Current Neurology and Neuroscience Reports

, Volume 11, Issue 1, pp 15–27

Stroke Prevention Treatment of Patients with Atrial Fibrillation: Old and New


  • Simerpreet Bal
    • Departments of Clinical Neurosciences, Calgary Stroke Program, Faculty of MedicineUniversity of Calgary
  • Pawan Ojha
    • Department of NeurologySir J.J. Group of Hospitals
    • Departments of Clinical Neurosciences, Medicine and Community Health Sciences, Faculty of MedicineUniversity of Calgary

DOI: 10.1007/s11910-010-0161-z

Cite this article as:
Bal, S., Ojha, P. & Hill, M.D. Curr Neurol Neurosci Rep (2011) 11: 15. doi:10.1007/s11910-010-0161-z


Atrial fibrillation is the most common cause of cardioembolic ischemic stroke and has a rising prevalence worldwide. Stroke prevention in this condition is poised to take a substantial leap forward with the evolution of new anticoagulant medications, with superior properties compared to vitamin K antagonists. New, safer and more effective chronic therapy is on the horizon. However, many issues surrounding the management of stroke prevention after an acute stroke and during the course of chronic anticoagulant therapy remain to be resolved.


AnticoagulationAtrial fibrillationThrombin inhibitorsStroke

Clinical Trial Acronyms


Atrial Fibrillation Clopidogrel Trial with Irbesartan for Prevention of Vascular Events


Atrial Fibrillation, Aspirin, and Anticoagulation Study


Atrial Fibrillation Follow up Investigation of Rhythm Management


Evaluating the Use of SR34006 Compared to Warfarin or Acenocoumarol in Patients with Atrial Fibrillation


Apixaban for the Prevention of Stroke in Subjects with Atrial Fibrillation


A Trial with Dronedarone to Prevent Hospitalization or Death in Patients with Atrial Fibrillation


Apixaban Versus Acetylsalicylic Acid to Prevent Strokes


Boston Area Anticoagulation Trial for Atrial Fibrillation


Birmingham Atrial Fibrillation Treatment of the Aged


Evaluation of Weekly Subcutaneous Biotinylated Idraparinux Versus Oral Adjusted-dose Warfarin to Prevent Stroke and Systemic Thromboembolic Events in Patients with Atrial Fibrillation


Effective Anticoagulation with Factor xA Next Generation in Atrial Fibrillation


Heparin in Acute Embolic Stroke Trial


Rate Control Versus Electrical Cardioversion for Persistent Atrial Fibrillation


Randomized Evaluation of Long Term Anticoagulant Therapy with Dabigatran Etexilate


Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared with Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation


Stroke Prevention in Atrial Fibrillation


Stroke Prevention in Nonrheumatic Atrial Fibrillation


Stroke Prevention Using an Oral Thrombin Inhibitor in Atrial Fibrillation


Atrial fibrillation (AF) is associated with a four- to fivefold increase in the risk of ischemic stroke, and 15% of all ischemic strokes are caused by AF. Importantly, this proportion increases substantially with age [1]. The prevalence of AF increases from less than 1% in patients younger than 60 years to almost 10% in patients over the age of 80 years. Similarly, the incidence of AF increases from 0.2% per year in men under the age of 40 years to more than 2% per year in men 80 to 89 years of age, with a lower age-adjusted incidence in women [2].

Risk factors for stroke in the setting of AF include age greater than 75 years, congestive heart failure or left ventricular dysfunction, hypertension, diabetes mellitus, smoking, and other structural lesions of the heart [3, 4]. Patients with a prior stroke or transient ischemic attack (TIA) and AF face the highest risk, more than twice that of AF patients without such a history. The task of preventing stroke in patients with “high-risk” AF therefore encompasses secondary stroke prevention as well as primary prevention in patients with one or more stroke risk factors. The risk of stroke among patients with paroxysmal or persistent AF is comparable to the risk among patients with permanent AF; importantly, the absence of symptoms does not imply a lower risk of thromboembolism [5].

The standard therapy to prevent stroke and systemic thromboembolism associated with AF has been anticoagulation with a coumarin derivative (or vitamin K antagonist [VKA]). However, coumarins have a narrow therapeutic index and require ongoing frequent monitoring. Newer antithrombotic agents are destined to provide stroke prevention therapy without the problems inherent in frequent monitoring of the International Normalized Ratio (INR). The aim of this review is to provide an update on newer antithrombotic therapy for stroke prevention in AF.

Classification of Atrial Fibrillation

Abnormal irregular heart rhythm with chaotic generation of electrical signals in the atria of the heart is called AF. Depending upon the frequency, duration, and recurrence, it can be classified as 1) first detected AF with only one diagnosed event of AF; 2) paroxysmal AF if the episodes are recurrent and self-terminate within 7 days; 3) persistent AF if it persists for more than 7 days and does not self-terminate eventually or requires cardioversion; or 4) permanent AF if cardioversion is unsuccessful or it is not attempted, and AF persists for a year or more.

There remain unresolved problems with this classification system. It is uncertain how important the duration of episodes of AF is in predicting thromboembolic risk. Holter monitoring may frequently identify a 20- or 30-second spell of AF. This system does not address cases in which AF is secondary to another condition (eg, myocarditis or thyrotoxicosis) [6]. Therefore, AF has also been categorized according to cause: 1) lone atrial fibrillation implies the absence of clinical or echocardiographic findings of other cardiovascular disease (including hypertension), related pulmonary disease, or cardiac abnormalities such as enlargement of the left atrium, or age less than 60 years [7]; 2) nonvalvular AF implies the absence of valvular heart disease (eg, rheumatic mitral valve disease, a prosthetic heart valve, or valve repair); and 3) secondary AF occurs in the setting of a primary condition that may be the cause of the AF, such as acute myocardial infarction, cardiac surgery, pericarditis, myocarditis, hyperthyroidism, pulmonary embolism, pneumonia, or other acute pulmonary disease.

Stroke Risk Stratification Models

The risk of stroke or systemic embolization is not uniform across patients with AF. Among nonvalvular AF patients, the average absolute risk of stroke is between 3% to 4% per year, but may vary 20-fold depending on age and other clinical features. In a systematic review of risk factors for stroke in patients with AF, four clinical features (prior stroke or TIA, older age, hypertension, or diabetes mellitus) were consistent independent risk factors for stroke and systemic thromboembolism [8]. Echocardiographic factors may also be used to stratify stroke risk. Moderate to severe left ventricular hypokinesis, left atrial thrombus, spontaneous echo contrast, and low left atrial appendage velocity all predict stroke and thromboembolism [9]. Thus, multiple risk stratification models have been developed [10, 11].

The CHADS2 risk stratification scheme, which applies to nonvalvular, persistent, or paroxysmal AF only, is well validated, easy to use, and has the greatest face validity. It uses a 6-point system in which 1 point is allocated for each of congestive heart failure, hypertension, age ≥75 years, and diabetes mellitus and 2 points are allocated for any prior stroke or TIA. The estimated risk of strokes per 100 patient-years in original cohorts as per CHADS2 stratification schemes are low risk (0 points; 1.2% to 3%), intermediate risk (1 point; 2.8% to 4.0%), and high risk (2 or more points; 5.9% to 18.2%) [12]. Other guidelines include the American College of Chest Physicians practice guidelines [13] and American Heart Association/European Society of Cardiology guidelines [14]. The clinical value of these schemes is dependent on their ability to identify patients at high enough risk that the benefit of coumarin-based anticoagulation is greater than the risk. Patients in the CHADS2 high-risk category are generally treated with anticoagulation.

Despite the limited consensus achieved with the CHADS2 score, there remain patient populations in which these risk stratification scores are inadequately discriminative. Factors such as concurrent heart failure and hypertension and how well they are treated may substantially influence the ongoing risk. Further, the safety of coumarins is highly dependent upon meticulous follow-up and systems of care; where these are inadequate, safety cohorts have shown unacceptably high rates of serious hemorrhage [15]. Newer scores, such as CHA2DS2-VASc score, have been developed [16], but have not yet become part of routine use. There remains scope for further work in this area.

Treatment of AF

The goals of treatment are to improve hemodynamics and prevent systemic embolization. Two approaches to treating the arrhythmia have been used: rate control, consisting of simply controlling fast ventricular response, and rhythm control, consisting of chemical or electrical cardioversion and the use of medication to keep the patient in sinus rhythm. Surgical (eg, Maze procedure) and catheter-based approaches (eg, radiofrequency ablation around the pulmonary veins) to rhythm control are effective for rhythm control but they do not protect against cardioembolism. Physiologically, it seems likely that rhythm control is quite important for patients with enough ventricular dysfunction that their diastolic ventricular filling is substantially dependent upon normal atrial contraction.

Two large trials have been conducted in this arena. The AFFIRM trial [17] with a primary end point of overall mortality showed no statistical difference between two groups (23.8% in rhythm control group vs 21.3% in the rate control group; P = 0.08) but with decreased hospitalization in the rate control group (80.1% vs 73%, P < 0.001). The RACE trial [18] with the combined primary composite end point of death from cardiovascular causes, thromboembolic complications, bleeding, need for pacemaker showed a 5.4% absolute difference in favor of rate control (22.6% rhythm control vs 17.2% rate control; absolute difference = −5.4%; 90% CI,−11.0 to 0.4). These trials suggest that, on average, simple rate control is the preferred initial approach to treatment [19, 20].

In slight contrast, a post hoc analysis of the ATHENA trial in which 4628 patients were randomized to placebo or dronedarone showed that dronedarone, a benzofuran derivative structurally similar to amiodarone, reduced the risk of stroke from 1.8% per year to 1.2% per year (hazard ratio, 0.66; 95% CI, 0.46–0.96; P = 0.027] [21]. Thus, it may be concluded that rhythm control with dronedarone confers a very slight advantage; however, this finding will need further confirmation.

Patients with rate or rhythm controlled AF still require anticoagulation to reduce the risk of stroke and systemic embolization. The risk of stroke or systemic embolization is the same among subjects with paroxysmal or chronic AF or atrial flutter. Chronic anticoagulation with coumarins, or in select cases with heparins, is a highly effective treatment approach with a 60% to 70% relative reduction in the risk of stroke or systemic embolism. The risk reduction is dependent upon maintaining the patient’s anticoagulation in the therapeutic range, which is challenging, and requires meticulous care and coordinated systems of care. In cases in which these are unavailable or the patients are unreliable, the risks of major hemorrhage may be far greater than the therapeutic risk reduction targeted. Several cohort studies have suggested that risk may be high in selected populations and places [22, 23]. Thus, the need for alternatives remains great, particularly when the prevalence of AF is rising.

Mechanical approaches to risk reduction including surgical removal of the atrial appendage at the time of coronary artery bypass graft and, more recently, endovascular occlusion of the left atrial appendage has been proposed. The latter has been tested in a randomized trial and shown to be feasible and potentially useful [2426]. This treatment is not yet adequately proven nor widely performed.

Pathophysiology of Antithrombotic Drugs

In the setting of AF all components of Virchow’s triad may be contribute to thrombogenesis. Abnormal blood flow/stasis may be due to atrial dilatation, poor ventricular contractility, valve disease, congestive heart failure, and cardioversion [27, 28]. These structural abnormalities may be associated with abnormal blood constituents such as elevated levels of β-thromboglobulin, D-dimers, thrombin-antithrombin (AT) complexes [2934], von Willebrand factor [31], or markers of inflammation such as interleukin-6, C-reactive protein, and plasma vascular endothelial growth factor, all of which predict the risk of stroke in AF [3538].

The available anticoagulant drugs inhibit various components of the clotting cascade (Fig. 1). Coumarins cause γ carboxylation and inhibition of vitamin K–dependent clotting factors (factors II, VII, IX, and X) and vitamin K–dependent anticoagulant factors (protein C and S) [3941]. Thus, the biological effect of coumarins is directly related to the normal half-life of these factors and, generally, the full anticoagulant effect is not seen for 4 to 7 days after commencing daily therapy [42]. In the early period after starting coumarins, there is a theoretical risk of inducing a procoagulant state because of differential reduction in these factors; among patients with known protein C deficiency, it is advisable to overlap treatment with parenteral heparin [39, 43].
Fig. 1

Shows extrinsic and intrinsic pathways of coagulation cascade with sites of action for numerous anticoagulants

Unfractionated heparin binds to AT, which inactivates thrombin and factor Xa. For inhibition of thrombin, heparin must bind simultaneously to AT, whereas this is not necessary for Xa inhibition. Low molecular weight heparins (LMWHs) contain a pentasaccharide molecule that binds factor Xa, but they lack enough length to also inhibit thrombin. LMWHs are in wide use for the treatment and prevention of venous thromboembolism and have been widely considered equivalent to unfractionated heparin for most clinical use. However, strictly speaking this is not true. Only one trial has examined their use in AF and stroke. The HAEST trial (discussed further below) showed that there was no difference in the rate of early recurrent stroke in the first 14 days after admission to hospital with a cardioembolic stroke or TIA due to AF [44].

The potential advantage of newer factor Xa inhibitors over heparin is that they inhibit both bound and free factor Xa and can be used in heparin resistance or heparin-induced thrombocytopenia. Parenteral agents are indirect Xa inhibitors (acting through AT), whereas oral agents are direct Xa inhibitors.

Existing Drugs: Coumarins, Heparin, and Antiplatelets

VKAs are the standard-of-care oral anticoagulants for long-term stroke prevention in AF. Efficacy in nonvalvular AF is well established from many trials (AFASAK, BAATAF, SPINAF, SPAF-I, II, III) [4550]. Reduction in the risk of stroke or death is estimated at 64% compared with placebo [5153]. A slightly smaller but still impressive risk reduction is seen when it is compared with aspirin [45, 48, 49].

Warfarin (adjusted INR 2–3) was recently compared with double antiplatelet therapy (acetylsalicylic acid [ASA]+clopidogrel 75 mg daily). In the ACTIVE-W trial, which was designed as a noninferiority study, the annual occurrence of stroke, systemic embolization, myocardial infarction, or vascular death was 3.9% in the warfarin group and 5.6% in the double antiplatelet group. The principal advantage of warfarin was a reduction of stroke (relative risk [RR], 1.72; 95% CI, 1.24, 2.37) and systemic embolization (RR, 4.66; 95% CI, 1.58, 13.8). The rates of major bleeding were similar in both groups: 2.42% per year with dual antiplatelets versus 2.21% per year with warfarin (RR, 1.1; 95% CI, 0.83, 1.45) [54•].

VKAs do not cause hemorrhage. If bleeding occurs, it is more serious and it is clear that small hemorrhages occur as part of normal life in selected patients (eg, gastrointestinal microhemorrhage and brain microhemorrhage), which may become large hemorrhages when the coagulation system is pharmacologically impaired. The risk of major bleeding is higher with coumarins compared with aspirin (2.2 vs 1.3 events per 100 patient-years) and risk increases with the INR value [55, 56]. Antiplatelet use, advanced age, supratherapeutic INR, and prior stroke increase the risk of intracranial hemorrhage (ICH) [5760]. Advancing age has been presumed to be a risk factor for bleeding during anticoagulation perhaps because of the accumulation of other risk factors such as stroke and more widespread use of antiplatelet agents. The BAFTA trial, with a mean subject age of 81 years, did not show excess bleeding in warfarin arm [61•]. A threefold increase in ICH has also been observed in patients taking combination of warfarin and ASA [62].

Coumarins have significant drug interactions [63] and their therapeutic effect will naturally vary according to the availability of vitamin K. Vitamin K arises from the diet and from normal intestinal flora. Thus, antibiotics that may affect normal flora or a change in dietary green vegetable intake may result in swings in the INR. Polymorphisms of P450 (CYP2C9) and vitamin K epoxide reductase complex 1 (VKORC1) enzyme recycling affect response to warfarin [39]. Rapid reversal of INR with vitamin K, fresh frozen plasma (FFP), and/or prothrombin complex concentrates (PCCs) (eg, Octaplex; Octapharma Canada, Brampton, Ontario [64]) in emergent situations is not fully effective [65]. Factor VIIa reverses the INR test in minutes but it has not yet been shown to completely reverse all aspects of VKA-associated hemorrhage [66]. Use of PCCs also normalizes the INR more rapidly than FFP but its effect on clinical outcome is not proved [6769]. It has been demonstrated in several studies that warfarin is underutilized due to these limitations [70, 71]. Underutilization is particularly prevalent in patients over the age of 80 years as they are perceived to be at high risk of bleeding due to frailty and comorbid illness. Paradoxically, these patients often have the most to gain from anticoagulation because they are at the highest risk of stroke [15].

Heparins have not been studied for long-term stroke prevention in AF. They are widely used as a transition or bridging medication before the level of VKAs becomes therapeutic, although the evidence for this is limited. This issue is discussed below.

Antiplatelet Therapy

ASA has been shown to have modest effect in preventing stroke in AF. The contraindications to chronic anticoagulation sometimes mandate use of ASA only. In the ACTIVE-A trial, 7554 patients unsuitable for warfarin therapy were randomized to clopidogrel or placebo over baseline aspirin (75–100 mg) treatment, and followed for a mean of 3.6 years. Dual antiplatelet therapy reduced the risk of stroke (RR, 0.72; 95% CI, 0.62–0.83) at a cost of a slight increase in the rate of major bleeding from 1.3% to 2% per year (RR, 1.57; 95% CI, 1.29–1.92) [72]. The combination of low-dose warfarin (adjusted INR target 1.5) and ASA versus warfarin (target INR 2–3) alone has been assessed in the SPAF-3 trial and found to be clearly inferior [50].

Novel Agents (Table 1)

Indirect Factor Xa Inhibitors

Idraparinux, a synthetic pentasaccharide, was evaluated for stroke prevention in 4576 patients with AF and at least one other risk factor for stroke in the AMADEUS trial using 2.5 mg injected subcutaneously once a week compared with adjusted dose warfarin. It was found to be noninferior to warfarin but caused more major bleeding (19.7% vs 11.3% per year; P < 0.0001) and ICHs (1.1% vs 0.4% per year; RR, 2.58; 95% CI, 1.18–5.63; P = 0.014). This along with other limitations (ie, renal clearance and lack of an antidote) led to early termination of the trial [73]. In view of these drawbacks, a biotinylated derivative of Idraparinux (ie, Idrabiotaparinux; antidote being avidin) has been developed. It is being evaluated in a phase 3 study, BOREALIS-AF, in which a 3-mg once-weekly subcutaneous dose is being compared with warfarin in about 9600 patients; results are expected in 2011.
Table 1

Pharmacologic characteristics and trials of some new anticoagulants compared with warfarin








Route of administration








Vitamin K epoxide reductase

Factor II

Factor Xa

Factor Xa

Factor Xa

Factor Xa

Onset of action

3–5 days

2 hours

3–6 hours

2–4 hours

3 hours

1–2 hours


Variable, OD

Twice daily

Once weekly

Once to twice daily

Twice daily

Once to twice daily


40 hours

14–17 hours

130 hours

7–11 hours

8–14 hours

9–11 hours

Drug interactionsa,b,c


Inhibitors of P-gp transporters


Inhibitors of CYP3A4 and P-gp transporters

Inhibitors of CYP3A4

Inhibitors of CYP3A4 and P-gp transporters

Renal clearance







Routine coagulation monitoring








Vitamin K






Phase 3 trials







Results in





2012 and 2010


CYP3A4 cytochrome P450 3A4, P-gp P-glycoprotein

aInhibitors of P-gp transporters (eg, amiodarone, quinidine) increase dabigatran levels

bInhibitors of CYP3A4 include antifungals, macrolide, and protease inhibitors

cInhibitors of CYP3A4 and P-gp transporters include antifungals and protease inhibitors

Direct Factor Xa Inhibitors

Rivaroxaban has a rapid onset of action, half-life of 5 to 9 hours, good oral bioavailability, and dual renal and hepatic excretion [74]. It is being evaluated in phase 3 trial of stroke prevention in AF (ROCKET-AF) in which 20 mg once a day of rivaroxaban (15 mg if glomerular filtration rate [GFR] between 30–49 mL/min) is being compared with adjusted-dose warfarin in 14,000 patients with nonvalvular AF and CHADS score of at least 2; results are expected in November 2010. A smaller similar trial has been conducted in Japan. Apixaban is a related oral thrombin inhibitor being evaluated in the ARISTOTLE study. A dose of 5 mg twice a day (bid) is being compared with warfarin for prevention of stroke or systemic embolism in 18,000 patients with nonvalvular AF and at least one more risk factor for stroke. The similar AVERROES study is comparing apixaban to aspirin for prevention of stroke or systemic embolism in 5600 patients with AF who have failed or are unsuitable for warfarin therapy. The results of both trials are expected in the latter half of 2010. Edoxaban is being evaluated by the TIMI group in the ENGAGE AF-TIMI 48 study. Edoxaban, 30 or 60 mg once daily orally, is being compared with adjusted-dose warfarin (INR 2–3) in 16,500 patients with AF. Results are expected in 2012. Betrixaban is a novel agent in the same class with a longer half-life of 15 hours and extrarenal clearance that may confer pharmacokinetic advantage over other xabans [75].

Thrombin Inhibitors

Thrombin (factor II) is a serine protease whose active site is surrounded by two exosites for attachment of blood proteins [7678]. Exosite I interacts with fibrinogen, factor V, and protein C and exosite II interacts with heparin. Direct thrombin inhibitors bind to inhibit the active site and exosite I, whereas indirect inhibitors bind to exosite II and AT [79]. By binding to thrombin directly they can inactivate clot-bound thrombin and are not inhibited by natural inhibitors such as platelet factor 4 or heparinase [8084]. They are also useful in uncommon conditions such as heparin-induced thrombocytopenia or acquired AT III deficiency.

Ximelagatran, dabigatran etexilate, and AZD0837 [85] are oral peptidomimetic direct and reversible thrombin inhibitors that have been evaluated for prevention of thromboembolism in AF. Ximelagatran was studied extensively in the SPORTIF trials and showed a 16% RR reduction (RRR) compared with adjusted-dose warfarin (INR 2–3) in the SPORTIF III and V trials [86, 87]. However, due to observations of transient elevation in liver enzymes in about 8% of patients, the drug has never been licensed [88].

Dabigatran etexilate is an orally administered prodrug of dabigatran with a peak blood level in 2 h, half-life of 14 to 17 h, with 80% renal excretion and a wide therapeutic window that allows fixed dosing above GFR greater than 30 mL/min. Although it causes prolongation of PT and PTT, neither assay can be used to monitor anticoagulant effect. Dabigatran was found to be as effective as enoxaparin for deep venous thrombosis prevention among orthopedic patients [89, 90]. The RE-LY study was a noninferiority trial assessing its safety and efficacy of two doses, 150 mg bid and 110 mg bid, compared with adjusted-dose warfarin (INR 2–3) in preventing stroke in 18,113 patients with nonvalvular AF and at least one other risk factor for stroke. The primary outcome was stroke and systemic embolism and safety end points included bleeding and abnormal liver functions. Dabigatran etexilate, 150 mg bid, was superior to warfarin in preventing stroke and systemic embolism (RR, 0.66; 95% CI, 0.53, 0.82) with no significant difference in incidence of major bleeding (3.11% per year on dabigatran 150 mg vs 3.36% per year on warfarin; P = 0.31).The lower dose of 110 mg bid was noninferior in efficacy to warfarin and showed a 20% RRR compared with the warfarin group (P = 0.003) for major bleeding. The occurrence of hemorrhagic strokes was lower with either dose of dabigatran (RRR 74% for 150 mg bid, P < 0.001; and RRR 69% for 110 mg bid; P < 0.001). In a post hoc analysis of the RE-LY trial, it was clear that the reduction in hemorrhage at the 150-mg-bid dosing of dabigatran was in large part related to the reduced time in the therapeutic range in the warfarin group. Patients with high INRs had a higher risk of ICH [91]. These results suggest that a major advantage of dabigatran is that it provides a stable and consistent anticoagulant effect. The mortality and liver function abnormality rates were not significantly different from the warfarin group. However, compared with warfarin, the 150-mg-bid dose was associated with a slightly but significantly increased risk of myocardial infarction (RR, 1.38; 95% CI, 1.0–1.91) but not statistically significant for the lower dose (RR, 1.35; 95% CI, 0.98–1.87). Dabigatran was associated with dyspepsia in 11.8% of patients on 150 mg bid and 11.3% on 110 mg bid compared with 5.8% on warfarin (P < 0.001). It needed to be discontinued for this reason in 2.2%, 2.1%, and 0.6% of patients on the 150-mg-bid and 110-mg-bid doses of the dabigatran and warfarin groups, respectively [92].

Anticoagulants Under Development

Thrombin inhibitors that would block only the procoagulant effect and spare its anticoagulant properties are being explored. Thrombin mutants and thrombomodulin mimetics are under development. Recombinant activated protein C (drotrecogin alfa) is already approved in reducing mortality due to sepsis-related coagulopathy [93]. Recombinant soluble thrombomodulin was found efficacious in preventing venous thromboembolism following total hip replacement surgery in a phase 2 trial [94]. Regarding factor IXa inhibition, REG1 is being evaluated for its anticoagulant properties in coronary artery disease [95]. It remains to be seen whether any of these agents might be suitable for stroke prevention in AF.

Dilemmas in the Clinical Management of Stroke Prevention in Stroke Patients with AF (Fig. 2)

Acute Stroke

In general, the risk of early recurrent stroke in the first 14 days after a stroke due to cardioembolism from AF is low to moderate ranging from 2.3% to 2.8% [96] to as high as 7.5% to 8.5% [44]. It remains entirely unclear if treatment with anticoagulants—usually unfractionated heparin or LMWH—will reduce that risk. In the HAEST trial, there was no benefit in risk reduction in the tinzaparin group, nor was there an increased risk of significant ICH [44]. The International Stroke Trial that tested unadjusted subcutaneous unfractionated heparin versus aspirin in acute stroke did show a slight increased risk of hemorrhagic transformation with heparin (1.2% risk of ICH) and suggestion of very modest benefit in preventing recurrent stroke [96]. Meta-analysis of randomized controlled trials showing efficacy and safety of anticoagulation after acute cardioembolic stroke has shown the rate of early recurrent stroke at the rate of 0.1% to 1.3% per day and also that there is no benefit of starting the anticoagulation in acute ischemic stroke to prevent early recurrence and improve functional outcomes [97]. Nevertheless, practicing stroke neurologists have all seen recurrent devastating cardioembolic stroke and there is a strong desire to prevent such events, which may account for the continued and ill-advised use of unfractionated heparin in this circumstance.
Fig. 2

Guidelines for antithrombotic therapy in nonvalvular atrial fibrillation (AF). Risk factors are not mutually exclusive and are additive to produce a composite risk. Echocardiogram is not needed for routine risk assessment but adds to certainty. In moderate-risk patients owing to lack of sufficient data from randomized trials, treatment might be decided on an individual basis, and the physician must balance the risks and benefits of warfarin versus aspirin. ASA—acetylsalicylic acid; INR—international normalized ratio; MCA—middle cerebral artery; PCA—posterior cerebral artery; TIA—transient ischemic attack. (Adapted from Lip and Lim [105])

An important competing issue is the risk of hemorrhage into the stroke. Petechial hemorrhage is part of the natural history of stroke and up to 70% of embolic infarcts will show some visible hemorrhagic conversion on brain imaging within 21 days of stroke onset [23]. Nearly 100% of pathologic examinations will show some petechial hemorrhage into an ischemic stroke. Devastating parenchymal hemorrhage is what must be avoided. Common teaching is that among large infarcts (~30 cm3 or greater), it is prudent to wait 7 to 10 days before commencing oral VKAs. However, this has never been formally tested and the common wisdom may be wrong. Some case series data suggest that immediate anticoagulation of cardioembolic stroke is safe [22] but these data have not been replicated in the modern era.

Bridging with Heparins

Because warfarin sodium takes several days to reach a therapeutic INR, different strategies are used during that period to achieve optimum anticoagulation: warfarin alone or with aspirin (no bridging); intravenous heparin sodium combined with warfarin (heparin bridging); and full-dose enoxaparin sodium (an LMWH), combined with warfarin (enoxaparin bridging). Given that the risk of recurrent stroke is so low, and one randomized trial has shown no benefit, it seems likely that bridging with heparins in the acute setting is simply not appropriate. However, it is still a common practice. One important issue to emphasize is that the risk of large parenchymal hemorrhage into the infarct bed is directly related to the peak value of the activated partial thromboplastin time (aPTT). Bolus injections of unfractionated heparin are to be avoided because the peak aPTT commonly rises greater than 150 s. It is uncertain whether the bolus dosing of LMWH increases risk in this setting, but it seems likely. Importantly, the effect size of the peak aPTT is much greater than the effect size of the volume of infarct in predicting hemorrhagic risk. Heparin bridging and enoxaparin bridging increased the risk for serious bleeding in one study [98].

Some physicians are reluctant to start warfarin treatment without bridging with either heparin or enoxaparin, because of the theoretical risk of a warfarin-related hypercoagulable state due to differential early reduction of the vitamin K–dependent anticoagulant factors, protein C and S. In reality, this is not a practical issue because abnormal clotting in dermal vessels causing skin necrosis is very rare and typically seen only in patients with protein C deficiency [99]. Currently, bridging patients to prevent skin necrosis when initiating warfarin in a maintenance dosage is not recommended by hematological guidelines unless the patient is known to have protein C deficiency [100].

Another common clinical problem is what to do with patients who are chronically anticoagulated for AF, but require a minor or major surgical procedure. Practice is highly variable in this circumstance. Some physicians will bridge patients with LMWH until the day of the procedure, and restart thereafter. Others will simply stop VKAs a week before and restart thereafter. Stroke neurologists know that a common story at presentation for cardioembolic stroke is the patient who stopped their VKA and had a colonoscopy yesterday or is due for a small biopsy but never gets there because they are admitted with a stroke.

There are reasonable data to suggest that minor procedures such as tooth extractions, skin biopsies, pacemaker placement, and cataract removal do not require cessation of VKAs [101104]. This information is not widely known and many surgeons refuse to perform such procedures while patients are on VKAs. Education is required. Practically, invasive procedures will require the cessation of VKAs but it remains unknown if bridging with LMWH is appropriate. It seems likely that a risk stratification process could be developed for this situation. This is an area in which a randomized trial is definitively needed.


Novel anticoagulant agents are going to revolutionize the treatment of stroke prevention in AF. The stroke community must seize this opportunity to develop trials to answer important questions about stroke treatment in both the acute phase and for secondary prevention.


Conflicts of interest: S. Bal: none; P. Ojha: none; M.D. Hill: is the director of the stroke unit at Foothills Medical Centre; he sits as a volunteer director of the board of the Heart & Stroke Foundation of Alberta (2010–2013) and is a consultant for Vernalis Group Ltd; he has received grants from the following companies: Hoffmann-La Roche Canada Ltd, Bayer Canada Ltd, and Merck Canada; has received honoraria from the following companies: Hoffmann La-Roche Canada Ltd, Stem Cell Therapeutics, Portola Therapeutics, Sanofi-Aventis Canada, Bristol-Meyers Squibb, and Merck Canada; he has received payment for educational presentations or served on the speakers’ bureau for Boehringer-Ingelheim Canada; he has stock in Calgary Scientific Inc; he has received travel expenses from GE Canada Healthcare to attend a GE conference; he has received a salary award from Alberta Heritage Foundation for Medical Research; and he is funded by the Heart & Stroke Foundation of Alberta Professorship in Stroke Research and is a Health Scholar of the Alberta Innovates Health Solutions.

Copyright information

© Springer Science+Business Media, LLC 2010