Fig. 1.1
Rate of stroke and systemic embolism in the five large studies on NOAC. Only AVERROES compared NOAC to ASS again showing platelet inhibition not to be an efficient stroke protection in patients with AF similar to earlier trials comparing warfarin and platelet inhibition like ACTIVE-W. All other studies compared NOAC to warfarin. Very similar data confirming these results are now available from large real-world registries on dabigatran, apixaban and rivaroxaban
Fig. 1.2
Rate of intracranial bleeding in the five large studies on NOAC. Only AVERROES compared NOAC to ASS. Interestingly, this study finds aspirin to have the same bleeding rate as apixaban in a low-risk population. All other studies compared NOAC to warfarin
Fig. 1.3
Major bleeding rates in large NOAC trials. Patients with a history of major bleeds within a certain time frame were excluded in some trials. While rivaroxaban had no advantage over warfarin, other NOAC’s had a statistically significant reduction in bleeding in this low-risk population (dabigatran 2 × 110 mg, apixaban, edoxaban). Similar data confirming these results are now available from large real-world registries on dabigatran, apixaban and rivaroxaban
Fig. 1.4
Matched pair analysis on real-life data from the US medicare registry on dabigatran 2 × 150 mg compared to warfarin therapy in 2 × 67,202 patients. The study finds similar results to the randomized trials with less intracranial bleeding yet slightly increased gastrointestinal bleeding. The rate of ischemic stroke is significantly reduced [10]
The only trial comparing a NOAC to platelet inhibition is discussed here in further detail: the AVERROES trial included 5591 patients randomized 1:1 to apixaban 2 × 5 mg or aspirin (80–320 mg/day; 91% of patients received between 81 and 162 mg aspirin). Mean CHADS2 score was 2.0 ± 1.1 vs 2.1 ± 1.1, respectively. 81% of patients were included due to “the assessment that INR could not or was unlikely to be measured at requested intervals; uncertainty about patients ability to adhere to instructions regarding OAT, expected difficulty in contacting patient for urgent change in OAT dose or CHADS2 score of one and VKA therapy not recommended by physician”. A serious bleeding event in the 6 months prior to study randomization or a high risk of bleeding were key exclusion criteria. 35% of patients were recruited within Western Europe and North America, while 65% were recruited from Latin America, Eastern Europe and Asia/South Africa. As expected, apixaban reduced the number of strokes significantly compared to aspirin (1.6 vs 3.4% per year, respectively) (Fig. 1.1). Bleeding events were statistically not different between aspirin and apixaban in this trial (Figs. 1.2 and 1.3). Two different definitions were employed: while “major” bleeding occurred at a rate of 1.4 and 1.2% per year, “clinically relevant non-major” bleeding occurred at 3.1 vs. 2.7% per year, respectively. These bleeding rates as well as the low CHADS2 score indicate that the study included low risk patients. The patients included in this study are therefore not representative of patients deemed to be unsuitable for oral anticoagulation in Western Europe or the US, where most patients are deemed to be ineligible for NOAC due to increased bleeding risk while INR measurements are readily available. Triple therapy combining dual antiplatelet and oral anticoagulation is a specific issue; it is now clear from various registries that there is no benefit in this treatment even regarding stent thrombosis while the bleeding risk is excessive. Combining clopidogrel and oral anticoagulation without aspirin appears to come with the lowest bleeding risk and the best outcome (Fig. 1.5).
Fig. 1.5
Rate of myocardial infarction/death, stroke, bleeding and all-cause death during follow-up in 11.480 patients with atrial fibrillation and an acute coronary syndrome. Triple therapy was set at a hazard risk of one, other therapeutic options were compared. Patients on clopidogrel + OAC had the best outcome. Source: M. Lamberts et al., J Am Coll Cardiol. 2013 Sep. 10;62(11):981-9
1.4 Risk of Bleeding with Oral Anticoagulation
Many general practitioners fear the risk of bleeding and other side effects when prescribing oral anticoagulation. Several surveys indicate that prescription of oral anticoagulation is overused in younger patients with AF and severely underused in elderly patients. This is dramatic since several large studies find both the risk of stroke as well as poor clinical outcome after stroke to increase with age; this paradigm is not altered by new treatment concepts for acute stroke including thrombolysis [11]. In conclusion, stroke prevention in patients with AF is in need of new treatment options since the currently available therapies face insurmountable practical hurdles that have only partially disappeared following the introduction of NOAC’s. To calculate the risk of bleeding, the 2010 ESC guidelines have introduced the HAS-BLED score based on a real-world patient cohort from the EuroHeart Survey on AF from 2005 [1]. This risk score has been derived from a patient cohort treated with warfarin. No larger studies have validated the HAS-BLED with NOAC therapy yet, however a substudy of the ARISTOTLE trial finds bleeding to increase with higher HAS-BLED score [12]. The randomized trials find extra-cranial bleeding events (i.e. gastrointestinal, bladder, skin) not to be reduced with NOAC therapy (Fig. 1.3).
Looking at NOAC prescription data in 2016 in Europe, dose reduction of NOAC therapy is performed in around 50% of patients. As the effect of NOAC’s is dose dependent, dose reduction will prevent bleeding but will also reduce the effectiveness of stroke prevention. Dose reduction is clearly defined with each NOAC; in general it should be limited to patients with reduced kidney function (usually GFR < 45 ml/min), low weight (<60 kg) and/or age >80 years. Dose reduction is comparible to the strategy of keeping patients at an INR of 1.8 at which increased numbers of stroke do occur according to large registry data [23]. This is therefore a group of patients in which interventional stroke prevention with a LAA occluder should be performed -similar to patients with a lack of compliance or concomitant long term platelet inhibition following acute coronary syndrome events (Fig. 1.6).
Fig. 1.6
Algorithm based on the NOAC-studies as well as the recent prospective studies on Watchman LAA occlusion. A large retrospective registry for the Amulet LAA occluder also supports this approach. LAA occluder therapy allows for stroke protection to the same degree as warfarin. Currently LAA occluder therapy is recommended for all patients that have a relative or absolute contraindication for warfarin and NOAC’s at full dose. LAA occluder is followed by 3 months of dual antiplatelet therapy in most European centers
1.5 Stroke Prevention in AF Patients Following Acute Coronary Syndromes
Patients with both acute coronary syndromes and atrial fibrillation are a specific group of patients with currently unclear treatment algorithms. Triple therapy (aspirin, clopidogrel, oral anticoagulation) was repeatedly shown to lead to unacceptable high bleeding rates (Fig. 1.5) also increasing clinical endpoints and mortality [13, 14]. The APPRAISE-2 study compared 2 × 5 mg apixaban on top of dual platelet inhibition after acute coronary syndrome to placebo and had to be stopped due to increased risk of bleeding including fatal and intracranial bleeding (1.3% vs. 0.5%, respectively) [15]. In contrast, the ATLAS ACS 2—TIMI 51 trial found a much lower dose of rivaroxaban (2 × 2.5 mg per day as compared to 1 × 20 mg used for stroke prevention) to be beneficial on top of dual platelet inhibition post-ACS with an acceptable increase in major bleeding (2.1 vs. 0.6% per year, respectively) but no increase in fatal bleeding [16]. Importantly, only patients with sinus rhythm were included in that trial; the low dose of rivaroxaban would not have been efficient for stroke prevention. Many centers in Europe have started to use the combination of clopidogrel and oral anticoagulation including NOAC. Several randomized trials analyzing NOAC therapy in combination with platelet inhibition are ongoing. However, the best treatment for secondary prevention of acute coronary syndromes remains the combination with new antiplatelet drugs such as prasugrel and ticagrelor. These drugs confer an even higher bleeding risk when combined with warfarin. LAA closure combined with dual antiplatelet therapy including i.e. ticagrelor or prasugrel long-term is therefore an ideal option for these patients [17] (Fig. 1.7).
1.6 Indications for LAA Occlusion in 2017
The ESC (European Society of Cardiology) guidelines on stroke prevention in AF 2016 recommend LAA occluder therapy for all patients with relative or absolute contraindication for oral anticoagulation; there is no room for platelet inhibition or reduced NOAC dose treatment (Fig. 1.8). The new oral anticoagulants show an improved efficacy and safety profile regarding stroke prevention in AF compared to warfarin. INR measurements are obsolete, which is an advantage over warfarin therapy. Due to the reduced half-life of 12 h, bridging with low-molecular weight heparin for elective surgery is not required. However, patients with bleeding events and patients in need of platelet inhibition are still in need of alternative therapies. The main reason not to start a patient on oral anticoagulation remains the risk of clinically relevant bleeding which also affects quality of life; a point especially relevant to the elderly. Comparing the risk of clinically relevant bleeding in the available trials demonstrates the current dilemma in that also the new anticoagulants come with a risk of 2–3% bleeding events per year (Fig. 1.3). Moreover, most studies have only included patients that were also eligible to warfarin therapy and have excluded patients with contraindications for oral anticoagulation or at “high” risk for bleeding events, i.e. recent clinically significant bleeding events [18–20]. The HAS-BLED score of patients included in these trials is not available, yet given the fact that the CHADS2 score was low with the exception of ROCKET-AF (CHADS2 score 3.4) it can be estimated to be around 2. Increased bleeding with an event rate >3% per year starts with a HAS-BLED of ≥3. This patient cohort is therefore not addressed through the current NOAC studies. Similarly, optimal treatment for the elderly >75 years remains unclear. Sub-analysis of the RE-LY trial shows a significant treatment-by-age interaction, in that the benefit of the low dose dabigatran scheme (2 × 110 mg) regarding major bleeding events was eliminated in patients >75 years. More precisely, dabigatran 2 × 110 mg vs. warfarin carried a risk of 5.1% vs. 4.4% major bleeding per year in these elderly patients. Major bleeding was reported at a rate of 3.3% vs. 5.2% per year comparing apixaban and warfarin in patients aged >75 years, respectively [19].
