Recent highlights on specific aspects of oral anticoagulation in difficult clinical scenarios from the International Journal of Cardiology Heart & Vasculature

In our series of spotlight reviews we highlight a selection of studies that consider specific aspects of oral anticoagulation in challenging clinical settings.

The global COVID-19 epidemic that erupted at the end of 2019 is now endemic and firmly entrenched as a comorbidity to be considered in patient care. Infection with the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has emerged as a new trigger of thromboinflammatory disease that can have long-term impact on existing and developing clinical conditions and the response to therapeutic intervention [1], [2], [3].

Elseidy et al. [4] provide a comprehensive overview of the constellation of cardiovascular complications of the post-acute COVID-19 syndrome, its temporal nature and the underlying mechanisms. The authors describe the distinct features of SARS-CoV-2 in terms of the acute and prolongued signaling and sequelae profiles compared to infection with other RNA viruses and delve into possible therapeutic management options. Anticoagulant therapy is a predominant topic in this context, given the causal link between virus infection and thromboembolism (TE). The report concludes with a recommended management map for patients with post-covid TE complications: (i) low molecular-weight heparin during hospitalisation, (ii) parenteral unfractionated heparin to be applied in cases of hemodynamic instability, morbid obesity or renal dysfunction; (iii) direct oral anticoagulants (DOAC) during the acute and recovery phases if hemodynamically stable and free of valve disease, with renal function being monitored in case of drug interactions with specific COVID-therapies. The efficacy of vitamin-K antagonist (VKA)-based therapy has not been verified and requires further study.

Anemia is another condition in which anticoagulation therapy is considered challenging. A new study by Tanaka et al., [5] however, indicates that there is no need to withhold guideline-conform OAC in patients with various grades of anemia. The authors examined findings from the J-ELD AF Registry, a multicenter prospective observational study, looking into one-year outcomes after on- label reduced dose dose apixaban in Japanese patients with atrial fibrillation (AF), aged ≥ 75 years. The patients were stratified according to mild, moderate-severe and no anemia. Moderate-severe anemia was identified to be an independent risk factor for total death (HR approximately 2.2), but not for strokes or systemic TE. In terms of bleeding, a greater risk was noted with moderate-severe anemia, but this was no longer significant in the adjusted analysis. In settings of severe anemia, where OAC are preferentially withheld, left-atrial (LA) appendage closure (LAAC) might provide an alternative means for stroke prevention, but given the lack of robust evidence, this approach is second-line only. Kany et al. [6] report on the prospective, multicentre observational Left-Atrium-Appendage Occluder Register-GErmany (LAARGE), and compared procedural safety and one-year outcomes of LAAC in patients with and without severe anemia. In-hospital safety outcomes – a composite of death, myocardial infarction (MI) and stroke – were similar between the groups, but those with severe anemia had a lower rate of other severe peri-procedural complications. At one-year follow-up, again mortality and TE-related event rates were comparable between the groups, but sub-analysis identified more severe bleeding in the non-anemic patients. Numerically, hazard ratio for mortality stroke and systemic TE did increase, but this did not reach statistical significance. The authors conclude with the reassuring message that pre-procedural anemia does not seem to impact detrimentally on procedural safety, but that a higher one-year mortality – driven mainly by co-morbidities - warrants consideration for follow-up management.

Because of lack of sufficient evidence, Nso et al. [7] pursued the question if occlusion of the LAA could completely replace OAC as an alternative for stroke risk prevention. The authors provide a systematic review and meta-analysis on the topic, comparing outcome in patients undergoing LAAC versus no LAAC, or DOAC, or VKA. Compated to no stroke preventative therapy at all, LAAC significantly reduces risk of stroke and major cardiovascular adverse events (MACE), with non-significant but numerically notable reductions in stroke and all-cause mortality. The incidence of post-operative AF is also reduced albeit non-significantly, while outcome in terms of other post-operative complications is unclear. Bleeding overall was not significantly different between LAAC vs. no LAAC. When comparing LAAC with DOAK or VKA the authors could found overall equivalent safety and efficacy with a trend to better outcome after LAAC, but the data are highly variable and the feasibility of complete replacement remains debatable.

Insight into the comparative efficacy of OAC strategies in patients with AF undergoing a cardiac implantable electronic device (CIED) procedure is provided by de Heide et al. [8]. Although CIED is generally considered to pose a low bleeding risk, a common complication in up to 16% of patients is a so-called device-pocket hematoma. This is per se not dangerous, but causes local discomfort, an increased risk of infection, prolongation of hospitalization and may lead to requirement for surgical intervention. For patients maintained on VKA, discontinuation with peri-procedural heparin bridging increases the bleeding risk by 5-fold versus continued VKA therapy, so this latter approach is now established and guideline-conform. For patients on DOAC, data on peri-procedural management are more obscure and the guidelines accordingly show a gap. Some evidence suggests that interrupted DOAC may reduce bleeding risk compared to continued DOAC therapy. The present study provides real-world data comparing continued VKA versus interrupted DOAC. Neither group developed a TE event, indicating sufficient supression of coagulation. Of note all documented haematomas occured in the VKA group, corresponding to an incidence of 2.5% vs. 0% in the interrupted DOAC group, although this difference was not statistically significant.

Paukovitch et al. [9] assessed outcome after transcatheter mitral valve repair. Of all patients, 13% suffered from a bleeding event, of which nearly half was site-related bleeding (36 % gastrointestinal and 18 % undefined). Interestingly, patients who developed bleeds were less likely to have been receiving DOAC prior to surgery (26% vs 42% who did not receive DOAC). A careful look at Table 2 of the study shows that the group WITH bleeding events had previously been prescribed DOAC to a lesser degree: specifically bleeding was – counterintuitively - less common in the group represented by those more frequently prescribed DOAC [9]. The authors speculate that a pre-existing high bleeding risk might have precluded OAC, and caution peri-procedural anticoagulation in this patient subset.

The choice of anticoagulant therapy is also important in settings of heart failure (HF). Koike et al. [10] note that manifest left ventricular (LV) diastolic dysfunction prior to ablation for AF worsens prognosis outcome, regarded as a composite of all-cause death, HF hospitalization and worsened HF symptoms requiring escalation of decongestive therapy. The group developing the primary outcome had a more frequent history of prior stroke, higher CHA₂DS₂ -VASc scores (2 vs. 1), were more frequently prescribed warfarin (64 vs 43 %). but less frequently DOAC (36% vs 57%). Although not identified as an independent predictor of the composite outcome, numerically the data might cautiously suggest that DOAC therapy is the best therapy in patients undergoing AF ablation who also exhibit diastolic dysfunction.

Bungo et al. [11] assessed patients with malignancy, who are highly prone to both TE and bleeding, as well the development of AF. This risk constellation arises from both hemostastic effects of cancer per se as well as from chemotherapy. The authors assessed a study cohort from the Surveillance, Epidemiologic, and End Results (SEER) database, a cancer registry in the United States that covers 35% of the population. Half of the cohort had diagnosed lung, colon, prostate or breast cancer, the other half were disease-free; patients with embolic or ischemic history were excluded, as were smokers. The study if incorporating cancer in the current CHA₂DS₂VASC score may improve stroke prediction in these patients. The implication is that a substantial number of patients who under the traditional score do not qualify for OAC, should in fact be anticoagulated. A small but by no means negligible 3.1% subset of cancer patients with concurrent AF had a CHA₂DS₂VASC score of ≤1, and would not be offered OAC. If the presence of cancer was specifically considered in the risk score, these patients might shift from a low-risk group to those who require stroke prevention, a concept that warrants expert discussion and clinical validation.

Risk scores are cornerstone aids for individualized initiation, duration and intensity of antithrombotic therapy. A novel concept for personalizing AF management is provided by Suzuki et al. [12] and Watanabe et al. [13] Both author groups apply hierarchical cluster analyses to identify subsets of patients with special needs as therapeutic management of AF, based on characteristic clinical bundles. This approach warrants further consideration, given the important impact on risk and outcomes. With specific relevance for antithrombotic therapy, Teppo et al. [14] describe a novel, to date largely underestimated, factor to be considered in personalized risk assessment: the socioecomomic status of the patient. The nationwide retrospective registry-based FinACAF cohort study encompasses all patients with AF from all levels of care in Finland, with the majority (around 70%) initiating OAC during follw-up. Disparities in income and educational levels are accepted to be linked with risk of stroke and death, but the authors now also noted a socio-economic component in the risk of adverse bleeding. Low income, more so that a low educational level, was independently associated with hazard of any bleeding as well as gastrointestinal and cerebral bleeding. The authors carefully dissect the possible reasons for this and specifically discuss the limitations of the study. Of note, clear income-related bleeding disparities were observed even after controlling for an extensive set of patient characteristics indicating that the bleeding risks associated with low income are not fully explained by the factors incorporated in the conventional bleeding risk stratification scores. Further on the topic of improved risk assessment, Chen at al. [15] introduced and validated a novel scoring system to predict stroke risk in AF. The authors actually begin with a note of caution: because information regarding AF subtypes, recurrence rates and methods of AF ablation were not available in the validation cohort, AF-related admissions were analysed as a surrogate for AF recurrences. The rationale for their overhaul of the risk score is that the long-term antithrombotic efficacy of rate or rhythm control in AF reaches a ceiling over the long-term. The authors criticize that current conventional stroke risk scoring systems CHA₂DS₂ and CHA₂DS₂-VASc leave a gap in this regard, since they do not account for actual age stratification and catheter ablation status at the time of the procedure. A new “AF-CA” score is proposed, which in the validation cohort disoplays a higher decimation ability to predict 1 year, 5 year and 10 year incident stroke than the conventional scores. Further, around 15–20% of ablated patients who were classified as carrying a moderate-to-high-risk based on CHA₂DS₂-VASc scores, would instead be designated as low-risk according to the AF-CA-Stroke score. The authors conclude that a larger fraction of patients could safely discontinue OAC after successful ablation than the conventional scores identify.

Danneberg et al. [16] report on the scoring system used to predict bleeding risk under dual antiplatelet therapy (DAPT). The authors examined if PRECISE-DAPT can also predict ischemic events and mortality over one-year follow-up after percutaneous coronary intervention (PCI), and in fact identify a potent discrimination capacity for MACE, especially ischemic events, with nearly identical cut-offs as for bleeding. The authors did not specifically address the question of concurrent anticoagulant therapy, but do in the limitations discuss the fact that a large proportion of patients require triple therapy, encompassing OAC on top of DAPT, which may (ind)directly affect platelet reactivity. Thrombin is the most potent platelet activator, and some DOAC like dabigatran can alter thrombin receptor expression on platelets [17]. DOAC-induced modification of this thrombin/receptor axis will conceivably skew platelet function and DAPT outcome. Drug interactions between DOAC and other medications commonly used in AF may also arise at the level of drug transport and metabolism. All DOAC are substrates for P-glycoprotein, and the FXa inhibitors are strongly metabolized by hepatic CYP enzymes. Both P-glycoprotein and CYP3A4 are to differing extents inhibited by common antiarrythmic drugs, potentially affecting DOAC pharmacokinetics. Shurrab et al. [18] now highlight how frequent this problem is likely to occur, by examining overlapping prescriptions for DOAC and either amiodarone and diltiazem. In a cohort of AF patients aged > 66 years, co-prescription for DOAC and amiodarone or diltiazem was identified in around 6% and 11% of patients, in whom bleeding risk should be monitored more carefully.

Bleeding risk is especially high in patients taking combined OAC and DAPT. Thus safety of drug combinations along with duration of individual therapy are key aspects to consider carefully. Wang et al. [19] report a meta-analysis and systematic review that seeks to examine how to best treat patients who develop AF within one year after PCI. The authors specifically considered subgroups with different risk profiles for thrombosis and bleeding, and essentially provide a roadmap for regional guidelines. In summary the authors identify four treatment strategies according to existing risk and prior history: (i) dabigatran low-dose (110 mg bid) + P₂Y₁₂ inhibitor for TIMI major or minor bleeding risk, (ii) FXa inhibitor + P₂Y₁₂ inhibitor for ISTH major bleeding risk, (iii): escalated dabigatran (150 mg bid) Xa inhibitor + P2Y12 inhibitor for risk of stroke plus all-cause death, (iv) VKA + P₂Y₁₂ inhibitor for risk of MI and stent thrombosis. The choice of aspirin rather than a P₂Y₁₂ inhibitor in this context remains controversial. Almas et al. [20] systematically reviewed the safety and efficacy of combining aspirin with DOAC in AF, since this therapeutic strategy has in recent years been replacing the traditional combination of VKA + aspirin. The authors assessed 2 randomised clinical trials and 7 observational studies, including COMPASS, AFIRE and PIONEER-AF-PCI. The key take-home messages were: (i) aspirin on top of DOAC significantly increases major bleeding, (ii) provides no superior prevention of MACE, and (iii) upon sensitivity testing, actually appears to increase stroke rate compared to monotherapy, although statistically there was no difference in terms of stroke or hospitalization, death, ischemia and MI. The authors critically discuss the implications and limitations of their analysis, and conclude that in the non-acute setting, the decision to add aspirin to the antithrombotic management of AF patients needs cautious assessment and consideration.

A clinical study in the pipeline might help to resolve the dilemma of best-practice antithrombotic therapy in these dually afflicted patients, who have to date been largely examined in restrospective substudy analyses. Riesinger et al. [21] outline the rationale and design details of the Apixaban versus PhenpRocoumon: Oral AntiCoagulation plus antiplatelet tHerapy in patients with Acute Coronary Syndrome and Atrial Fibrillation (APPROACH-ACS-AF) trial (NCT02789917), a prospective randomized parallel-group, open-label, blinded-endpoint, superiority, multicenter-trial of triple versus dual therapy in patients with AF and ACS. This study will be the first trial dedicated to ACS patients, testing whether a DOAC-based dual approach is superior to a VKA-based triple approach in high-risk ACS patients with AF in terms of bleeding. In the context of combined APT and anticoagulant therapy however, Fiocca et al. [22] raise the important issue of adherence to therapy. The Italian group assessed cessation patterns and causes of ticagrelor discontinuation in the AD-HOC (ADHerence of ticagrelOr in real-world patients with aCute coronary syndrome) cohort, and identified OAC as an independent predictor of premature ticagrelor cessation. The hazard ratio for premature discontinuation associated with OAC was nearly 11, much higher than that associated with advanced age > 80 years (3.3) or female gender (1.6). As mentioned above, fear of bleeding is likely the most critical driver of inappropriate therapy cessation, although in this cohort the increased risk of new ischemic events also played a major role.

The most feared forms of adverse bleeding to occur under oral anticoagulant therapy is intracranial hemorrhage (ICH). Hesitancy to reinitiate OAC after such a devastating event is understandable, both from both the physician’s and patient’s perspectives. Given the absence of clear guidelines in this setting, the decision if, when and how to recommence OAC is highly variable. The Korean study by Moon et al. [23]. comprehensively reviewed data on ICH survivors with AF, who resumed DOAC, VKA or APT, or alternatively did not reinitiate any antithrombotic therapy. Study enpoints were severe thrombotic and bleeding events. Compared to non-users, both OAC approaches and APT showed significant benefits in terms of TE or hemorrhage after adjustment, with no discernible difference between the different medications. The take-home message is that recommencing antithrombotic therapy at 6–8 weeks after the index ICH substantially lowers the risk of all-cause mortality and severe TE, while the adjusted risk of hemorrhage is higher if antithrombotic therapy is initiated earlier, and that DOAC recipients fare best. The apparently lower bleeding risk with re-initiated antithrombotic therapy compared to non-users (approximate HR for OAC or APT use vs. non-use 0.6) is curious and warrants deeper interpretation and discussion than provided in the report. The study data do not inform why antithrombotic therapy was withheld in the non-users. The patient characteristics in fact show that non-users had a consistently lower history of a wide spectrum of cardiovascular, thrombotic, metabolic and neurodegenerative diseases, so there is actually no obvious clinical reason why antithrombotic therapy was withheld.

Oral anticoagulation remains a critical cornerstone for stroke prevention. Comorbidities and polypharmacy complicate patient management by further skewing thrombotic and bleeding risks. This selection of studies recently published in the International Journal of Cardiology Heart & Vasculature highlight some of the challenges and gaps that warrant further expert discussion and clinical validation, and provide novel avenues for improved risk assessment and patient care.

Disclosures

None.

Sources of funding

Deutsche Forschungsgemeinschaft (DFG, FE 1365/4-1, to A.F.), National Institutes of Health (R01-HL131517, R01-HL136389, R01-HL089598, R01-HL163277, and R01-HL160992 to D.D.), and European Union (large-scale integrative project MAESTRIA, No. 965286 to D.D).