Bjom-240.

qxd 5/23/02 9:08 AM Page 248

British Journal of Oral and Maxillofacial Surgery (2002) 40, 248–252 BRITISH
© 2002 The British Association of Oral and Maxillofacial Surgeons
doi: 10.1054/bjom.2001.0773, available online at http://www.idealibrary.com on Journal of
Oral and
Maxillofacial
Surgery

Can warfarin be continued during dental extraction? Results of

a randomized controlled trial

I. L. Evans,* M. S. Sayers,* A. J. Gibbons,* G. Price,* H. Snooks,† A. W. Sugar*

*Maxillofacial Unit, Morriston Hospital, Swansea, UK; †Mid and West Wales Research and Development
Support Unit, School of Postgraduate Studies in Medical and Health Care, University of Wales, Swansea, UK

SUMMARY. A randomized controlled trial was set up to investigate whether patients who were taking warfarin
and had an International Normalised Ratio (INR) within the normal therapeutic range require cessation of
their anticoagulation drugs before dental extractions. Of 109 patients who completed the trial, 52 were allo-
cated to the control group (warfarin stopped 2 days before extraction) and 57 patients were allocated to the
intervention group (warfarin continued). The incidence of bleeding complications in the intervention group
was higher (15/57, 26%) than in the control group (7/52, 14%) but this difference was not significant. Two
patients in the study required hospital review for bleeding and all other episodes of bleeding were controlled by
patients at home. Continuing warfarin when the INR is :4.1 may lead to an increase in minor post-extraction
bleeding after dental extractions but we found no evidence of an increase in clinically important bleeding.
As there are risks associated with stopping warfarin, the practice of routinely discontinuing it before dental
extractions should be reconsidered. © 2002 The British Association of Oral and Maxillofacial Surgeons

INTRODUCTION bacterial endocarditis may increase the effects of war-

Warfarin is the most commonly prescribed oral anti-
coagulant. At present over 300,000 people in the UK
are taking oral anticoagulants1 and the treatment is
underused in some conditions.2 With an ageing popula-
tion in the UK and a greater proportion of this popula-
tion retaining their teeth, the number of patients taking
warfarin who require dental extractions is likely to
increase.
Therapeutic levels of warfarin are measured by the
International Normalised Ratio (INR). The British
Society of Haemotology has published guidelines on
anticoagulant control which recommend a maximum
target INR of 3.5, with a range of 3–4.3 For dental
extractions, patients who have been taking warfarin are
at an increased risk of perioperative thromboembolism if
the drug is stopped but may be at an increased risk of
bleeding if it is continued.4 Patients may bleed from
extraction sockets and may also bleed into the medial
pterygoid muscle if an inferior dental nerve block is
given. A small bleed can produce trismus but a large
bleed could embarrass the airway.5 On the other hand,
discontinuing warfarin can cause serious embolic com-
plications6 and may lead to a rebound hypercoagulable
state.7–10 In addition, several antibiotics that are pre-
scribed as prophylaxis during dental extraction against
248
farin and the risk of bleeding. Indeed, several case
reports have been published of antibiotic-induced bleed-
ing in patients who were taking warfarin after dental
procedures.11
Customary practice in the UK has been to stop war-
farin treatment 2 days before extractions, to do an INR
on the day of operation and to proceed if the INR is
:2.1. Warfarin is started again later the same day.12
Several authors have suggested that the anticoagulant
regimen does not require alteration for dental extrac-
tions if the INR is :4.0.1,12–16 Additional measures that
can be taken to ensure haemostasis include packing
sockets with oxidized cellulose gauze and suturing all
sockets.1,12 Tranexamic acid has been used as a mouth
rinse to reduce haemorrhage further postoperatively.15,17
Because of the risks associated with either stopping or
continuing warfarin for dental extraction, general dental
and medical practitioners routinely refer patients who
are taking warfarin to maxillofacial units for this proce-
dure. If patients could be treated without altering their
anticoagulant regimen, then it is possible that most
dental extractions could be done in general dental prac-
tice on the same day as regular INR blood monitoring.
This would often be more convenient and quicker for the
patient, cost-effective and would help to reduce hospital
waiting lists.
Bjom-240.qxd 5/23/02 9:08 AM Page 249
PATIENTS AND METHODS
The trial lasted from May 1999 to October 2000. A total
of 117 consecutive patients who were referred by their
general dental or medical practitioners to the Maxillo-
facial Unit at Morriston Hospital, Swansea, and its
related departments at the West Wales General Hospital,
Camarthen, and Neath General Hospital, who were taking
warfarin and required dental extractions were included
in the study. Patients who could not give informed con-
sent, did not have access to a telephone, had an INR 94,
had a history of liver disease or coagulopathies,
or could not attend for follow-up were excluded from
the study.
At the initial consultation a full medical history was
taken and an orthopantogram and clinical examination
were done. Those patients who met the study criteria
were given verbal and written information on the study
and asked for their consent. Randomization was then by
consecutive sealed envelopes containing allocation to
the two study groups. A register was kept to confirm the
sequence of allocation.
A preoperative INR and coagulation screen were
arranged on the day of operation.
The intervention group continued taking warfarin as
usual (anticoagulant group). If a patient’s INR was 94
on the day of operation they were to be withdrawn from
the study. The control group stopped taking warfarin
2 days before their dental extractions. If a patient’s INR
was 92, a further dose of warfarin was omitted and the
operation was rebooked for the following day.
Extractions were done only if the INR was -2.
Antibiotic prophylaxis was given for patients at risk
of endocarditis in both groups in accordance with British
National Formulary guidelines. Dental extractions were
done under local anaesthesia using 2% (20 mg/ml) ligno-
caine hydrochloride with 1/80,000 (12.5 ␮g/ml) adrena-
line. Local infiltrations and regional blocks were used in
the mandible and maxilla as appropriate. Dental extrac-
tions were by forceps and elevators. Surgeons were
instructed to use as atraumatic a technique as possible.
Where a surgical approach was required the minimum
size of mucoperiosteal flap was raised and the least
amount of bone was removed. Each extraction socket
was packed with oxycellulose dressing (Surgicel®) and
sutured with 3/0 polyglactin 910 (Vicryl®) sutures.
Patients were then given a gauze swab to bite on for 10
minutes and they were observed for a further 10 minutes
before being discharged.
If haemostasis was not achieved after biting on a
gauze swab for 10 minutes this was recorded as an
immediate bleed. All patients were given verbal and
written information with a telephone number to contact
the on-call maxillofacial senior house officer in case of
any emergencies. A logbook was given to each patient to
Can warfarin be continued during dental extraction? 249
record any bleeding episodes, whether pressure was
used to stop bleeding, and whether professional help was
required to deal with problems. Paracetomol 1 g 6 hourly
as required was prescribed and patients were advised not
to use any other form of analgesia. After the extractions
the control (warfarin withdrawn) group were told to
resume taking their warfarin on the same day. All
patients were given a review appointment for a week
after surgery. Those who failed to attend were sent a
further appointment and, if necessary, contacted by
telephone.
Data were collected from the clinical notes made at
each appointment and from the logbooks completed by
patients. All complications were recorded and were cate-
gorized as bleeding problems, prescription of a course
of antibiotics, or prescription of additional analgesia
for pain.
Data were analysed using the Statistical Package for
Social Scientists (SPSS) on an ‘intention to treat’ basis.
Descriptive statistics are presented with ␹2 and unpaired
Student’s t tests used to compare proportions and means
as appropriate; 95% confidence intervals (CI) are also
presented.
Ethical approval for the study was granted by the
Iechyd Morgannwg Local Research Ethics Committee.
The use of tranexamic acid as a mouth rinse was not
permitted by the committee.
RESULTS
Of the 117 patients, 3 were excluded, one because of
liver disease and 2 because they refused consent. Of the
114 patients who were randomized, 5 were withdrawn,
leaving 109 who completed treatment (Table 1). Of
these, 57 were allocated to the anticoagulant group and
52 to the warfarin withdrawn group. One hundred
patients attended for review in clinic and 9 patients were
reviewed by telephone. There were no significant
differences between the groups in the mean age, propor-
tions of men, number of teeth extracted, or the number
of patients given prophylactic antibiotics. Naturally,
the mean INR for the anticoagulant group was signifi-
cantly higher at 2.5 than that for the control (warfarin
Table 1 Trial recruitment
Parameter Continued warfarin Stopped warfarin
(60 patients) (54 patients)
Withdrawn INR9 4 1 0
Stopped warfarin 1 0
Withdrew consent 1 0
Died before treatment 0 1
Emergency extraction 0 1
Extraction completed 57 52
and patient followed up
Bjom-240.qxd 5/23/02 9:08 AM Page 250

250 British Journal of Oral and Maxillofacial Surgery

withdrawn) group, at 1.6 (P:0.001) (Table 2). One
patient who continued warfarin had an INR level 94
and should have been excluded preoperatively. Two
patients in the control group had INR levels 92 (2.1 and
2.3 respectively) and should have discontinued warfarin
for a further day before operation. These three patients
were included in their allocated study groups on the
intention to treat basis.
Table 3 summarizes the main outcomes. The only two
patients who sought help for bleeding were in the anti-
coagulant group. The first of these had an oroantral com-
munication (OAC) that developed at the time of removal
of an upper third molar; it had been closed immediately
by advancing a buccal flap. Bleeding had started again
from this site 6 hours later and was stopped by resuturing
under local anaesthetic and applying pressure. The patient
was admitted overnight for observation and had no fur-
ther bleeding episodes. The second patient had bled from
an upper second premolar socket, which was stopped
with local anaesthetic, packing, suturing and pressure.
The patient was discharged home from the accident and
emergency department and had no further bleeding.
The overall rate of bleeding complications in the anti-
coagulant group was higher than in the control group
Table 2 Baseline characteristics of patients in the two groups
(26% compared with 14%, ␹2 2.8, P:0.1; 95% CI
0–27%). The difference was not significant. The three
immediate bleeds were controlled with additional
sutures and pressure and all but the two delayed bleeds
described above were controlled by patients biting on
gauze squares at home.
Five patients had other complications (Table 2).
In the anticoagulant group two patients were prescribed
antibiotics and one was prescribed additional analgesia
by clinicians unconnected with the study. One patient in
each group was prescribed antibiotics for mild local
infections of the extraction sockets by clinicians
involved in the study.
We found no association between the risk of bleeding
and the giving of prophylactic antibiotics. Antibiotics
were given to 7 of 22 patients who experienced a bleed-
ing complication compared with 18 of 84 patients who
did not (␹2 1.23, P:0.3). In addition, no association
was found between the risk of bleeding and the number
of teeth extracted. Those who experienced bleeding after
extraction had a mean of three teeth extracted, compared
with a mean of two in the patients who did not experi-
ence bleeding (P:0.2). However, the numbers of
patients were low, and small to moderate associations
would not be identified.

Parameter Anticoagulant group Control group P value

(n:57) (n:52) DISCUSSION
Number of men (%) 36 (63) 37 (71) 0.38
Mean age (range) 67 (36–92) 66 (30–93) 0.48 Previous prospective trials comparing perioperative anti-
Mean number of 2 (1–7) 3 (1–9) 0.24 coagulant regimens for dental extractions have been too
teeth extracted
(range) small to detect significant differences in bleeding5,12,14
Prophylactic 13 12 0.26 or have not been adequately randomized.17
antibiotics prescribed In our study there was a higher risk of bleeding in the
Mean INR (range) 2.5 (1.2–4.7) 1.6 (1.2–2.3) :0.001
anticoagulant group (15 patients in the anticoagulant
group and 7 patients in the control group), although the
difference was not significant. All but two of the bleed-
Table 3 Comparison of complications in study groups
ing complications were dealt with by the patients at
Complications Anticoagulated Control Total home, suggesting that even if postoperative bleeding
group group episodes do occur more commonly when patients con-
Sought help for bleeding 1 0 1 tinue anticoagulant treatment, most can be controlled by
at hospital the patient with no professional help. This supports pre-
Sought help for bleeding 1 0 1 vious research findings. Bailey and Fordyce5 found a
at hospital and admitted
Immediate postoperative 3 0 3 delayed bleeding tendency in patients who continued to
bleeding take warfarin, but all episodes were controlled by local
Delayed postoperative 9 7 16 measures. Devani et al.12 found a very low incidence of
bleeding treated by patient
at home postoperative bleeding in anticoagulated patients under-
Sought help for bleeding by 1 0 1 going dental extractions. Wahl6,13 in a review found
telephone at home there was little or no difference in terms of blood loss
Postoperative antibiotics 3 1 4
prescribed after dental surgery between patients who were receiv-
Additional analgesia given 1 0 1 ing anticoagulants and those whose coagulation was
for pain normal. In 2400 documented dental operations, he found
None 38 44 82
that only 12 patients experienced bleeding that was
Total 57 52 109
uncontrolled by local measures. In 7 of these 12 the
Bjom-240.qxd 5/23/02 9:08 AM Page 251
extent of anticoagulation was above recommended ther-
apeutic limits.
We have not found any reports of patients experiencing
serious harm from postoperative bleeding after dental
extractions while continuing anticoagulants.6 On the
other hand, stopping warfarin does not lead to a pre-
dictable fall in INR. The INR often falls below 1.5 and
exposes the patient to risks of thromboembolism.12 Four
patients have been reported to have had fatal embolic
complications after anticoagulant treatment was with-
drawn,18–21 supporting the view that anticoagulant treat-
ment should not be stopped before dental extractions
without good reason.
We did not find a connection between taking prophy-
lactic antibiotics and increased bleeding. Various authors
have suggested that antibiotics may cause an increase in
bleeding in patients taking warfarin and having dental
operations.11,22 The effects of warfarin may be magni-
fied by the suppression by antibiotics of intestinal bacte-
rial flora that can provide a supplementary source of
vitamin K when the patient’s intake is low. In addition,
erythromycin stimulates liver enzymes and potentiates
the effects of warfarin, and metronidazole inhibits the
metabolism of warfarin, again potentiating its effects.
Multiple extractions of teeth are generally considered
to be a good test of the efficacy of the haemostatic mech-
anism.5 However, we found no correlation between post-
operative bleeding and the number of teeth removed.
Tranexamic acid used as a mouth rinse has been shown
to prevent bleeding after oral surgery in patients who are
taking anticoagulants by inhibiting plasminogen activation
and fibrinolysis and acting as a clot stabilizer.17,23 In ani-
mal experiments tranexamic acid mouth rinses combined
with Surgicel® packing proved effective in reducing
blood loss.24 In human trials, Blinder et al. did not find
this technique better at reducing bleeding than gelatin
sponge and sutures alone.14 However, Ramstrom et al.
found that in patients who continued taking warfarin at
the time of dental extractions, those who used tranexamic
mouth rinses had significantly less bleeding than those
who did not.25 More recently the use of Surgicel® soaked
in tranexamic acid, sutures and tranexamic acid mouth
rinses has proved particularly effective in preventing
postoperative bleeding.26 The use of tranexamic acid
mouth rinse was considered in the current study.
However, and somewhat surprisingly, tranexamic acid
does not have a product licence for use as a mouth rinse
in the UK27 and can be used only for named patients. The
local research ethics committee therefore did not think its
use was appropriate.
Our study has shown that although there may be an
increase in the number of patients experiencing postoper-
ative bleeding when anticoagulation is continued, this
does not appear to be a clinically important problem.
Nearly all patients controlled bleeding by simple pressure
Can warfarin be continued during dental extraction? 251
and in the two cases that did require professional help,
bleeding was stopped by local treatment of the extrac-
tion socket.
The cost and inconvenience of referring a patient to
hospital to have dental extractions is justified only if
there is a substantial benefit to the patient. Patients who
take warfarin would certainly benefit if dental extractions
could be performed safely in general practice. Our study
has shown that dental extractions can be done safely in a
hospital without stopping warfarin treatment if the
patient’s INR is :4.1. It remains to be established if this
finding can be translated to general dental practices.
ACKNOWLEDGEMENTS
We thank Kathy Wareham, clinical research scientist, Clinical
Research Unit, School of Postgraduate Studies, University of Wales
Swansea, for methodological advice at the outset of the study; Alan
Watkins, senior lecturer in statistics, European Business Management
School, University of Wales Swansea, for providing statistical advice;
Simon Hodder, consultant maxillofacial/head and neck surgeon,
Morriston Hospital for permission to recruit his patients into the study
and Judy Jones, research secretary, Maxillofacial Unit, Morriston
Hospital, for organizing data collection and recording.
REFERENCES
1. Webster K, Wild J. Management of anticoagulation in patients
with prosthetic heart valves undergoing oral and maxillofacial
operations. J Oral Maxillofac Surg 2000; 38: 124–126.
2. Connolly SJ. Anticoagulation for patients with atrial fibrillation
and risk factors for stroke. BMJ 2000; 320: 1219–1220.
3. Haemostasis and Thrombosis Task Force for the British
Committee for Standards in Haemotology. Guidelines on oral
anticoagulation, 3rd ed. Br J Haematol 1998; 101: 374–387.
4. Drugs in the peri-operative period: 4 – Cardiovascular drugs.
Drug Ther Bull 1999; 37: 89–92.
5. Bailey BMW, Fordyce AM. Complications of dental extractions in
patients receiving warfarin anticoagulant therapy – a controlled
clinical trial. Br Dent J 1983; 155: 308–310.
6. Wahl MJ. Myths of dental surgery in patients receiving
anti-coagulant therapy. J Am Dent Assoc 2000; 131: 77–81.
7. Poller L, Thomson J. Evidence for ‘rebound’ hypercoagulability
after stopping anticoagulants. Lancet 1964; ii: 62–64.
8. Grip L, Blomback M, Schulman S. Hypercoaguable state and
thromboembolism following warfarin withdrawal in post
myocardial infarct patients. Eur Heart J 1991; 12: 1225–1233.
9. Harenberg J, Haas R, Zimmerman R. Plasma hypercoagulability
after termination of oral anticoagulants. Thromb Res 1983;
29: 627–633.
10. Mulligan R. Response to anticoagulant drug withdrawal. J Am
Dent Assoc 1987; 115: 435–438.
11. Wood GD, Deeble T. Warfarin: dangers with antibiotics. Dental
Update 1993; 20: 350–353.
12. Devani P, Lavery KM, Howell CTJ. Dental extractions in patients
on warfarin: is alteration of anticoagulant regime necessary?
Br J Oral Maxillofac Surg 1998; 36: 107–111.
13. Wahl MJ. Dental surgery in anticoagulated patients. Arch
Intern Med 1998; 158: 1610–1615.
14. Blinder D, Manor Y, Martinowitz U, Taicher S, Hashomer T.
Dental extractions in patients maintained on continued oral
anticoagulant: comparison of local hemostatic modalities.
Oral Surg Oral Med Oral Pathol 1999; 88: 137–140.
15. Lippert S, Gutschik E. Views of cardiac-valve prosthesis patients
and their dentists on anticoagulation therapy. Scand J Dent Res
1994; 102: 168–171.
Bjom-240.qxd 5/23/02 9:08 AM Page 252

252 British Journal of Oral and Maxillofacial Surgery

16. Beirne OR, Koehler JR. Surgical management of patients on 27. Evans IL, Sugar AW. Re: Management of anticoagulation in
warfarin sodium. J Oral Maxillofac Surg 1996; 54: 1115–1118. patients with prosthetic heart valves undergoing oral and
17. Souto JC, Oliver A, Zuazu-Jausoro, Vives A, Fontcuberta J. Oral maxillofacial operations. Br J Oral Maxillofac Surg 2001;
surgery in anticoagulated patients without reducing the dose of 39: 245.
oral anticoagulant; a prospective randomized study. J Oral
Maxillofac Surg 1996; 54: 27–32.
18. Akbarian M, Austen WG, Yurchak PM, Scannell JG.
Thromboembolic complications of prosthetic cardiac valves.
Circulation 1968; 37: 826–831. The Authors
19. Behrman SJ, Wright IS. Dental surgery during continuous
anticoagulation therapy. J Am Dent Assoc 1961; 62: 172–180. I. L. Evans BSc, FDS, RCS(Eng)
20. Marshall J. Rebound phenomena after anticoagulant therapy in M. S. Sayers FDS, RCS(Edin)
cerebrovascular disease. Circulation 1963; 28: 329–332. Senior House Officers
21. Ogiuchi H, Ando T, Tanaka M et al. Clinical reports on dental A. J. Gibbons MA(Cantab), FDS, RCS(Edin), FRCS(Edin)
extractions from patients undergoing anticoagulant therapy. Bull Specialist Registrar
Tokyo Dent Coll 1985; 26: 205–212. G. Price RGN
22. Bandrowsky T, Vorono AA, Borris TJ, Marcantoni HW. Sister
Amoxicillin-related postextraction bleeding in an anticoagulated A. W. Sugar FDS, RCS(Eng), FDS, RCS(Edin)
patient with tranexamic acid rinses. Oral Surg Oral Med Oral Consultant Oral and Maxillofacial Surgeon
Pathol 1996; 82: 610–612. Morriston Hospital
23. Sindet-Pedersen S, Ramstrom G, Bernvil S, Blomback M. The Swansea, Wales, UK
effect of tranexamic acid mouthrinse in anticoagulant treated H. Snooks PhD
patients undergoing oral surgery. N Engl J Med 1989; 320: Senior Lecturer in Health and Social Care Research
840–843. Mid and West Wales Research and Development Support Unit
24. Vinckier F, Vermylen J. Blood loss following dental extraction in Morriston Hospital
anticoagulated rabbits: effects of tranexamic acid and socket Swansea, Wales, UK
packing. Oral Surg Oral Med Oral Pathol 1985; 59: 2–5.
25. Ramstrom G, Sindet-Pedersen S, Hall G, Blomback M, Alander U. Correspondence and request for offprints to: Mr A. W. Sugar FDS,
Prevention of postsurgical bleeding in oral surgery using tranex- RCS, Consultant in Oral and Maxillofacial Surgery, Maxillofacial
amic acid without dose modification of oral anticoagulants. J Oral Unit, Morriston, Swansea, SA6 6NL, UK. Tel:;44(0) 1792 703077;
Maxillofac Surg 1993; 51: 1211–1216. Fax:;44(0) 1792 703068; E-mail: adrian.sugar@swansea-tr.
26. Bal BJ, Hardee PS. Efficacy and cost effectiveness of tranexamic wales.nhs.uk
acid mouthrinse for oral surgery in warfarinized patients. Br J Oral
Maxillofac Surg 2000; 38: 390. Accepted 6 December 2001