Introduction

Many patients undergoing skin surgery are prescribed at least one anticoagulant or antiplatelet agent.¹ Skin cancer is rising around the world, with over 2 million new diagnoses of skin cancer each year and a death attributed to melanoma every 4 minutes.^{2, 3} As a consequence, skin surgery is becoming increasingly more common. For best patient outcomes and in order to minimize patient harms, it is imperative that the skin surgeon is aware of patient factors that may influence the safety of the procedure and perioperative course.

Complicating the landscape is the advent of newer anticoagulant agents, principally the direct oral anticoagulants (DOACs). Warfarin, a vitamin K antagonist (VKA), has been the mainstay of oral anticoagulation over the last five decades.⁴ However, since the pivotal studies that demonstrated the safety, non-inferiority, and increased convenience of DOAC agents, DOACs have been increasingly prescribed at the expense of VKAs.⁵

The benefit of cessation or continuation of these during skin surgery is conflicting. There is limited evidence resulting in variations between surgeons regarding patient instructions perioperatively. A recent survey of dermatologists in the United Kingdom demonstrated that 78% of them advised patients to continue their DOAC as usual, with differences depending on the procedure (Mohs micrographic surgery [MMS] vs simple excision) and the optimal timeframe in which to recommence the DOAC if it was paused.⁶ Of note, 82% of respondents were unaware of a recent update in guidance by the National Institute for Health and Care Excellence (NICE), which stipulated that procedures that have “no clinically important bleeding risk”, such as skin excisions, can be performed either just prior to the next dose (with no interruption to DOAC therapy) or by omitting only a single dose.⁷

Many guidelines and reviews heavily support individualizing one's anticoagulation in the perioperative period on a case-by-case basis or simply recommend their continuation throughout the skin surgery period.^8-10 This may be due to the definition of skin surgery varying between simple skin biopsies, to flaps and grafts, to MMS.

Commonly reported limitations of reviews in the field are a paucity of studies that seek to comprehensively describe the effects of DOAC therapy in skin surgeries.^11-13 The aim of this systematic review was to evaluate the literature on hemorrhagic and thromboembolic complications associated with all perioperative DOAC management in those undergoing skin surgery.

Materials and methods

Results

Across the three databases, 7,741 abstracts were identified. Individual screening of the titles and abstracts by authors PI & LB revealed 35 articles for further full-text analysis. Further screening of the full texts of these articles by PI identified 13 articles that met inclusion criteria and were subsequently eligible for analysis and synthesis (Figure 1). Of the 13 articles, four were prospective observational cohort studies, one was a retrospective case–control study, and the remaining eight were retrospective observational cohort studies (Appendix S1).^15-27

Over 12 years, there were 10,750 patients who underwent cutaneous surgery, with 1,312 of them performed on patients prescribed a DOAC. The mean participant age ranged from 68 to 78.8 years of age, with a male predominance (49.9–81.8%). The most common indication for anticoagulation with a DOAC was atrial fibrillation. The most common disease being treated was basal cell carcinoma (BCC) and, if location was stated, most commonly involved the head, face, and/or scalp. MMS was the most common cutaneous procedure featured in the relevant articles (7/13 studies), followed by direct excision with a variety of methods of closure (direct closure, healing by secondary wound intention, flaps, and grafts). Complexity and duration of procedures were not described in any included articles.

Given that all the literature eligible for inclusion were observational cohort studies, the Newcastle-Ottawa Assessment tool was used to appraise the risk of bias. The majority of studies had a low-medium risk of bias, with two having a high risk of bias.^{18, 20} Meta-analysis was not feasible due to the significant heterogeneity in methodologies and outcomes.

The included articles studied four areas: (1) complication rate in those who continued DOAC therapy,^{15, 16} (2) complication rate in those who ceased DOAC therapy prior to undergoing skin surgery,^{17, 18} (3) complication rate between patients who had their DOAC therapy ceased vs continued,¹⁹ and (4) comparing complication rate in patients anticoagulated with a DOAC agent vs a control group not on a DOAC agent (Table 1).^20-27

Outcomes

DOAC vs. control

Eight of the 13 articles evaluated the rate of hemorrhagic complications in patients undergoing skin surgery (Table 1), comparing those on a DOAC with a control group (Figure 2).^20-27 Studies with non-events were excluded from graphical comparison.

Discussion

This systematic review analyzed 13 articles evaluating 10,750 patients who underwent cutaneous surgery. Of these, there were 1,312 procedures performed on patients who were prescribed a DOAC. As a whole, the evidence surrounding the perioperative management of DOACs in patients who undergo skin surgery offered no single unifying recommendation.^15-27 Many authors deemed the continuation of DOAC therapy throughout the procedure to be safe or that there was no increased risk of hemorrhagic complications.^{15, 16, 23, 25} In contrast, various authors found that interrupting DOAC therapy conferred little thromboembolic risk while minimizing hemorrhagic complications.^17-19 Two studies concluded that DOAC therapy was associated with a significantly increased risk of bleeding when compared to a control group.^{24, 26} Many of the authors go on to describe that given the small sample size of patients on DOAC therapy or those who experienced complications, limited conclusions could be derived from the studies and that further research is required for more definitive recommendations.^{16-18, 25}

The optimal management of a patient's anticoagulation is a difficult balance for the clinician to perfect; on one hand, arterial or venous thromboembolic complications are a very real and devastating concern. Conversely, continuing one's anticoagulation unnecessarily perioperatively can predispose to serious sequelae, including hemorrhage, poor wound healing, and infection.²⁸ In the skin surgery setting, this can lead to poor cosmetic outcomes, dehiscence, and tissue necrosis requiring extended antimicrobial regimes and take-back procedures.^{28, 29}

There exists a paucity in the literature of RCTs that evaluate the periprocedural management of DOACs in skin surgery, and hence there were none eligible for inclusion. Several studies have addressed the issue from a broader perspective, providing recommendations for use in varying clinical contexts based on findings derived from other fields of medicine.^{30, 31} Unfortunately, they often provide little stratification for the bleeding risk of dermatologic specific procedures.

Skin procedures are diverse and heterogeneous in nature, ranging from small biopsies and simple excisions to reconstructive plastic surgeries requiring large flaps and full-thickness grafts which may pose a higher risk of bleeding. Despite this, the vast majority of cutaneous procedures are defined as those posing a minimal or low bleeding risk, described as a 0–2% chance of incurring major bleeding, commonly defined as bleeding into a critical anatomic site, lowering hemoglobin by ≥20 g/l or requiring ≥2 units of packed blood cell transfusions.³²

In light of the literature, we recommend an approach that considers these issues in the perioperative period: (1) What is the inherent bleeding risk associated with the procedure? and (2) What is the thromboembolic risk in this patient?

For procedural bleeding risk in skin surgery, the authors recommend the following approach detailing two levels of bleeding risk. The first level includes minor dermatologic procedures such as cutaneous biopsies (excluding full-thickness excision), excision with primary side-to-side closure, and procedures with expected duration <30 minutes. The second level includes procedures where advanced closure is required, including grafts and flaps, or where duration is >30 minutes.

Secondly, determining the thromboembolic propensity of the patient is vital in underpinning optimal perioperative anticoagulation management. Patients can be defined as low, moderate, or high risk of thromboembolism (Table 2).³¹

Based on bleeding and thromboembolic risk, the authors propose recommendations based on these two levels and subsequent perioperative management of DOACs (Figure 3).

Please note that following this decision tree is to be used as a guide only and should not supersede clinical gestalt. If the decision to withhold anticoagulants is made, counseling the patient on the increased risk of thromboembolic disease must be carried out. In addition, the review undertaken and the recommendations made are only applicable to low-risk dermatologic procedures involving DOACs. This decision process should not be applied to any other type or procedure or any other antithrombotic agent due to differences in pharmacokinetics and bleeding risk. The decision to continue or interrupt one's DOAC regime based on clinical decision tools needs to be validated and ideally further studied in RCTs.

It was the view of the authors to include all types of skin procedures (ranging from minor biopsies to full-thickness skin grafts) that have served as delineation points in other reviews, so as to capture all relevant evidence. Subspecialist procedures relating to the skin that were not covered by skin surgeons, such as reconstructive oculoplastic procedures performed by ophthalmologists, were deemed less relevant to the skin surgeon and were thus excluded from the review. Of note, however, it was recently reported that there are “no studies investigating the safety of continuing DOACs perioperatively in patients undergoing oculoplastic and lacrimal surgery”.³³

Many studies included did not have a sufficient sample size of patients receiving DOAC therapy to amount to statistical power,^{16-18, 20, 25} and either did not have a comparator control cohort^15-18 or had a control cohort without clinician blinding.^{21, 22, 27} Often, there was very little definition provided for hemorrhagic complication and the level of intervention required to achieve hemostasis, while only one study attempted to stratify patients based on the dose of their DOAC (low-dose vs high-dose apixaban and rivaroxaban).¹⁹ Only one study attempted to account for confounders associated with hemorrhagic or thromboembolic complications (e.g., smoking, comorbidities, procedure, and closure type).¹⁹ Studies could serve to benefit from detailing complexity of procedure and procedure length, as while patients who were on DOACs may not have experienced postoperative hemorrhagic complications, procedural length may have increased in order to achieve intraoperative hemostasis.³⁴ Direct comparison of outcomes was difficult across studies given the heterogeneity in reporting outcomes and measured parameters.

In studies nonspecific to dermatological or cutaneous surgeries, often no subgroup analysis of cutaneous procedures was performed or reported, meaning that any complications and conclusions made were extrapolated based on findings of procedures deemed to be of a similar risk profile.¹⁹ The methodology implemented for the identification and reporting of complications in the postoperative period was extremely heterogeneous and subject to bias and imprecision. While some studies had a follow-up organized at a specified time with the reporting clinician, one study relied on patients to re-present within 3 weeks if they noticed a complication.²¹ The majority of the procedures were MMS and skin cancer excisions performed on either the head, face or scalp, meaning that any conclusions drawn should only be applied to similar clinical contexts.

Indeed, in the absence of high-quality evidence or consistent international guidelines providing a consensus on ideal management of DOAC therapy in cutaneous surgery, several surveys of dermatologist practice undertaken globally reflect a trend toward individualized approaches. Chiang et al. surveyed 67 dermatologists across the UK with respondents generally recommending to their patients the continuation of DOAC therapy pre- and postoperatively in low bleeding-risk procedures such as minor skin excisions.⁶ Logically, as the size of the excision or complexity of the surgical procedure increased, the proportion of dermatologists recommending uninterrupted DOAC therapy decreased. Similarly, a more recent survey of over 250 German dermatological surgeons found the vast majority (>70%) continued rivaroxaban in the setting of small (e.g., nevus) and large (e.g., melanoma) skin excisions with practices remaining consistent between hospital and office-based clinicians.³⁵ For larger excisions and sentinel lymph node biopsies (procedures only performed by the hospital-based dermatologists surveyed), the proportions of those continuing DOAC therapy was reduced to between 15 and 30%. Han et al. found dermatological surgeons were more likely to continue DOAC therapy when compared with plastic and reconstructive surgeons in a survey of US-based clinicians performing cutaneous surgery, likely reflecting the increased complexity and invasivness of cutaneous surgery performed by plastic surgeons.³⁶

While there is no unifying recommendation, guidelines lean in favor of continuing DOACs in skin surgery in the majority of instances. Given this, one has to ask the question as to why many clinicians continue to withhold DOACs in skin procedures. As previously described, DOACs have been in one's armamentarium for as little as 10 years, suggesting that the prejudice of dermatologists to discontinue these agents may be based on their familiarity with medications such as warfarin and aspirin. Some studies document that the risk of hemorrhagic complications in patients taking warfarin can be up to three-fold higher than their non-anticoagulated counterparts, warranting consideration for interruption perioperatively.³⁷ Efforts need to be made to reassure dermatologic surgeons that minor skin procedures can be performed in patients without changes to their DOAC regime, especially in those with higher thromboembolic burden.

Conclusion

The evidence characterizing the literature of anticoagulants in skin surgery is conflicting, often differing based on the antithrombotic agent and the nature of the procedure. This systematic review largely found that either the continuation or cessation of DOAC therapy in the perioperative period was safe, with minimal complications of hemorrhage or thromboembolism, often comparable to the general population.

The authors present a set of recommendations based on two levels of skin surgery, defined as simple or those requiring advanced closure. Those performing skin surgery should feel reassured that continuing a patient's DOAC will be safe and practical in the majority of procedures. Future studies validating this approach would serve to benefit the sphere of literature on perioperative DOAC management.

Author contributions

All authors satisfied the ICMJE criteria for authorship. Patrick Ireland – conceptualization, selection criteria, data screening, data collection and synthesis, manuscript drafting, manuscript editing. Luca Borruso – conceptualization, data screening, data collection and synthesis, manuscript drafting. Sascha Spencer – conceptualization, manuscript editing. Richard Rosen – conceptualization, manuscript editing. Robert Rosen – conceptualization, selection criteria, manuscript editing, supervision. All authors were involved in the organization, coordination, conduction, critical review of the manuscript, and approval of the final version.