Xiaomeng Liu1,2, Luca Lammers3, Patty J. Nelemans4, Klara Mosterd1,2 and Nicole W. J. Kelleners-Smeets1–3

1Department of Dermatology, 2GROW, School for Oncology and Developmental Biology, Maastricht University Medical Centre, P. Debyelaan 25, NL-6229 HX Maastricht, 3Department of Dermatology, Catharina Hospital, Eindhoven, and 4Department of Epidemiology, Maastricht University, Maastricht, The Netherlands. E-mail: xiaomeng.liu@mumc.nl

Accepted Mar 3, 2015; Epub ahead of print Mar 10, 2015

Many patients presenting for dermatologic surgery use antithrombotic medication. It remains controversial whether or not this medication should be withheld prior to cutaneous surgery. Different studies, including one meta-analysis, have assessed the risk of bleeding when antithrombotic medication was continued and showed conflicting results (1–8). Furthermore, case reports illustrated the potential life-threatening risk of temporary discontinuation of antithrombotic medication (9–11).

We evaluated the frequency of postoperative haemorrhage and thrombotic events in patients continuing their antithrombotic medication with those temporarily withholding this medication prior to Mohs micrographic surgery.

Material and Methods

We retrospectively retrieved information from patients who underwent Mohs micrographic surgery from July 2010 to July 2012 at the Maastricht University Medical Centre. During the first year a regional guideline required platelet inhibitors (PI’s) to be discontinued 7 to 10 days prior to surgery and Vitamin K antagonists (VKA’s) to be tapered to reach an International Normalized Ratio of 2.0 or less 24 h prior to surgery. As a result of a national discussion, the guideline was changed from July 2011. Patients did not need to discontinue their antithrombotic medication with the exception of surgery performed at the medial canthus of the eye. In addition to patient and operation characteristics, we extracted information on postoperative adverse events including bleeding, infection, necrosis and wound dehiscence as well as vascular events within 7 days post-surgery. Patients with known bleeding diathesis (such as hemophilia or von Willebrand disease) were excluded. Bleeding was defined as severe if it required coagulation and/or suturing, as moderate if it required a new pressure dressing, and as mild if it could be resolved by telephone advice. Frequency of bleeding was compared amongst the 3 groups: patients continuing their anti-thrombotic medication, patients discontinuing their anti-thrombotic medication and patients not using anti-thrombotics (controls). This study protocol followed the principles of the 1975 Declaration of Helsinki and was approved by the institution’s medical ethical committee.

The χ2-square test or Fisher’s exact test was used to test for statistical significance of differences in proportions. For continuous variables the t-test for independent samples or Mann-Whitney U test was used. Multivariate logistic regression analysis was performed to evaluate the relationship between haemorrhage and antithrombotic therapy and to adjust for potential confounders. p-values ≤ 0.05 were considered to indicate statistical significance. Analysis was performed using IBM SPSS Statistics version 20.0.0.1.

Results

A total of 598 Mohs procedures in 534 patients were performed in the study period. After excluding 175 procedures (151 patients) due to unclear documentation regarding the use and/or management of antithrombotic medication, 423 procedures (383 patients) were left for analysis. Baseline characteristics are shown in Table I. Patients using antithrombotic medication were significantly older and were more often male than patients not using them. In the group that discontinued the antithrombotic medication, more operations were done in the eye region, reflecting adjustment to the protocol. There was no statistically significant difference in the diameter of the surgical defect or the closure type among all 3 groups.

A total of 18 haemorrhages in 18 patients were recorded (4.3% of procedures). The haemorrhage rates and their severity are shown in Table II. No life-threatening bleeding occurred. After adjustment for age, which was a significant co-variant, the relative risk (RR) of haemorrhage in patients discontinuing their medication compared to controls was 3.52 (95% confidence interval, CI 0.93–13.38) and in patients continuing their medication compared to controls 3.84 (95% CI 1.13–12.98). No wound dehiscence, necrosis or infection occurred in patients who experienced postoperative haemorrhage1.

1In patients using PI (n = 88) the RR of postoperative haemorrhage was 1.20 (95% CI 0.12–12.21) when comparing patients continuing versus discontinuing. In patients using VKA (n = 54), the RR of postoperative haemorrhage in continuing compared to discontinuing was 1.32 (95% CI 0.19 –9.24).

Two vascular events were identified. One patient withheld acenocoumarol 4 days preoperatively and developed a cerebellar infarction three days after surgery. Another patient discontinued carbasalatecalcium 7 days preoperatively and experienced a myocardial infarction during the operation.

Discussion

For daily practice, the RR of post-operative haemorrhage in patients continuing versus discontinuing their antithrombotic medication is of major relevance. Only one prior study focussed on this comparison and found similar risks in both groups which is in agreement with our findings (12). In this earlier study, the decision to stop antithrombotic medication or not was left to the treating physician which could potentially have introduced a selection bias. In our study this selection bias was minimized due to a uniform protocol. However, patients using antithrombotics are probably more aware of the risk of bleeding and are more likely to contact a doctor. This could also have contributed to the higher rate of registered bleeding in patients using antithrombotics.

Some studies have concluded that the use of VKA did increase the risk of bleeding while that of aspirin (a PI) did not (4, 13). In our study, both patients on VKA and PI had an increased risk of bleeding compared with control patients (results not shown). Nonetheless, we have shown that the haemorrhage rate was similar in patients continuing and discontinuing their VKA. This is important for decision making prior to dermatologic surgery. Two other studies have suggested that combinations of different agents could increase the rate of bleeding complications (7, 8). In this study, none of the patients who received a combination of antithrombotics had experienced a haemorrhage. This group was however small.

In our study, two patients who had discontinued their anticoagulation developed a vascular event within one week after surgery. Although these complications are rare, they should not be ignored due to the life-threatening potential. On the contrary, bleeding complications are mostly not life-threatening but can be extremely stressful for the patient and a potential threat for wound healing. Nevertheless, in this study, no dehiscence or necrosis appeared in the group with post-operative haemorrhage, indicating that this threat is minimal.

There are limitations to this study that should be noted. First of all, due to the retrospective design, data are missing and a limited sample prohibited potentially relevant sub-group analyses, such as the method of wound closure. The limited sample size is also associated with a reduced power to detect statistically significant results. Secondly, the recorded severity of the bleeding is subjective, but the potential misclassification of bleeding severity does not affect the results concerning total number of haemorrhages. Other possible confounders, such as patients coping ability or surgeon’s skill could not be addressed in this paper due to its retrospective design.

In conclusion postoperative haemorrhage balancing both risks and vascular events, favours continuation of anticoagulants in dermatologic surgery. Future prospective studies are needed to confirm our results.

The authors declare no conflicts of interest.

References