Author + information
- Received July 19, 2004
- Revision received August 31, 2004
- Accepted September 6, 2004
- Published online January 4, 2005.
- Jonathan L. Halperin, MD, FACC* ()
- ↵*Reprint requests and correspondence:
Dr. Jonathan L. Halperin, The Zena and Michael A. Wiener Cardiovascular Institute, Mount Sinai Medical Center, 1 Gustave L. Levy Place, New York, New York 10029-6574
Atrial fibrillation (AF) causes 50,000 to 100,000 ischemic strokes annually in the U.S., most of which could be prevented by oral anticoagulant treatment of the highest-risk patients. The greatest barrier to such treatment is the narrow therapeutic index of the vitamin K antagonists ([VKAs]: warfarin and related coumarin derivatives), the only oral anticoagulant agents currently available. Safe and effective treatment with the VKAs requires careful monitoring, because they interact with many other drugs and foods, and their anticoagulant action is unpredictable. Besides vitamin K, candidate targets for anticoagulant therapy include thrombin, a key prothrombotic mediator. Ximelagatran, the oral direct thrombin inhibitor at the most advanced stage of clinical development, is rapidly absorbed and bioconverted to its active moiety, melagatran—a potent, competitive inhibitor of both free and clot-bound thrombin. Two large clinical trials have demonstrated that fixed-dose oral ximelagatran, 36 mg twice daily, administered without coagulation monitoring, prevents stroke and systemic embolic events in patients with nonvalvular AF as effectively as well-controlled, adjusted-dose warfarin (international normalized ratio 2.0 to 3.0). The overall risk of bleeding was lower with ximelagatran than warfarin, although differences in rates of major hemorrhage were not statistically significant. Elevation of serum alanine aminotransferase levels above 3× the upper limit of normal occurred in approximately 6% of ximelagatran-treated patients but typically returned toward pretreatment levels without associated symptoms. In terms of preventing thromboembolism without hemorrhage, ximelagatran may have a more favorable benefit:risk profile than warfarin for patients with AF.
Nonvalvular atrial fibrillation (AF) is associated annually with 50,000 to 100,000 ischemic strokes in the U.S. and more than a million worldwide. Among anticoagulant-naive patients with no history of cerebrovascular events, AF increases the annual risk of ischemic stroke by five-fold, to ∼5% (1). Because emboli from intracardiac thrombi of diameter >5 mm are large enough to occlude a major cerebral artery (2), AF-related cardioembolic strokes are associated with poor functional outcomes (3), a high risk of permanent, severe disability (4), and high early mortality (5).
Atrial fibrillation is among the most common dysrhythmias seen in clinical practice. The Framingham Heart study investigators reported the lifetime risk of developing AF as approximately one in four (6). During 18 months' follow-up of 1.89 million adults in California, Go et al. (7) identified 17,974 patients with diagnosed AF, a period prevalence of 0.95%. Using 1995 U.S. census data, the authors estimated the national population of patients with AF as ∼2.3 million. The prevalence of AF is age-related, increasing from <1% among those younger than 50 years to ∼10% among those older than 80 years (1,7). As the mean age of the population increases, the number of patients with AF is projected to double over the next two generations (8,9), reaching ∼5 million in the U.S. by 2050 (7).
Current, evidenced-based practice guidelines recommend antithrombotic therapy for patients with AF and additional thromboembolic risk factors (1,10). Warfarin and other vitamin K antagonists (VKAs) are the antithrombotic agents of choice for high-risk patients, as they prevent disabling cardioembolic stroke far more effectively than aspirin. Intention-to-treat (ITT) pooled analysis of five randomized primary prevention trials showed that warfarin reduces stroke risk by 68% (11). In a meta-analysis of six trials, including one trial of secondary prevention, the relative risk reduction (RRR) was 62% (12) by ITT analysis, but ∼80% by on-treatment analysis, thus approaching complete reversal of the excess attributable risk (13). Nevertheless, in 1999 to 2000, VKAs were prescribed for fewer than 50% of a sample of ambulatory high-risk patients with AF (14).
Barriers to VKA use include high pharmacokinetic variability, a marked propensity for drug-drug and drug-food pharmacokinetic interactions, and a narrow therapeutic index (15). In comparisons of warfarin with other anticoagulants, the gradient of the dose-response curve (the Hill coefficient [HC]) correlated with the number of coagulation factors inhibited (Fig. 1)(16):
HCWARFARIN (↓ FII, FVII, FIX, FX)= 3.6
>HCheparin(↓ FII, FX)= 1.8
>HCmelagatran(↓ FII)= 1.2
∼HCinogatran(↓ FII)= 1.1
where the downward arrow (↓) denotes decreased coagulation-factor activity, through suppression of synthesis (warfarin) or inhibition, whether indirect (heparin) or direct (melagatran, inogatran). Among these agents, warfarin inhibits the greatest number of coagulation factors (four) and exhibits the steepest dose-response curve. By contrast, the agents with the shallowest dose-response curve, the direct thrombin inhibitors (DTIs) melagatran and inogatran (an early DTI), inhibit a single coagulation factor (16). These findings suggest that the narrow therapeutic index of warfarin is related to nonselective inhibition of all four vitamin K-dependent coagulation factors. They also suggest that rationally targeted inhibition of thrombin, a key mediator of thrombosis, might be associated with a wider therapeutic index.
The DTI at the most advanced stage of clinical development is ximelagatran (Exanta, AstraZeneca, Charnwood, United Kingdom and Mölndal, Sweden) (Fig. 2),a lipophilic molecule that readily penetrates cultured human intestinal epitheliocytes (17). In vitro, melagatran inhibits free and clot-bound thrombin rapidly, potently (Ki, 2 nmol/l), and reversibly, by direct, competitive binding to the catalytic site (18) (Fig. 3).In healthy volunteers, at least 40% to 70% of an oral ximelagatran dose is absorbed (19) and rapidly bioconverted to the active moiety, melagatran (17). Of more than 30,000 participants in the ximelagatran clinical trial program, at least 17,000 have received ximelagatran or melagatran, and treatment has continued for at least six months in more than 6,000 patients.
Prevention of venous thromboembolism (VTE) after elective hip- or knee-replacement surgery
In several European countries, ximelagatran is approved for prevention of VTE in patients undergoing elective hip- or knee-replacement surgery. A meta-analysis (20) of three European trials in this setting reported that subcutaneous melagatran followed by oral ximelagatran was effective and generally well-tolerated. In the U.S., four trials in patients undergoing elective hip- or knee-replacement surgery demonstrated that the efficacy of ximelagatran, 24 mg twice daily, initiated postoperatively, was similar to that of standard therapy (21–24). Two trials reported that ximelagatran, 36 mg twice daily, was more efficacious than warfarin (target international normalized ratio [INR], 2.5) in patients undergoing elective knee-replacement surgery (24,25).
Treatment and secondary prevention of acute VTE.
Ximelagatran 24, 36, 48, or 60 mg twice daily was compared with dalteparin followed by warfarin in patients with acute lower-extremity deep venous thrombosis (26). The venographically determined course of thrombosis and the rate of discontinuation due to bleeding were comparable in all treatment groups; bleeding risk did not vary with ximelagatran dose. A double-blind, phase III trial of secondary VTE prevention reported similar efficacy with ximelagatran, 36 mg twice daily, and standard enoxaparin/warfarin therapy (27).
Prevention of thromboembolic events after myocardial infarction
The multinational, randomized Efficacy and Safety of the Oral DTI XimElagatran in Patients with REcent Myocardial Damage (ESTEEM) trial (n = 1,883) enrolled patients within 14 days after myocardial infarction with or without ST-segment elevation (28). Patients were randomized either to ximelagatran in doses of 24, 36, 48, or 60 mg twice daily, or to placebo. All received aspirin, 160 mg daily (1,10). Over six months, ximelagatran plus aspirin was superior to aspirin alone in preventing the composite primary end point (death, nonfatal myocardial infarction, and severe recurrent ischemia), with an absolute risk reduction of 3.6% (hazard ratio, 0.76; 95% confidence interval, 0.59 to 0.98; p = 0.036). Rates of major hemorrhage were 1.8% and 0.9% (hazard ratio, 1.97; 95% confidence interval, 0.80 to 4.84) in the ximelagatran (all doses combined) and the placebo (aspirin only) groups, respectively. Ximelagatran was not associated with serious adverse events. The dose associated with optimum balance between efficacy and safety was 24 mg twice daily. Overall, this dose-guiding study demonstrated that, among survivors of acute myocardial infarction, ximelagatran plus aspirin offered better secondary prevention than aspirin alone (28). A larger, more definitive trial is planned.
The Stroke Prevention with an Oral Thrombin Inhibitor in AF (SPORTIF) clinical trial program
Phase I: Clinical pharmacology: SPORTIF-VI
The open-label, 6-day SPORTIF-VI trial compared the pharmacodynamic and pharmacokinetic profiles of ximelagatran in 12 patients with AF and 12 age- and gender-matched healthy volunteers (29,30). All participants were administered a single 2.66-mg dose of melagatran by intravenous infusion on day 1 and oral ximelagatran, 36 mg twice daily, on days 2 to 6.
Between day 1 (baseline) and day 6, coagulation and thrombin generation were suppressed to a similar degree, in both study groups (Table 1).Over the same period, the mean proportion of platelets expressing P-selectin decreased from 7.5% to 7.0% among healthy volunteers and from 10.9% to 9.2% among patients (29). These results demonstrate that ximelagatran exerts both an anticoagulant action, associated with suppression of thrombin generation, and a distinct antiplatelet action.
In the SPORTIF-VI trial, the Cmaxfor ximelagatran was 150 to 250 nmol/l after the first dose, and the time to Cmax(tmax) was ≤1 h both in patients with AF and in healthy volunteers (Table 2)(30). Melagatran (tmax, 3.0 h in patients vs. 2.9 h in healthy volunteers) was the overwhelmingly predominant species in plasma samples drawn 2 h or more after ximelagatran dosing. The mean terminal elimination half-life for melagatran (t1/2, 3.5 to 4.0 h) did not differ significantly between study groups, and the bioavailability of melagatran was consistent with that reported previously (31). The estimated renal clearance of melagatran approximated to the calculated creatinine clearance, suggesting that melagatran is eliminated primarily by glomerular filtration. In healthy volunteers, melagatran Cmaxwas ∼20% higher at plasma steady-state on day 6 than after initial ximelagatran dosing on day 2; in patients, the corresponding increase was ∼30% (30). These findings suggest that the pharmacokinetic characteristics of ximelagatran do not undergo clinically significantly changes during repeated dosing.
A population mathematical model of ximelagatran pharmacokinetic in patients with AF was derived from pooled results of three SPORTIF clinical trials (II, IV, and VI), in which follow-up continued for up to two years. The model predicts single-compartment disposition, with a lag-time and a first-order absorption-rate constant, independent of both treatment timing and dose (32). By covariate analysis, the model estimates the pharmacokinetic variability of melagatran as ∼23% within and ∼15% between individuals. The interindividual variability in melagatran bioavailability is 18%, while total variability in the area under the concentration-time curve is 45%. The mean t1/2for melagatran is 5 h, which permits ximelagatran dosing twice daily. The volume of distribution correlates linearly with body mass (156 l at 85 kg), and melagatran clearance correlates with renal function. In the pooled studies, the bioavailability of melagatran was not significantly altered by concomitant administration of angiotensin-converting enzyme inhibitors, beta-blockers, loop diuretics, verapamil, or dihydropyridine derivatives.
Similarly, studies in healthy volunteers (33) found that the pharmacokinetic profile of ximelagatran was not altered by concomitant administration of diclofenac, nifedipine, diazepam (34), aspirin (35), digoxin (36), or atorvastatin (37). A pharmacokinetic interaction between ximelagatran and erythromycin was associated with only a minor change in activated partial-thromboplastin time (38). Although ximelagatran metabolism appears independent of the P450enzyme system (35), the mechanism of interaction with erythromycin is under investigation. The pharmacokinetic profile of ximelagatran is not significantly influenced by age, gender (39), ethnicity (40), obesity (41), or intake of food (42) or alcohol (43).
Phase II Tolerability and general safety: SPORTIF-II and -IV
The tolerability and safety of long-term ximelagatran therapy in patients with AF were evaluated in the SPORTIF-II trial, a 12-week dose-guiding trial (44) at 37 centers in 11 countries. Patients with AF and at least one additional thromboembolic risk factor were randomized to either fixed-dose ximelagatran, 20, 40, or 60 mg twice daily without coagulation monitoring, or adjusted-dose warfarin (target INR, 2.0 to 3.0). As the inclusion criteria for the SPORTIF-II trial and for the subsequent phase III SPORTIF-III and -V trials were based on current indications for long-term anticoagulation in patients with AF, these SPORTIF study populations broadly resembled those in earlier placebo-controlled trials of warfarin. The primary end point was the number of thromboembolic and major bleeding events.
Patients completing the SPORTIF-II trial were eligible for its open-label extension, SPORTIF-IV, with follow-up continuing up to five years. Patients in the SPORTIF-IV trial received the originally randomized antithrombotic agent, but all patients assigned to ximelagatran coalesced to a dose of 36 mg twice daily. Interim analysis was conducted two years after randomization (45). In the warfarin group (n = 67), the proportion of patients with INR values in the target range increased from 34% at entry (reflecting previous warfarin treatment) to 57% at SPORTIF-II study completion (12 weeks after randomization). Mean adherence to ximelagatran treatment (n = 187), estimated by tablet counts, was 100%, 96%, and 98% in the 20-, 40-, and 60-mg dose groups, respectively.
Four nonfatal primary end point events occurred in the SPORTIF-II trial: one ischemic stroke and one transient ischemic attack (TIA) in the ximelagatran group and two TIAs in the warfarin group (44). One major hemorrhage occurred in a warfarin-treated patient, but none in patients randomized to ximelagatran. In both treatment groups, ∼50% of patients reported adverse events during the study; the incidence of most adverse events did not differ significantly between groups. Serum alanine aminotransferase elevations above 3× the upper limit of normal (ULN) (>3×), however, were reported only among ximelagatran-treated patients (8 of 187 = 4.3%). In all eight of these patients, serum alanine aminotransferase levels returned to normal whether treatment was continued (five patients) or interrupted (three patients) (44).
In the SPORTIF-IV interim analysis, two years after randomization, two nonfatal ischemic strokes had occurred during 231 ximelagatran treatment-years (0.9% per year), compared with two fatal hemorrhagic strokes during 76 warfarin treatment-years (2.6% per year). For TIAs, the rates were 0.4% per year and 2.6% per year, respectively (45). Rates of major bleeding were 0.9% per year (ximelagatran) versus 2.6% per year (warfarin). Five patients died, including the two warfarin-treated patients with hemorrhagic strokes; of three deaths in the ximelagatran group, none was attributed to study treatment. Twelve patients had alanine aminotransferase elevations >3× ULN on ximelagatran, but alanine aminotransferase level returned to normal in all 12 (8 who continued and 4 who discontinued treatment) (45).
Phase III: Efficacy—SPORTIF-III and -V
Two randomized phase III trials (combined, n > 7,000) evaluated the efficacy of ximelagatran in patients with AF at elevated thromboembolic risk: SPORTIF-III, at 259 centers in Europe, Asia, and Australasia, and SPORTIF-V, at 409 centers in North America (46). The protocols differed only as regards treatment blinding: open-label administration of ximelagatran or warfarin in the SPORTIF-III trial, and double-blind anticoagulation in the SPORTIF-V trial. Pooled analysis of the results from both trials was prespecified to assess heterogeneity, compare the effects of the anticoagulants on events occurring at low frequencies, and examine patient subgroups.
Given the proven efficacy of the VKAs in patients with AF, a placebo would have been unethical, and an active-control design was required; the SPORTIF-III and -V trials, therefore, tested the noninferiority, within a margin of 2% per year, of ximelagatran relative to adjusted-dose warfarin (46), based upon the hypothesis that fixed-dose ximelagatran, 36 mg twice daily, without coagulation monitoring, prevents all stroke (ischemic or hemorrhagic) and systemic embolic events (SEEs) at least as effectively as well-adjusted warfarin (target INR, 2.0 to 3.0). In the combined study population, the mean age was 71.6 years, 69% were male, and 75% had at least two thromboembolic risk factors beyond AF. Treatment allocation was balanced according to aspirin therapy at entry and history of stroke or TIA. In both trials, local evaluation of primary end point events was masked, and all such events were independently assessed by a Central Event Adjudication Committee to reduce bias, particularly in the open-label trial (47).
At study entry, typical patients in the SPORTIF-III trial (n = 3,407) were elderly, hypertensive, high-risk white men tolerating warfarin anticoagulation (46,47). After randomization, the demographic and risk-factor profiles of the treatment groups were similar. There were 1,158 men (68%) in the ximelagatran group (n = 1,704) and 1,196 (70%) in the warfarin group (n = 1,703). Mean age was 70.1 years (standard deviation, 8.6) in the ximelagatran group, compared with 70.3 years (standard deviation, 8.6) in the warfarin group. In both groups, 88% of patients were white, and 34% were ≥75 years old. Thromboembolic risk factors, such as left ventricular dysfunction, diabetes mellitus, and hypertension, were distributed evenly; mean systolic blood pressure was 139 mm Hg in both groups. Multiple risk factors were present in 70% and 68% of patients in the ximelagatran and warfarin groups, respectively. Use of aspirin at entry was 20% in the ximelagatran group and 21% in the warfarin group. Continued use of aspirin, in doses of up to 100 mg per day in addition to randomized anticoagulant therapy was permitted for patients with concomitant coronary disease, in keeping with guidelines (1,10).
The mean follow-up in the SPORTIF-III trial was 17.4 months; INR values in patients assigned to warfarin (mean, 2.5 ± 0.7) were within the target range (2.0 to 3.0) for 66% of the entire duration of exposure and within the extended range of 1.8 to 3.2 for 81%. Adherence to ximelagatran therapy (tablet counts) was 94% (47).
By ITT analysis, primary end point events occurred in 40 patients randomized to ximelagatran, during 2,446 patient-years at risk (1.64% per patient-year) (47). Thirty-two had ischemic stroke (1.3% per patient-year), four had hemorrhagic stroke (0.2% per patient-year), and four had SEE (0.2% per patient-year). Primary events occurred in 56 patients randomized to warfarin, during 2,440 patient-years at risk (2.29% per patient-year). Forty-six had an ischemic stroke (1.9% per patient-year), nine had hemorrhagic stroke (0.4% per patient-year), and two had SEE (0.1% per patient-year). Compared with warfarin, ximelagatran was associated with an absolute risk reduction of 0.65% (95% confidence interval, −0.1 to 1.4; p = 0.1) (Table 3,Fig. 4)(47,48), which met the predefined criterion for noninferiority (≤2% per year) relative to warfarin. The RRR was 29% (95% confidence interval −6.5 to 52). Kaplan-Meier curves depicting the cumulative percentages of patients experiencing primary events during 24 months follow-up are presented in Figure 5.
By secondary on-treatment analysis, 29 primary end point events occurred during 2,286 patient-years (1.27% per year) in the ximelagatran group, compared with 52 events during 2,352 patient-years (2.21% per year) in the warfarin group. Ximelagatran was associated with an absolute risk reduction of 0.9% per patient-year (0.2 to 1.7; p = 0.0180) and an RRR of 43% versus warfarin (10 to 63). Event-rate comparisons for both the ITT and on-treatment analyses of primary outcomes, together with prespecified sensitivity analyses for each comparison, are depicted in Figure 4.
End point events were associated with discontinuation of study drug in similar proportions of patients: ximelagatran group, 3%; warfarin group, 4%. Only 138 patients (4.1%) withdrew from the study for reasons other than death, and the rate of premature, permanent discontinuation was lower than in earlier studies of warfarin. Efficacy was consistent across all predefined subgroups, irrespective of risk factors.
Treatment groups did not differ significantly with respect to all-cause mortality (3.2% per year), fatal or nonfatal disabling stroke, or composite secondary end points (47). Rates of bleeding (major plus minor, the latter including purpura, epistaxis, occult fecal hemoglobin, and microscopic hematuria) (47) were 25.8% per year with ximelagatran (478 patients) versus 29.8% per year with warfarin (547 patients)—an RRR of 14% (95% confidence interval, 4% to 22%; p = 0.0065) favoring ximelagatran (47). In both groups, total bleeding rates among patients with creatinine clearance <80 ml/min (ximelagatran, 31% per year; warfarin, 29% per year) were similar to those with better renal function (ximelagatran, 20% per year; warfarin, 32% per year) (47). Hemorrhagic stroke, fatal hemorrhagic stroke, other major bleeding, and study drug discontinuation related to major bleeding occurred at similar rates in the two groups (47).
During the trial, concomitant aspirin use was more frequent in the ximelagatran group (337 of 1,704; 20%) than the warfarin group (290 of 1,703; 17%; p = 0.042) (47,49,50). Among patients taking aspirin, 14 primary events occurred during 477 patient-years (2.94% per year) in the ximelagatran group versus 16 events during 399 patient-years (4.01% per year) in the warfarin group, an absolute risk reduction of 1.08% per year (95% confidence interval, −3.57 to 1.42). Among patients not taking aspirin, 26 primary events occurred during 1,969 patient-years (1.32% per year) with ximelagatran versus 40 events during 2,042 patient-years (1.96% per year) with warfarin, an absolute risk reduction of 0.64% per year (95% confidence interval, −1.43 to 0.15). Statistical testing did not suggest any significant correlation between study drug and concomitant use of aspirin (p = 0.85). The higher rate of primary events among patients taking aspirin may reflect a greater intrinsic risk in this patient subgroup (50).
Adverse events were reported in 1,472 patients (87%) in the ximelagatran group, compared with 1,452 patients (85%; p = 0.228) in the warfarin group. As in the SPORTIF-II and -IV trials, between-group differences in the frequency of particular adverse events were not significant, except for serum alanine aminotransferase elevations. Elevations >3× ULN developed in 107 patients (6%) in the ximelagatran group versus 14 (1%; p < 0.0001) with warfarin. Typically, these began one to six months after initiation of treatment, were not associated with specific symptoms, and returned toward baseline without clinical sequelae whether therapy was continued (59 patients) or discontinued (48 patients).
The composite rate of death, primary events, and major bleeding during treatment was used as a measure of net clinical benefit in an exploratory, post-hoc analysis. The composite rate was 4.6% per year with ximelagatran (104 events) versus 6.1% per year with warfarin (143 events), RRR 25% (95% confidence interval, 4 to 42; p = 0.019) favoring ximelagatran (47).
In the double-blind SPORTIF-V trial (n = 3,922), primary end point events occurred in 88 patients during 6,405 patient-years of exposure (mean, 20 months per patient) (51). By ITT analysis, rates of primary events were 1.2% per year (warfarin) versus 1.6% per year (ximelagatran), absolute risk reduction 0.4% per year (95% confidence interval, 0.13% to 1.03% per year; p = 0.13). This result met the prespecified criterion for noninferiority of ximelagatran relative to warfarin, confirming the principal result of the SPORTIF-III trial.
By on-treatment analysis, the absolute risk reduction for primary end point events was 0.55% per year (95% confidence interval, −0.06 to 16; p = 0.089). For the composite of primary events and all-cause mortality, the difference was not significant (0.10% per year; 95% confidence interval, −0.97% to 1.18% per year; p = 0.86). Rates of disabling or fatal stroke did not differ between groups.
As in the SPORTIF-III trial, the quality of warfarin anticoagulation in the SPORTIF-V trial was better than in earlier trials (47,51,52), and far better than in usual practice. For warfarin-treated patients, the INR (mean, 2.4 ± 0.8) was within the target range (2.0 to 3.0) 68% of the time and within the extended range (1.8 to 3.2) 83% of the time. The combined rate of bleeding (major plus minor) was significantly lower with ximelagatran (37% per year) than with warfarin (47% per year; p < 0.0001). There were no significant between-group differences in the rates of hemorrhagic stroke or major bleeding. Within the first six months of treatment, alanine aminotransferase elevation >3× ULN developed in 6.0% versus 0.8% (p < 0.001) of patients with ximelagatran and warfarin, respectively. Typically, elevations were not associated with specific symptoms, and alanine aminotransferase returned toward the pretreatment baseline, whether treatment was continued or discontinued. One 80-year-old patient, however, developed fatal gastrointestinal bleeding during corticosteroid therapy for hepatitis associated with ximelagatran.
The prespecified pooled analysis of the SPORTIF-III and -V trials (51) encompassed a total exposure of 11,346 patient-years (mean, 18.5 months). Treatment groups did not differ with respect to rates of primary events or strokes of presumed cardioembolic or noncardioembolic origin. In the ITT populations, the primary event rates were 1.62% per year versus 1.65% per year in the ximelagatran and warfarin groups, respectively, absolute risk reduction with ximelagatran of 0.03% per year (95% confidence interval, −0.44% to 0.50% per year; p = 0.941). In the subpopulation with a history of stroke or TIA, primary event rates were 2.83% (ximelagatran) versus 3.27% (warfarin), an absolute risk reduction of 0.44% per year (95% confidence interval, −0.98 to 1.86; p = 0.625). Ischemic stroke (mostly noncardioembolic) occurred at rates of 0.75% per year (ximelagatran) versus 0.92% per year (warfarin). Rates of cardioembolic stroke were low: 0.62% per year (ximelagatran) versus 0.53% per year (warfarin). The annual incidence of bleeding (minor plus major) was lower with ximelagatran (32%) than warfarin (39%; p < 0.0001). The rate of intracranial hemorrhage was 0.11% per year with ximelagatran versus 0.20% per year with warfarin (51). Aspirin was used concomitantly by 22.7% and 17.2% of patients with and without a history of stroke, respectively.
Potentially fatal or disabling cardioembolic stroke continues to threaten a large proportion of patients with AF, chiefly because VKA oral anticoagulation is used in fewer than half of eligible patients despite numerous trials demonstrating the efficacy of warfarin. Recognized barriers to wider use of VKAs include their unpredictable anticoagulant action, pharmacokinetic variability, and propensity for drug-drug and drug-food pharmacokinetic interactions, as well as the ongoing need for regular therapeutic monitoring. Paramount among the limitations of the VKAs is a narrow therapeutic index that reflects nonselective inhibition of multiple coagulation factors. In this context, selective targeting of a single prothrombotic mediator has emerged as the key therapeutic principle guiding the rational design of next-generation antithrombotic agents.
The oral DTI ximelagatran has received regulatory approval in various European countries for the prevention of thromboembolic events in patients undergoing elective total hip- or total knee-replacement surgery, but the compound has not gained regulatory approval in the U.S. for any indication to date. In the SPORTIF-III and -V trials, two large phase III trials with a combined study population of more than 7,000 patients with nonvalvular AF, oral ximelagatran, 36 mg twice daily, administered at a fixed dose without coagulation monitoring, prevented stroke as effectively as well-controlled warfarin (INR 2.0 to 3.0), without increased bleeding (47,51,52). Serum transaminase levels increased in a small percentage of patients during the early months of ximelagatran treatment but further studies are needed to establish the frequency or occurence of more serious liver disease. Nevertheless, by eliminating both coagulation monitoring and dose adjustment, ximelagatran has the potential to encourage the use of oral anticoagulation in a larger proportion of the eligible, at-risk population of patients with AF (53–55). Were ximelagatran approved for oral anticoagulation in this setting, then the guidelines for management of patients with AF would need to be revised (56).
The author gratefully acknowledges the assistance of Dr. Garry Sandison, MB, in the preparation of this manuscript.
Dr. Halperin is the recipient of consulting fees from AstraZeneca, L.P., as Co-Chairman of the Executive Steering Committee for the SPORTIF-III and -V clinical trials, which address the issue of antithrombotic therapy for prevention of thromboembolism in patients with atrial fibrillation, and Chairman of the Executive Steering Committee for the EXPECT trial investigating this issue in the context of cardioversion. AstraZeneca is the manufacturer of ximelagatran.
- Abbreviations and acronyms
- atrial fibrillation
- direct thrombin inhibitor
- international normalized ratio
- relative risk reduction
- systemic embolic event
- Stroke Prevention with an ORal Thrombin Inhibitor in Atrial Fibrillation
- transient ischemic attack
- vitamin K antagonist
- venous thromboembolism
- Received July 19, 2004.
- Revision received August 31, 2004.
- Accepted September 6, 2004.
- American College of Cardiology Foundation
- Hart R.G.,
- Palacio S.,
- Pearce L.A.
- Lloyd-Jones D.M.,
- Wang T.J.,
- Leip E.P.,
- et al.
- Steinberg J.S.
- Stewart S.,
- Murphy N.,
- Walker A.,
- McGuire A.,
- McMurray J.J.V.
- Fuster V.,
- Rydén L.E.,
- Asinger R.W.,
- et al.
- ↵Hart RG, Benavente O, McBride R, Pearce LA. Antithrombotic therapy to prevent stroke in patients with atrial fibrillation: a meta-analysis. Ann Intern Med 1999;131:492–501.
- ↵Hart RG, Halperin JL. Atrial fibrillation and thromboembolism: a decade of progress in stroke prevention. Ann Intern Med 1999;131:688–95.
- Eriksson U.G.,
- Bredberg U.,
- Hoffmann K.-J.,
- et al.
- ↵Cohen AT, Agnelli G, Dahl OE, et al. Efficacy and safety of the treatment regimen of melagatran and ximelagatran for prevention of thromboembolic events after total hip or knee replacement: a meta-analysis of 3 randomized, double-blind studies, to evaluate the influence of time and dose (abstr CD-ROM). J Thromb Haemost 2003;1 Suppl 1.
- Colwell C.W.J.,
- Berkowitz S.D.,
- Davidson B.L.,
- et al.
- Colwell CW, Berkowitz SD, Comp PC, et al. Randomized, double-blind comparison of ximelagatran, an oral direct thrombin inhibitor, and warfarin to prevent venous thromboembolism (VTE) after total knee replacement (TKR): EXULT B (abstr). Blood 2003;102. Online only. Available at: http://www.abstracts2view.com/hem/.
- Eriksson H.,
- Wåhlander K.,
- Gustafsson D.,
- Welin L.,
- Frison L.,
- Schulman S.
- Francis C.W.,
- Ginsberg J.S.,
- Berkowitz S.D.,
- et al.
- Grind M.,
- Eriksson U.G.,
- Hamrén B.,
- Bååthe S.,
- Wollbratt M.
- Dorani H.,
- Schützer K.-M.,
- Sarich T.C.,
- Wall U.,
- Ohlsson L.,
- Eriksson U.G.
- Johansson L.C.,
- Andersson M.,
- Fager G.,
- Gustafsson D.,
- Eriksson U.G.
- Petersen P.,
- Grind M.,
- Adler J.,
- SPORTIF II Investigators
- ↵Petersen P. A 2-year follow-up of ximelagatran as an oral anticoagulant for the prevention of stroke and systemic embolism in patients with nonvalvular atrial fibrillation (abstr). Neurology 2002;58 Suppl 3:A477.
- Halperin J.L.,
- Executive Steering Committee on behalf of the SPORTIF III and V Study Investigators
- Executive Steering Committee on behalf of the SPORTIF III Investigators
- ↵Albers GW, for the Executive Steering Commitee on behalf of the SPORTIF Investigators. Pooled analysis of the SPORTIF III and V trials: secondary stroke prevention and stroke subtypes (abstr). Stroke 2004;35:242–3.
- Executive Steering Committee on behalf of the SPORTIF V Investigators
- Hankey G.,
- Klijn C.J.,
- Eikelboom J.
- Rockson S.G.,
- Albers G.W.