Anticoagulation in dental surgery: Is it rude to interrupt?

A DECISION-TREE REANALYSIS

In 2010, Balevi published a decision-tree analysis that slightly favored withdrawing warfarin for dental surgery, but he stated that the analysis “can be updated in the future as more accurate and up-to-date data for each of the variables in the model become available.”¹⁶ Now that there are more accurate and up-to-date data, it is time to revisit this decision-tree analysis.

In Balevi’s analysis, major bleeding is not defined. But major bleeding after dental surgery should be defined as any bleeding requiring more than local measures for hemostasis. In calculating probabilities for the analysis, Balevi cited studies allegedly showing high incidences of major bleeding after dental extractions with warfarin continuation.^17,18 There were some minor bleeding complications necessitating additional local measures for hemostasis in these studies, but no major bleeding complications at all in the warfarin- continuation or warfarin-interruption group. There were no significant bleeding events in either study, and the differences in bleeding rates were not significantly different between the two groups. In both studies, the authors concluded that warfarin interruption for dental surgery should be reconsidered.

Similarly, Balevi accurately asserted that there has never been a reported case of fatal bleeding after a dental procedure in an anticoagulated patient, but “for the sake of creating balance,”¹⁶ his decision-tree analysis uses a fatal bleeding probability of 1%, based on an estimated 1% risk for nondental procedures (eg, colorectal surgery, major abdominal surgery). It is unclear how a 1% incidence creates “balance,” but dental surgery is unlike other types of surgery, and that is one reason there has never been a documented postdental fatal hemorrhage in an anticoagulated patient. Major vessels are unlikely to be encountered, and bleeding sites are easily accessible to local hemostatic methods.

Balevi used an embolic complication incidence of 0.059% with warfarin interruption of 3 days. Perhaps he used such a low embolic probability because of his incorrect assertion that “there has been no reported case of a dental extraction causing a cardiovascular accident in a patient whose warfarin was temporarily discontinued.”¹⁶ In fact, our group has now identified at least 22 reported cases of embolic complications after temporary interruption of warfarin therapy in patients undergoing dental surgery.⁴ These included 12 embolic complications (3 fatal) after interruption periods from 1 to 5 days.^19,20 In addition, there are numerous cases of embolic complications reported in patients whose warfarin was temporarily interrupted for other types of surgery.^21,22

The literature shows that embolic complications after temporary warfarin interruption occur at a much higher rate than 0.059%. Many documented embolic complications have occurred after relatively long warfarin interruption periods (greater than 5 days), but many have occurred with much shorter interruptions. Wysokinski et al²¹ showed that there was a 1.1% incidence of thromboembolic events, more than 18 times greater than Balevi’s incidence, in patients whose warfarin was interrupted for 4 or 5 days with or without bridging therapy. One of these patients developed an occipital infarct within 3 days after stopping warfarin without bridging (for a nondental procedure). Garcia et al²² showed that of 984 warfarin therapy interruptions of 5 days or less, there were 4 embolic complications, a rate (0.4%) more than 6 times greater than that reported by Balevi.

Even if one were to accept a 0.059% embolic risk from interruption of warfarin, that would mean for every 1,700 warfarin interruptions for dental procedures, there would be one possibly fatal embolic complication. On the other hand, if 1,700 dental surgeries were performed without warfarin interruption, based on the literature, there may be some bleeding complications, but none would be fatal. If airline flights had a 0.059% chance of crashing, far fewer people would choose to fly. (There are 87,000 airline flights in the US per day. A 0.059% crash rate would mean there would be 51 crashes per day in the United States alone.)

But regardless of whether the embolic risk is 0.059% or 1%, the question comes down to whether an anticoagulated patient should be subjected to a small but significant risk of death or permanent disability (if anticoagulation is interrupted) or to a small risk of a bleeding complication (if anticoagulation is continued), when 100% of cases up until now have apparently resulted in a full recovery.

As a result, the decision-tree analysis was fatally flawed by grossly overestimating the incidence of fatal bleeding when warfarin is continued, and by grossly underestimating the incidence of embolic complications when warfarin is interrupted.

IS WARFARIN CONTINUATION ‘TROUBLESOME’?

An oral surgeon stated, “My experience and that of many of my colleagues is that even though bleeding is never life-threatening [emphasis mine], it can be difficult to control at therapeutic levels of anticoagulation and can be troublesome, especially for elderly patients.”²³ The American College of Chest Physicians stated that postoperative bleeding after dental procedures can cause “anxiety and distress.”³ Patients with even minor postoperative bleeding can be anxious, but surely, postoperative stroke is almost always far more troublesome than postoperative bleeding, which has never been life-threatening. Although other types of surgery may be different, there is no need to interrupt lifesaving anticoagulation for innocuous dental surgery.

My mother was right—it can be rude to interrupt. Anticoagulation should not be interrupted for dental surgery.