing tooth extraction(s). Studies of patients taking forms of anticoagulant medication other than or in addition to warfarin, undergoing oral surgical procedures other than extractions or having an increased bleeding tendency because of chronic liver or renal disease or genetic bleeding disorders were excluded.
Intervention: Discontinuation of warfarin, any hemostatic agent including (but not limited to) tranexamic acid mouthwash, Histoacryl glue, gelatin sponges, resorbable cellulose meshes, resorbable sutures, autologous fibrin glue and commercial fibrin adhesives.
Control: Continuation of warfarin or dose alteration, addition of a hemostatic agent.
Outcome measures: Postoperative bleeding following tooth extraction and outcomes of bleeding events.
Studies: RCTs
Data Sources and Search Methods
A literature search was conducted using PubMed, EMBASE and the Cochrane Central Register of Controlled Trials databases from the time of their inception to April 2014. This was done with the help of a University of Toronto research librarian. Articles were retrieved by combining database-specific search terms for warfarin, bleeding or hemorrhage, specific hemostatic agents and hemostatic agents in general and tooth extraction. The search was then limited to RCTs and human studies. See Appendix 1 for the detailed search strategy used in PubMed.
Selection of Studies and Data Abstraction
At all stages — review of title, review of abstract, full-text review, assessment of quality and data abstraction — each study was assessed independently by 2 reviewers (4 reviewers in total: NJ, JC, DD, AK). Disagreements were resolved by consensus and consultation with other authors (YA, LA, AA). When deemed necessary, the team contacted the corresponding authors of the published articles.
We assessed the quality of the trials included in the final review using the Cochrane Risk of Bias Tool.13 In brief, we evaluated 6 domains (namely sequence generation, allocation concealment, blinding, incomplete outcome data, selective outcome reporting and other issues) based on what was reported to have happened in the study. Then, we assigned a judgement of the risk of bias (high, low or unclear) to each domain. Information retrieved from each selected study included sample size, demographic data (age and gender), INR values at baseline, number of teeth extracted per patient, description of the hemostatic interventions, outcomes data and conclusions. The plan was to conduct a meta-analysis to evaluate the effectiveness of hemostasis if any of the studies compared similar hemostatic agents.
Results
Selection of Studies
After removing duplicates, our search identified 95 potentially relevant studies (Fig. 1). A review of titles and abstracts eliminated 85 studies, resulting in 10 articles for a full-text review. Four additional studies were excluded at this stage.14-16 One14 was not a true randomized study as patients were assigned to treatments on a “simple alternating basis.” The second15 was an early report of a larger RCT17 subsequently included in our review. The third16 did not specify which type of anticoagulant treatment patients were taking. The fourth18 defined patients’ antithrombotic drug regimen as “aspirin and coumarin” without clear specification of what proportion of patients used only 1 of these and what proportion, if any, used both. Hence, 6 studies were included in this research synthesis.17,19-23 (Table 1)
Records identified through database searching
(n = 111)
(PubMed 29, EMBASE 83)
Additional records identified through other sources
(n = 0)
Records after duplicates removed
(n = 95)
Records screened
(n = 95)
Records excluded
(n = 85)
Full-text articles assessed for eligibility
(n = 10)
Full-text articles excluded
(n = 4)
- Not an RCT (1)
- Early report of an RCT (1)
- Unspecified antithrombotic agent (1)
- Unclear drug regimen
Studies included in qualitative synthesis
(n = 6)
Systematic Reviews
We identified 2 potentially relevant systematic reviews closely related to the topic.9,24 The systematic review and meta-analysis by Nematullah and colleagues9 included 5 RCTs and evaluated the effect of continuing versus stopping the regular dose of warfarin on risk of bleeding in patients undergoing minor dental surgical procedures. In this review, the hemostatic agents were not the main focus and were described as “co-interventions.” Two of the studies in this systematic review were also identified in our search and were included in our final review.17,22 The systematic review by Patatanian and Fugate,24 which included 8 studies (2 prospective observational studies and 6 RCTs), evaluated the efficacy and safety of local hemostatic agents in patients taking oral anticoagulants and undergoing dental extractions. Patients were using warfarin exclusively in only 3 of these studies,24 2 of which were also identified in our search and were included in our systematic review.20,21
Quality of Selected Studies
Blinding was the single domain where all studies had a high risk of bias (Table 2). A low risk of bias was assigned to domains that evaluated completeness of outcome data, selective reporting and other sources of bias.
| Author, year, country | Treatment groups | Sample size | Age range (mean), years | Male: female | INR range (mean) | Mean no. extracted teeth/ patient* |
|---|---|---|---|---|---|---|
| Note: n/a = not available, INR = international normalized ratio, UK = United Kingdom. | ||||||
| *Range is listed when the mean number was not reported. | ||||||
| Al-Belasy and Amer19 2003, Egypt | Experimental | 15 | 50–64 (56.5) | 10:5 | 1.9–4.3 (2.5) | 6.0 |
| Control | 15 | 53–65 (58.0) | 9:6 | 1.7–4.1 (2.4) | 6.3 | |
| Negative control | 10 | 49–67 (57.2) | 5:5 | 0.9–1.3 (1.0) | 6.5 | |
| Al-Mubarak et al.17 2007, Saudi Arabia | Group 1 | 48 | 38.0–66.6 (52.3) | 22:26 | n/a (1.8) | 1–5 |
| Group 2 | 58 | 37.0–66.4 (51.7) | 27:31 | n/a (2.4) | 1–5 | |
| Group 3 | 56 | 35.6–61.8 (48.7) | 25:31 | n/a (1.9) | 1–5 | |
| Group 4 | 52 | 39.4–66.8 (53.1) | 24:28 | n/a (2.7) | 1–5 | |
| Carter and Goss20 2003, Australia | Group 1 | 43 | 21–77 (65.2) | 22:21 | n/a (2.7) | 2 |
| Group 2 | 42 | 24–86 (65.7) | 32:10 | n/a (2.8) | 3 | |
| Carter et al.21 2003, Australia | Group 1 | 26 | 24–85 (n/a) | 16:10 | 2.3–4.0 (3.0) | 1–13 |
| Group 2 | 23 | 40–83 (n/a) | 15:8 | 2.1–4.0 (3.1) | 1–18 | |
| Evans et al.22 2002, UK | Experimental | 57 | 36–92 (67.0) | 36:21 | 1.2–4.7 (2.5) | 2 |
| Control | 52 | 30–93 (66.0) | 37:15 | 1.3–2.3 (1.6) | 3 | |
| Halfpenny et al.23 2001, UK | Experimental | 20 | 33.4–83.4 (66.5) | 13:17 | 2.0–4.1 (2.7) | 2 |
| Control | 26 | 38.2–79.3 (64.8) | 17:9 | 2.1–4.1 (2.9) | 1.5 | |
| Al-Belasy and Amer19 | Al-Mubarak et al.17 | Carter and Goss20 | Carter et al.21 | Evans et al.22 | Halfpenny et al.23 | |
|---|---|---|---|---|---|---|
| *+ = high risk, − = low risk, ? = uncertain. | ||||||
| Random sequence generation | − | − | − | − | + | − |
| Allocation concealment | ? | ? | − | ? | − | − |
| Blinding Participants | ? | + | + | + | + | + |
| Personnel | − | ? | ? | ? | + | + |
| Outcome assessment | + | − | ? | + | + | + |
| Incomplete outcome data | − | − | − | − | − | − |
| Selective reporting | − | − | − | − | − | − |
| Other sources of bias | − | − | − | − | − | − |
Effectiveness of Hemostatic Interventions after Dental Extraction
Compared interventions and study outcomes are summarized in Table 3. Oral or local agents were compared in 4 studies19-21,23 where patients continued taking warfarin before and after the procedure. Two studies17,22 compared warfarin continuation with temporary discontinuation. In general, all 6 studies compared different interventions. This heterogeneity in treatment modalities precluded us from conducting a meaningful meta-analysis.
| Article | Compared interventions | Outcomes: no. (%) patients with postoperative bleeding | Conclusions | ||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Al-Belasy and Amer19 | Experimental: Patients continued warfarin, and sockets were dressed with Histoacryl glue (n-butyl-2- cyanoacrylate) after extractions Control: Patients continued warfarin, and sockets were dressed with gelatin sponge after extractions Negative control: Patients were never on warfarin, and sockets were dressed with gelatin sponge after extractions | Experimental: 0 (0.0%) Control: n = 5 (33.3%) (2 on day 2; 2 on day 3; 1 on day 5) Negative control: 0 (0.0%) p = 0.016, control vs. experimental p = 0.046, control vs. negative control | Patients on warfarin can safely undergo extractions without changing their anticoagulant regimen, provided that an effective local hemostatic method is used, such as Histoacryl glue. | ||||||||||||||||||||||||
| Al-Mubarak et al.17 | 4 study groups were compared: |
|
Dental extractions can be carried out safely among patients on warfarin without alteration of the anticoagulant dosage, as long as the INR is ≤ 3.0 and effective hemostasis is achieved. Suturing plays no significant role in achieving hemostasis and should be used only when needed. | ||||||||||||||||||||||||
| Group 1. Stopped warfarin treatment 2 days before extractions and resumed after 12 h; no sutures placed | |||||||||||||||||||||||||||
| Group 2. Continued warfarin treatment; no sutures placed. | |||||||||||||||||||||||||||
| Group 3. Stopped warfarin treatment 2 days before extractions and resumed after 12 h; sutures placed | |||||||||||||||||||||||||||
| Group 4. Continued warfarin treatment; sutures placed | |||||||||||||||||||||||||||
| Carter and Goss20 | Group 1: Patients continued warfarinandused4.8% tranexamic mouthwash for 2 days after tooth extraction, 4 times/day for 2 minutes Group 2: Patients continued warfarin and used 4.8% tranexamic mouthwash for 5 days after tooth extraction, 4 times/day for 2 minutes | Bleeding on day 2 after extractions: Group 1: 2 (4.7%) Group 2: 1 (2.4%) p = 0.57 | Both regimens of 4.8% tranexamic mouthwash were effective at controlling local hemostasis in patients on warfarin. | ||||||||||||||||||||||||
| Carter et al.21 | Group 1: Patients continued warfarin; sockets were irrigated with 4.8% tranexamic acid mouthwash; resorbable cellulose mesh soaked in the mouthwash was then placed in each socket; and mouthwash rinse was continued for 7 days after extraction, 4 times/day for 2 minutes Group 2: Patients continued warfarin; 80 mL blood sample was collected from patients 1–2 weeks before surgery to prepare autologous fibrin glue (AFG), which was then applied to socket walls, wound site and over sutures | Bleeding on day 2 after extractions: Group 1: 0 (0.0%) Group 2: 2 (8.7%) p = 0.12 | Dental extractions can be performed without alteration to existing anticoagulant treatment. Both tranexamic acid mouthwash and AFG are similarly effective in preventing postoperative bleeding. | ||||||||||||||||||||||||
| Evans et al.22 | Experimental: Patients continued warfarin Control: Patients discontinued warfarin 2 days before extractions and resumed same dose after the extraction | Postoperative days when bleeding occurred were not specified. Experimental: 15 (26%) Control: 7 (14%) p = 0.10 | The observed difference in bleeding rates was not clinically significant. Dental extractions can be safely done in a hospital setting without changing warfarin regimen if INR is < 4.1. The routine discontinuation of warfarin before dental extractions should be reconsidered. | ||||||||||||||||||||||||
| Halfpenny et al.23 | Experimental: Patients continued warfarin; after the extraction, sockets were dressed with Beriplast P, a fibrin adhesive Control: Patients continued warfarin; after the extraction, sockets were dressed with Surgicel, a resorbable oxycellulose dressing | Bleeding on day 1 after extractions (numbers too small to test for significance of difference): Experimental: 2(10%) Control: 1(3.8%) | Both local hemostatic agents, Beriplast P and Surgicel, are effective in achieving local hemostasis among patients on warfarin. | ||||||||||||||||||||||||
Studies that Compared Oral or Local Hemostatic Agents while Patients Continued Warfarin
The study by Al-Belasy and Amer enrolled patients in experimental, control and negative control (i.e., those never taking warfarin) groups.19 After extractions, sockets were dressed with Histoacryl glue in the experimental group and with gelatin sponge in the other 2 groups. Minor postoperative bleeding was observed in only 5 patients (33%) in the control group (p < 0.05 compared with other groups) and was successfully stopped with local hemostatic treatment with 5% tranexamic acid. The authors concluded that patients taking warfarin can safely undergo dental extractions without any change of regimen if an effective local hemostatic agent, such as Histoacryl glue, is used.19
Studies by Carter and Goss20 and Carter et al.21 investigated the effect of tranexamic mouthwash after tooth extraction. The first study20 randomly assigned patients to a 5-day or a 2-day regimen of 4.8% tranexamic acid mouthwash applied 4 times a day. Minor bleeding was observed in only 3 patients and there were no significant differences between the groups. The second study21 compared a 7-day regimen of 4.8% tranexamic acid mouthwash with autologous fibrin glue that was applied to the extraction sockets and over the sutures. Minor bleeds were observed in only 2 patients in the autologous fibrin glue group, with no significant differences among the groups. Both studies concluded that the compared interventions were similarly effective.
Halfpenny et al.23 compared the effectiveness of the fibrin adhesive, Beriplast P (CSL Behring, King of Prussia, Penn.), with the resorbable oxycellulose dressing, Surgicel (Johnson & Johnson, New Brunswick, N.J.) after dental extractions. After extraction, sockets were dressed with 1 of the hemostatic agents and then sutured with softgut. Bleeding was observed in 3 patients in total: 2 required additional suturing and 1 required admission to hospital because of more significant bleeding. The low bleeding rate did not allow statistical testing of the difference. The authors concluded that both treatments were equally effective in preventing bleeding after dental extractions.
Studies that Compared Warfarin Continuation and Discontinuation
The study by Evans et al.22 randomly assigned patients to continue or discontinue warfarin 2 days before dental extraction. The authors reported a higher rate of bleeding in patients who continued warfarin (26%) than in those who stopped (14%), although the difference was neither statistically nor clinically significant. Two patients in the warfarin continuation group required a hospital visit to stop the bleeding. The authors concluded that warfarin can be safely continued in patients who undergo dental extraction in a hospital setting if their INR is below 4.1.
The study by Al-Mubarak et al.17 applied a factorial design of 4 possible combinations of warfarin continuation and discontinuation and suturing and no suturing. The observed bleeding rate on the first postoperative day was slightly but not significantly higher in the 2 groups that continued warfarin compared with the groups that discontinued warfarin treatment. Bleeding rates significantly diminished by day 7 with no significant differences among the groups on any of the postoperative days. Suturing did not play any role, and wound healing was similar across the groups. All bleeding events were described as “of the mild transient type.”17 The authors concluded that warfarin therapy can be safely continued in patients during dental extractions if the INR is 3 or lower; however, suturing or any other invasive manipulation should be used only when required.17
Discussion
Optimal management strategies for patients taking oral anticoagulants and undergoing dental procedures have been discussed extensively in the past.2,8-10,24,25 However, there is still a controversy over whether warfarin should be routinely discontinued or if additional hemostatic agents should be used.9,10 Our systematic review summarizes RCTs that compared any hemostatic management of patients taking warfarin and undergoing single or multiple dental extractions. Only 6 RCTs, conducted between 2001 …
