PRIMARY STROKE PREVENTION INNONVALVULAR ATRIAL FIBRILLATION: IMPLEMENTING THE CLINICAL TRIAL FINDINGS

Transcription

1 Therapeutic Controversies PRIMARY STROKE PREVENTION INNONVALVULAR ATRIAL FIBRILLATION: IMPLEMENTING THE CLINICAL TRIAL FINDINGS Patricia A Howard and Pamela W Duncan OBJECTIVE: To review the clinical trials evaluating warfarin for primary stroke prophylaxis in nonvalvular atrial fibrillation (NY AF), to discuss the relative benefits and risks of warfarin versus aspirin therapy, and to review the clinical practice guidelines and identify potential barriers to their implementation in clinical practice. DATA SOURCES: A MEDLINE literature search was performed to identify clinical trials of antithrombotic therapy for NYAF, clinical practice guidelines, studies evaluating physician practices and attitudes, cost-effecti veness studies, and pertinent review articles. Key search terms included atrial fibrillation, stroke, antithrombotic, warfarin, aspirin, and cost-effectiveness. DATA EXTRACTION: Prospective, randomized clinical trials were selected for analysis. Clinical practice guidelines from recognized panels of experts were reviewed. Comprehensive review articles were selected. DATA SYNTHESIS: NY AF is a common arrhythmia that is associated with a substantial risk for stroke. Seven prospective, randomized, clinical trials have conclusively demonstrated the efficacy of warfarin for stroke prevention. The greatest benefits are achieved in older patients and those with comorbidities that increase their risk for stroke. The potential benefits of preventing a devastating stroke, however, must be weighed against the potential for bleeding complications. Warfarin has been shown to be cost-effective in high-risk patients, provided the rate of complications is minimized. Nonetheless, many physicians remain hesitant to implement warfarin therapy in older, high-risk patients. The clinical data on aspirin are less consistent than those observed with warfarin. Aspirin appears to be most effective in younger individuals or those considered to be at low risk for stroke. CONCLUSIONS: In patients with NYAF, the personal, social, and economic consequences of stroke are often devastating. Clinical trials have provided definitive proof that the risks of stroke can be significantly reduced through the use of appropriate antithrombotic therapy. Despite this evidence and the recommendations of a Patricia A Howard PhannD FCCP BCPS. Associate Professor. Department of Pharmacy Practice, School of Pharmacy, University of Kansas Medical Center, Kansas City, KS Pamela W Duncan PT PhD, Associate Professor, Health Services Administration, School of Pharmacy, University of Kansas Medical Center Reprints: Patricia A Howard PharmD FCCP BCPS, Department of Pharmacy Practice, University of Kansas Medical Center Rainbow Blvd. Kansas City, KS FAX 913/ number of clinical practice guidelines, variations in care exist that continue to place patients at risk. Additional outcomes research is needed to evaluate the impact of the clinical trial findings and practice guidelines on clinical practice and to develop methods for overcoming barriers to implementation. KEY WORDS: arrhythmias, atrial fibrillation, warfarin, aspirin, prophylaxis. Ann Pharmacother 1997;31: ATRIAL FIBRILLATION (AF) affects approximately 2.2 million people in the US. The median age is 75 years and approximately 70% of individuals with AF are between 65 and 85 years of age. I In the past, valvular heart disease was the most common condition associated with AF; however, nonvalvular AF (NVAF) now accounts for approximately 80% of all cases. In North America, an estimated strokes occur each year in individuals with AF. It has been estimated that 15-30% of all ischemic strokes are associated with AF.2Overall, the risk of stroke with AF is approximately 5% per year; however, this represents a wide spectrum from less than 1% to more than 8%, depending on other risk factors.' For many patients, the consequences of a first stroke are devastating. The mortality rate with stroke is 17-34% in the first 30 days and 25-40% during the first year. Approximately 25-50% of survivors have permanent neurologic damage, which may result in partial or total dependence in activities of daily living.' The total annual cost for stroke in the US has been estimated at $30 billion.' Because of the enormous emotional, social, and economic impact of stroke, primary prevention is of the utmost importance. A number of randomized clinical trials have demonstrated the benefits of warfarin and aspirin therapy in patients with NVAF. Professional organizations and expert panels have developed clinical practice guidelines for stroke prophylaxis. In general, these guidelines recommend warfarin as first-line therapy for high-risk patients. The AnnalsofPharmacotherapy 1997 October, Volume 31 JJ87

2 Aspirin is recommended as alternative therapy for low-risk patients or those unable to take warfarin. Despite the overwhelming evidence in favor of stroke prophylaxis for patients with NVAF, there are significant variations in the use of antithrombotic agents in eligible patients. The objectives of this article are to critically evaluate and compare the relative benefits and risks of warfarin versus aspirin therapy in NVAF, review the clinical guidelines, and discuss the applicability of the clinical trial evidence and practice guidelines to clinical practice. literature Review Seven prospective, randomized, clinical trials have evaluated the efficacy of warfarin and/or aspirin for primary stroke prevention in NVAF, as summarized in Table The Boston Area Anticoagulation Trial for Atrial Fibrillation (BAATAF) compared the effects of low-dose warfarin therapy with control patients who received no treatment but were allowed to take aspirin on their own," The primary end point was ischemic stroke, and death was evaluated as a secondary end point. Warfarin reduced the annual incidence of stroke by 86% (95% CI 51% to %%; P=0.0022). The study population had relatively few comorbidities, reflected in the low incidence of stroke in the control group. The death rate was reduced with warfarin from 5.97% to 2.25%, for a risk reduction of 62% (95% CI 18% to 83%; P =0.005). Major bleeding defined as that leading to hospitalization or transfusion occurred in 2 patients receiving warfarin and in 1 control patient. One fatal hemorrhage occurred in each group. Minor bleeding occurred in 38 patients receiving warfarin versus 21 control patients, for an incidence ratio of 1.62 (95% CI 0.95% to 2.74%). Ten percent of warfarin patients discontinued therapy during the study. The control group was potentially compromised in that patients were allowed to take aspirin on their own and 46% did so regularly. The Copenhagen Atrial Fibrillation, Aspirin, Anticoagulation (AFASAK) study randomized patients to receive open-label warfarin, or double-blind aspirin or placebo.' The primary end points were thromboembolic complications defined as stroke, transient ischemic attack (TIA), or embolic complications of the viscera and extremities. The secondary end point was death. The study found no statistically significant differences between aspirin and placebo. The yearly rate of thromboembolism was 2.0% (95% CI 0.6% to 4.8%) with warfarin versus 5.5% for both aspirin and placebo (95% CI 2.9% to 9.4%). Of the patients randomized to receive warfarin, 38% withdrew due to adverse events or the inconvenience of drawing blood. The overall incidence of bleeding was 6.3% with warfarin and 0.6% with aspirin. No bleeding occurred with placebo. The first Stroke Prevention in Atrial Fibrillation (SPAF I) study compared either open-label warfarin or double-blind aspirin with placebo," The study did not have sufficient sample size to directly compare the two drugs. Group I pa- Table 1. Efficacy of Warfarin and Aspirin for Primary Stroke Prevention in Nonvalvular Atrial Fibrillation EMBOLIC STUDY PATIENTS PRIMARY EVENTS MEAN FOLLOW-UP EMBOLIC (Annual %) % TRIAL DESIGN n AGE(y) %MEN INTERVENTION MEAN(y) ENDPOINTS W C A REDUCTION BAATAF (1990) NB.RM W: PTR S " C: no therapy or A AFASAK (1989)' RM,PC b 54 W: INR S. SE, TIA w vs. C: 59" W:NB A: 75 mg/d A vs. C: 18 A:DB SPAF I (1991)8 RM.PC W: PTR S, SE w vs. C: 67" W:NB A vs. C: 44" A:DB CAFA (1991)' RM,DB,PC W: INR S, SE SPINAF (1992)10 RM.DB,PC W: PTR S " SPAFIl (1994)" RM.PL W: PTR or 3.1 S, SE w vs. A: 33 INR vs A: 325 mg/d Wvs. A: 27 SPAF 111 (1996)12 RM,NB W: INR vs. 1.1 S,SE 1.9 W vs. (W + A): 74" W: INR A 325 mg/d (W+A) Pooled analysis'y Wvs.C S w vs. C: 68" A vs.c A vs. C: 36" A = aspirin; AFASAK = Copenhagen Atrial Fibrillation, Aspirin, Anticoagulation study; BAATAF = Boston Area Anticoagulation Trial for Atrial Fibrillation; C = control group; CAFA = Canadian Atrial Fibrillation Anticoagulation study; DB = double-blind; INR = international normalized ratio; NB = nonblinded; PC = placebo-controlled; PL = parallel design; PTR = prothrombin time ratio; RM = randomized; S = ischemic stroke; SE = systemic embolism; SPAF = Stroke Prevention in Atrial Fibrillation study; SPINAF = Stroke Prevention in Nonrheumatic Atrial Fibrillation; TIA = transient ischemic attack; W = warfarin. "Statistically significant at p < bmedian age. 'Pooled analysis ofbaataf. AFASAK, CAFA, SPAF I, and SPINAF. l188 The Annals ofpharmacotherapy 1997 October, Volume 31

3 Stroke Prevention in Non valvular Atrial Fibrillation tients were randomized to receive aspirin or warfarin. Group II patients were considered ineligible for warfarin and received aspirin or placebo. Ineligibility criteria included patients unwilling to take warfarin or undergo regular laboratory monitoring, patients with a high risk of hemorrhage, patients older than 75 years, and patients at low risk for embolism such as those with lone AF (atrial fibrillation in the absence of underlying cardiovascular disease). The primary end points were ischemic stroke and systemic embolism. Compliance rates with both aspirin and warfarin were approximately 80-88%. Approximately II % of warfarin patients and 5% of aspirin patients were withdrawn from medication during the trial. Compared with placebo, the rates of primary events were reduced 67% (95% CI 27% to 85%; p =0.01) with warfarin and 42% (95% CI 9% to 63%; p =0.02) with aspirin. The annual risks of significant bleeding were 1.5%, 1.4%, and 1.6% with warfarin, aspirin, and placebo, respectively. A fatal intracranial hemorrhage (lch) occurred in both the warfarin and aspirin groups. The Canadian Atrial Fibrillation Anticoagulation (CAFA) study was designed to assess the efficacy of warfarin in 630 patients with NVAP However, the trial was stopped for ethical reasons after 378 patients were randomized due to publication of the positive findings from the AFASAK and SPAF I trials. The primary outcomes were nonlacunar ischemic stroke, systemic embolism, and intracranial or fatal hemorrhage. Compared with placebo, the annual rate of primary events was reduced 37% (95% CI -63.5% to 75.5%; P =0.17) with warfarin. This difference was not statistically significant, which was most likely due to the early termination of the trial. Fatal or major bleeding occurred at an annual rate of 2.5% with warfarin versus 0.5% with placebo. One fatal ICH occurred in a warfarin-treated patient. Minor bleeding occurred in 16% of warfarin patients and 9% of placebo patients. The Stroke Prevention in Nonrheumatic Atrial Fibrillation (SPINAF) trial, also known as the Veterans Affairs Cooperative study, evaluated the efficacy of warfarin versus placebo. to The primary end point was cerebral infarction and secondary end points were cerebral hemorrhage or death. The risk reduction for primary events with warfarin was 0.79 (95% CI 0.52 to 0.90; P = 0.001). Warfarin was withdrawn from 81 patients during the study compared with 56 patients receiving placebo. One ICH occurred in a 73-year-old warfarin-treated patient. Other major hemorrhages occurred at an annual rate of 1.3% for warfarin versus 0.9% for placebo. Minor hemorrhages occurred 115 times in warfarin-treated patients compared with 96 times in placebo patients (p = 0.04). There were 15 deaths in the warfarin group versus 22 deaths in the placebo group, for a nonsignificant risk reduction of 0.31 (p = 0.19). The second Stroke Prevention in Atrial Fibrillation trial (SPAF II) was designed to assess the relative efficacy of aspirin and warfarin in different age groups. 11 Warfarin was compared with aspirin in two parallel trials involving 715 patients aged 75 years or less (group I) and 385 patients older than 75 years (group II). The study population included 416 of the original SPAF I warfarin-eligible patients who continued their original therapy, plus 265 addi- tional patients primarily from the SPAF I placebo group who were rerandomized. The primary events were ischemic stroke and systemic embolism. During the trial, 10.5% of warfarin-treated patients and 4.3% of aspirintreated patients withdrew from therapy. Compared with aspirin, the annual relative risk of primary events was 0.67 in the younger patients taking warfarin (p = 0.24) and 0.73 in the older patients taking warfarin (p =0.39). The annual rates of major hemorrhages for warfarin and aspirin, respectively, were 1.7% versus 0.9% (p =0.17) for younger patients and 4.2% versus 1.6% (p =0.04) for older patients. The rates of major hemorrhages with warfarin were higher in the older patients despite a similar intensity of anticoagulation (p = 0.008). In the younger age group, six ICHs occurred in warfarin patients compared with two in aspirin patients. For the older patients, seven ICHs occurred with warfarin versus three with aspirin. The findings of SPAF I and II suggested that warfarin was more effective than aspirin in high-risk patients; however, warfarin therapy is complicated by the risk of hemorrhagic complications and the need for frequent laboratory monitoring. The third Stroke Prevention in Atrial Fibrillation trial (SPAF III) was designed to determine whether a combination of low-intensity, fixed-dose warfarin and aspirin would provide a safe and effective alternative to traditional adjusted-dose warfarin in these high-risk patients." The study population included patients with NVAF and at least one of the following additional thromboembolic risk factors: impaired left ventricular function, previous thromboembolism, systolic blood pressure above 160 mm Hg at enrollment, or being a woman older than 75 years. The primary end point was ischemic stroke or systemic thromboembolism. Combination therapy consisted of aspirin 325 mg daily plus fixed doses of warfarin ranging from 0.5 to 3.0 mg daily (mean 2.0 mg) to increase the international normalized ratio (INR) to Although the INR was monitored, the warfarin dose was not changed unless the INR was greater than 3.0 or bleeding occurred. Patients assigned to adjusted-dose warfarin received doses (mean 3.9 mg) sufficient to prolong the INR to a target range of For this group, the INR was checked monthly and dosage adjustments were made as needed. Aspirin compliance was assessed by pill count. The trial was stopped prematurely after the second interim analysis showed the superiority of adjusted-dose warfarin over the combination regimen. After adjustment for baseline differences in blood pressure, the relative risk reduction with adjusted-dose warfarin was 74% (95% CI 50% to 87%; p < ). The number of disabling strokes was also significantly less with adjusted-dose warfarin compared with combination therapy (1.7% vs. 5.6%, respectively; p = ). The substantial number of primary events in patients with INRs between 1.2 and 1.5 suggested that this intensity of anticoagulation was inadequate for stroke prophylaxis in high-risk patients. In patients with INRs below 1.2, an annual primary event rate of II % was observed. Five ICHs (3 fatal) occurred with combination therapy compared with three ICHs (2 fatal) in patients receiving adjusted-dose warfarin therapy. Minor bleeding leading to discontinuation of therapy occurred in 1.2% of patients re- The AnnalsofPharmacotherapy 1997 October, Volume 3] 1189

4 ceiving combination therapy versus 0.7% of those receiving adjusted-dose warfarin. The low incidence of hemorrhagic complications in this trial should be interpreted with caution, since the early termination was based on differences in thromboembolic rates rather than hemorrhagic complications. The wide confidence intervals for the rates of hemorrhage suggest that the incidence of bleeding may not have been accurately defined. After completion of the first five randomized trials (i.e., BAATAF, AFASAK, SPAF I, CAFA, SPINAF), a pooled analysis was conducted to compare the relative efficacy of warfarin or aspirin versus placebo." Comorbidities for randomized patients included a history of hypertension (50%), heart failure (20%), angina (23%), and diabetes (14%). Three of the trials (i.e., BAATAF, SPAF I, SPINAF) used the prothrombin time ratio (PTR), whereas two trials (i.e., AFASAK, CAFA) used the INR to adjust the warfarin dose. Aspirin was administered at either 75 or 325 mg daily. The primary end points for this collaborative analysis were ischemic stroke and major hemorrhage, which was defined as bleeding resulting in an ICH, hospitalization, or transfusion. Compared with placebo, the risk of stroke was reduced 68% (95% CI 50% to 79%) with warfarin and 36% (95% CI 4% to 57%) with aspirin. The annual rate of major hemorrhages was 1.0% for both placebo and aspirin and 1.3% for warfarin. ClinicalTrialEvidence The pooled analysis" of the original five trials was conducted after a determination that the trial protocols and baseline characteristics of the randomized patients were similar. All of the trials were terminated early due to the overwhelming evidence of the efficacy of stroke prophylaxis compared with control patients. The inclusion and exclusion criteria were relatively similar for all of the trials except that AFASAK and SPINAF excluded patients with intermittent AF. Although these were primary prevention trials, the percentage of patients with a prior history of stroke or TIA varied from 3% in BAATAF to 8% in SPINAF. Warfarin therapy was double-blind in the CAFA and SPINAF studies. The collaborative analysis was performed with and without data from BAATAF patients who chose to use aspirin on their own and, since the results were similar, all patients were included in the pooled analysis. An important variable in these studies is the intensity of anticoagulation. The recommended method for monitoring warfarin is the INR, which is a mathematical correction that adjusts for variations in the sensitivity of the thromboplastin reagent used to perform the PT test. If the INR is not used, two PTs of the same value (e.g., for 15 sec) may actually reflect different intensities of anticoagulation, depending on the sensitivity of the reagent used." This is a particular problem when the tests are performed TRIAL BAATAF (1990)" AFASAK (1989)' SPAFI (1991)8 CAFA (1991)" SPINAF (1992)10 SPAF II (1994)11 SPAF III (1996)12 Pooled analysis lj b by multiple laboratories, as was the case for most of these trials. Two of the studies" adjusted warfarin dosage on the basis of the INR. The remaining three studies 6,8.lo used the PTR and then later tried to estimate the INRs based on reagent data obtained from the laboratories or by using sensitivity values for typical thromboplastins. However, a survey conducted during approximately the same time period found that thromboplastin reagents commonly used in the US had International Sensitivity Index values ranging from 1.2 to Thus, there is considerable variation and uncertainty in the actual level of anticoagulation achieved in these trials. Furthermore, substantial numbers of patients were not actually in the target PT ranges during parts of the trials (Table 2). Analysis suggests that many thrombotic strokes occurred when the PTR was low, ranging from 1.0 to 1.2. Conversely, many hemorrhagic events occurred when the PTR or INR was above the target range. In addition, noncompliance and discontinuation of warfarin were problems in all of the trials, and many strokes occurred at a time when patients were not actually receiving therapy. In SPAF II,II for example, 12 of the 28 primary events in the warfarin group occurred when patients were not taking warfarin. Despite these limitations, the findings of the randomized trials were remarkably consistent. Warfarin was effective in all subgroups, including both young and old patients. In the pooled analysis," the risk reduction with warfarin was 84% (95% CI 55% to 95%) for women and 60% (95% CI 35% to 76%) for men. In contrast to the findings obtained with warfarin, the aspirin data were considerably more variable. The pooled data" were derived from two studies, AFASAK' and SPAF Table 2. Hemorrhagic Complications with Warfarin in the Nonvalvular Atrial Fibrillation Trials % OF TIME % ANNUAL BLEEDING MEAN WARFARIN WAS EVENTS WITH WARFARIN AGE(y) IN RANGE ABOVE RANGE MAJOR ICH MINOR " NR NR NR NR AFASAK =Copenhagen Atrial Fibrillation. Aspirin, Anticoagulation study; BAATAF =Boston Area Anticoagulation Trial for Atrial Fibrillation; CAFA =Canadian Atrial Fibrillation Anticoagulation study; ICH =intracranial hemorrhage; INR =international normalized ratio; NR =not reported; PTR =prothrombin time ratio; SPAF =Stroke Prevention in Atrial Fibrillation; SPINAF=Stroke Prevention in Nonrheumatic Atrial Fibrillation. "Median age. bpooled analysis ofbaataf, AFASAK. CAFA. SPAF I. and SPINAF The AnnalsofPharmacotherapy 1997 October, Volume 31

5 Stroke Prevention in Nonvalvular Atrial Fibrillation 1,8 which found a risk reduction of 18% (p = NS) and 44% (p = 0.02), respectively, compared with placebo. These variable findings have not been fully explained; however, there were important differences in the two studies. Patients in AFASAK received a lower dose of aspirin than did those in SPAF (75 vs. 325 mg, respectively). This appears to be an unlikely explanation since the benefits of antiplatelet therapy with similar low doses of aspirin have been demonstrated for other indications such as unstable angina and myocardial infarction. IS A more likely explanation is the fact that the AFASAK patients were significantly older and had more underlying conditions such as heart failure and hypertension that increased their overall risk for stroke. Subgroup analysis of the control patients taking aspirin in the BAATAF study demonstrated no benefit; however, definitive conclusions cannot be drawn since the aspirin use was uncontrolled. SPAF Il" found that aspirin was less effective than warfarin, particularly in older patients; however, the risk of major hemorrhage was also decreased. None of the observed differences in primary event rates in the SPAF II trial showed a statistically significant difference between warfarin and aspirin. The relative efficacy of warfarin and aspirin may depend primarily on the type of stroke. Investigators from the SPAF I trial reported that aspirin prevented noncardioembolic strokes more effectively than cardioembolic strokes, compared with placebo (p = 0.01).16 Similarly, analysis of the SPAF II data found that warfarin was significantly more effective than aspirin in preventing cardioembolic strokes (p = 0.005).17 Since it has been estimated that as many as 65-75% of all strokes in patients with AF are cardioembolic.v? this may explain the greater efficacy observed with warfarin. The benefits of antitbrombotic therapy must be weighed against the potential risks for bleeding complications, especially with warfarin (Table 2). The major concern is ICH. In the pooled analysis, the six patients who had ICH while taking warfarin had both higher systolic blood pressure (p = 0.001) and higher diastolic blood pressure (p = 0.016) at the time of randomization compared with those without ICH.l3 The mean age for patients with and without ICH was 73 and 69 years, respectively. The age difference was not statistically significant; however, the small number of patients with ICH precludes any reliable conclusions about the effects of age. At the time of the six ICHs, two patients had INRs of and four had PTRs of These findings suggest that the risk of bleeding may be minimized by using a lower intensity of anticoagulation. The most serious concerns about the safety of warfarin were raised by the SPAF II trial." In the subgroup of patients older than 75 years, the beneficial effects of warfarin were negated to a large extent by the high incidence of ICH. It has been suggested that the higher hemorrhagic rates in SPAF II were directly attributable to the target anticoagulation ranges (PTR ; approximate INR ), which were relatively high compared with those of other trials. A multivariate analysis was performed on the SPAF II data to identify potential risk factors for bleeding." The only baseline characteristics that significantly and independently predicted warfarin-related major bleeding were advanced age (p = 0.006) and increasing number of prescrip- tion medications (p =0.007). The latter finding suggests that drug interactions or patient confusion due to complex drug regimens may have played a significant role. The intensity of anticoagulation, as measured by the mean PTR, was also found to be an independent risk factor for bleeding in patients older than 75 years (p =0.02). The annual rate of major bleeding was 2.7% (95% CI 1.4% to 5.5%) in patients with a mean PTR of 1.5 (approximate INR 3.0) or less compared with 9.0% (95% CI 4.5% to 18%) for those whose mean PTR was over 1.5. From the individual trials and the pooled analysis, it is clear that the risk of stroke for patients with NYAF is not uniform.v" Several important risk factors were identified, including advancing age, previous stroke or TIA, diabetes, and hypertension. Patients less than 60 years of age with no additional risk factors (lone AF) had an annual risk of stroke of 1% or less. Thus, it is unlikely that this subgroup could benefit from anticoagulation therapy. Patients between 60 and 75 years of age with lone AF had a 2% annual risk of stroke. These patients could be adequately protected with aspirin. On the other hand, patients with additional risk factors and those older than 75 years have at least a 5-8% annual risk for stroke and clearly derive the most benefit from warfarin therapy. The annual stroke rate for patients receiving warfarin in the five pooled trials was less than 2% for all subgroups. Overall, 47% of the strokes that occurred were fatal or resulted in functional impairment. Generalizability ofresults In many respects the patients enrolled in the NYAF trials were typical of those seen in clinical practice. Trial patients were elderly, with a number of comorbid conditions such as hypertension, diabetes, heart failure, and coronary artery disease. Unlike many previous cardiovascular trials, enough women (- 25%) were enrolled to demonstrate benefit in this subgroup. Despite these similarities, one cannot exclude the possibility that individuals who agree to participate in a long-term clinical trial may not represent the typical patient. In addition, the age distribution of patients in the trials did not accurately represent that seen in clinical practice. In the trials, the mean age was 69 years, with 20% of those receiving warfarin older than 75 years. In the US, the median age for individuals with AF is 75 years. Approximately 84% are older than 65 years and 32% are older than 80 years.' Thus, practitioners see a population that is considerably older than those in the clinical trials. The safety of warfarin in patients older than 75 years remains controversial. In a recent case-control study, age and the intensity of anticoagulation were identified as risk factors for ICH.19 In the SPAF II trial," bleeding complications were more frequent in patients older than 75 years. In contrast, a pooled analysis of patients older than 75 years from the other NVAF trials reported a 0.3% (95% CI 0.04% to 2.1%) annual rate of ICH with warfarin." When generalizing the trial results, one must keep in mind that the exclusion rates for the trials ranged from 53% to 93%. Most of these patients were excluded because they were considered to be at high risk for bleeding complications. There are a number of additional factors that may limit the applicability of these findings to clinical practice. Pa- TheAnnalsofPhannacotherapy 1997 October, Volume

6 tients enrolled in these trials were seen on a regular basis and their anticoagulation was closely monitored. Even under these ideal conditions, a significant number of patients discontinued warfarin or were noncompliant with their medication regimen. A comparison of the efficacy results for the intention-to-treat versus the on-treatment groups illustrates the importance of compliance. The intention-totreat results, however, are probably closer to reality. In clinical practice, many physicians do not yet use the INR or adhere to the recommended target ranges for anticoagulation. 21,22 Some physicians continue to follow the outdated practice of using loading doses of warfarin, which may further increase the risk for bleeding complications. The need for compliance and frequent laboratory monitoring, coupled with warfarin's potential for many significant drug interactions, may decrease the overall effectiveness of therapy in clinical practice. In particular, the concurrent use of nonsteroidal antiinflammatory agents (NSAIDs) in the elderly for the treatment of arthritis may significantly increase the potential for bleeding complications. To some extent, these problems may be minimized by extensive patient education and by referring patients whenever possible to anticoagulation clinics. Gottlieb and Salem-Schatz" conducted a retrospective analysis at a large health maintenance organization (HMO) to determine whether the results of the NVAF trials translated into effectiveness in clinical practice. Of238 patients with AF, 198 were eligible for warfarin therapy (no contraindications and had not refused entry in the study). Among these patients, 168 were offered anticoagulation (84.8%) and 156 were actually receiving anticoagulation, representing 78.8% of the eligible patients. Compared with the patients in the randomized trials, there were no significant differences in mean age or gender distribution; however, the HMO patients had a greater prevalence of comorbidities that are considered risk factors for stroke. The routine JYf monitoring interval for HMO patients was days compared with days for study patients. The JYfRs for HMO patients were in the "target range" ( times control) 50% of days compared with 68% of days for the clinical trials. The annual rates for stroke, major bleeding, and minor bleeding for the HMO patients were 1.3%, 0.6%, and 13.6%, respectively, which were not statistically different from the average rates for the pooled analysis" of 1.3%, 1.1%, and 11.1 %, respectively. This study" has several important limitations. First, although published in 1994, the study was conducted prior to the publication of all the clinical trials and the pooled analysis. Second, the fact that the INR was not in use prohibits definitive conclusions about the actual intensity of anticoagulation. The study did not perform a subset analysis on the oldest patients (i.e., >75 y). Finally, the HMO in this study had served as a recruitment site for the BAATAF trial, which may have increased physician awareness and acceptance of the clinical trial findings. A retrospective analysis evaluated the use of warfarin for NVAF in residents of a skilled nursing facility." A total of 76 residents older than 60 years had a diagnosis of chronic AF. Of these residents, 15% were taking warfarin, I% was receiving heparin, and 29% were administered aspirin. The remaining 55% were not taking any antithrombotic agent, despite the fact that only 28% of these residents had a documented contraindication. Anticoagulant use was not further stratified on the basis of age. In a follow-up survey of the facility medical staff, 86% of responding physicians believed that the benefits of warfarin therapy outweighed the risks. However, only 29% of these physicians would use warfarin in an AF patient with dementia or Parkinson's disease. The finding that warfarin and aspirin were underused in elderly patients with AF was further supported by a study conducted in a geriatric rehabilitation hospital." Of 102 NVAF patients with a median age of 82 years, 19 were receiving warfarin and 32 were taking aspirin. Although the incidence of potential contraindications for warfarin was high, 23% of the patients without a contraindication were not receiving anticoagulation therapy. The use of antithrombotic therapy for AF has also been evaluated in university teaching hospitals." Researchers reviewed the medical records of 309 NVAF patients from six medical centers that are members of the University HealthSystem Consortium. The mean age of the patients was 72 years and 82% of the patients had at least one additional risk factor for stroke. Contraindications for anticoagulation were present in 44% of the patients. At discharge, 31% of the patients with no contraindications were receiving warfarin, 32% were receiving aspirin, and 8% were taking both drugs. Interestingly, among patients with contraindications, the corresponding rates were 22%, 40%, and 6%, respectively. Physicians were more likely to prescribe warfarin to AF patients who had additional risk factors for stroke (p < 0.00 I). The number of patients receiving combined therapy with warfarin and aspirin is cause for concern in view of the findings of SPAF Physician attitudes about the use of anticoagulation in patients with NVAF were recently evaluated in a national study," Surveys were sent to 1189 randomly selected office-based physicians in three strata: primary care, neurology, and cardiology. Among the 450 respondents, 2.7% of cardiologists, 21% of neurologists, and 19.2% of primary care physicians reported that they would not use oral anticoagulation for patients with NVAF. Overall, only 52.4% of physicians were very likely and an additional 25.3% were somewhat likely to prescribe anticoagulation. Fewer physicians were willing to prescribe anticoagulation for a 75-year-oldpatient with the same clinical scenario as a 65 year-old patient. No characteristics of the physicians distinguished those who would or those who would not prescribe anticoagulation. The majority of physicians believed that warfarin was beneficial but was associated with a high risk of hemorrhage, especially in older patients. Interestingly, only 50% of the physicians used the recommended INR range of for their patients and most used a lower target range for a 75-year-old patient than for a 65 year-old patient (p < 0.04). Nonetheless, the results of this survey are considerably more encouraging than those of two older surveys performed in 1990 and 1991 prior to completion of the randomized trials. 28,29 Stafford and Singer" reviewed data from the National Ambulatory Medical Care Surveys in 1980, 1981, 1985, lj92 The Annals ofpharmacotherapy 1997 October, Volume 31

7 Stroke Prevention in Nonvalvular Atrial Fibrillation and 1989 through 1993 to analyze the use ofantithrombotic therapy for AF. Data were collected from 1062 patient visits to randomly selected office-based physicians. Warfarin use increased from 7% in 1980 and 1981 to 32% in 1992 and 1993 (p < 0.00 I). In 1992 and 1993, patients 80 years of age or older were less likely to be prescribed warfarin compared with younger patients (19% vs. 36%, respectively) although overall, more older patients were taking warfarin than in Cardiologists and general internists were more likely to prescribe warfarin than were family practitioners. Warfarin use also varied with geographic location. The percentage of patients receiving warfarin ranged from 46% for patients in the Midwest to 16% for patients in the South. There was no significant difference in the use of warfarin in men and women. From 1980 to 1993, the use of aspirin for AF increased from 3% to 10% (p = 0.00 I). A potential limitation of this study design is that drug use was based on the number of office visits rather than on the number of patients. Thus, the extent of warfarin use could have been overestimated, since these patients would be expected to have more office visits than those not taking warfarin. In another study of physician attitudes about the management of AF, surveys were analyzed from 904 physicians (20% response rate) including 385 cardiologists, 326 internists, and 193 electrophysiologists." For patients with chronic AF, approximately 83% of the surveyed physicians indicated that warfarin was the antithrombotic agent of choice. The majority of physicians (80%) favored a target INR range of Some physicians indicated that they preferred a combination of warfarin and aspirin; however, specific dosage ranges were not reported. Based on findings of the physician surveys, it appears that many physicians arbitrarily use a lower target INR range for elderly patients in the belief that this will provide adequate prophylaxis while minimizing the risk of bleeding complications. A recent case--control study" of 74 patients with NVAF suggests that INRs below 2.0 may not provide adequate protection against stroke. Compared with an INR of 2.0, the adjusted odds ratio for stroke increased from 2.0 (95% CI 1.6 to 2.4) if the INR was 1.7 to 6.0 (95% CI 3.6 to 9.8) if the INR was 1.3. In this study an INR greater than 2.0 conferred no additional efficacy. Based on event rates from the randomized NVAF trials, this study had statistical power equivalent to approximately 6000 person-years of prospective follow-up. Previous work by these same investigators suggested that the risk of ICH increased exponentially at INRs greater than These findings support the use of the currently recommended INR range of and suggest that tight control of the INR is needed to provide optimal efficacy and safety. In summary, despite the abundance of clinical trial data supporting the use of warfarin in patients with NVAF, there continue to be a number of barriers that may prevent effective application of these results to clinical practice. Cost-Effectiveness In the current healthcare climate, a key issue for any prophylactic therapy is that of cost-effectiveness. A 1991 Swedish study evaluated the cost-effectiveness of primary stroke prevention with oral anticoagulants or aspirin in patients with AF.33 The analysis was based on four of the randomized clinical trials'" and observational studies assessing the prevalence of AF and the incidence of stroke. 34,35 The study concluded that primary prevention was cost-effective, provided the risk of hemorrhagic stroke was low. With warfarin therapy, the total cost per stroke prevented varied from a savings of $16.00 (US$) if the rate of ICH was 0.3% to a cost of $43.00 if the rate of ICH was 2%. A net cost reduction was observed when the hemorrhagic rate was below 1.3%, which was similar to the rate observed in the pooled analysis of the clinical trials. At this complication rate, it was estimated that warfarin would prevent 1000 strokes per year and reduce the costs of stroke by approximately $29 million per year. Aspirin saved $43.00 per stroke prevented. This study has some important limitations. Like most economic analyses, estimates were calculated on the basis of assumptions from multiple studies, including two observational studies. The calculations did not take into account additional risk factors for stroke or medication noncompliance. Estimates for the lifetime cost of stroke were derived from unpublished data, prohibiting further analysis. A second cost-effectiveness study by Gage et al." evaluated quality-adjusted survival and marginal cost-effectiveness of antithrombotic therapy. A Markov model was used to analyze the expected outcomes from warfarin, aspirin, or no therapy over a lo-year period. The probabilities for stroke, hemorrhage, and death were obtained from randomized, controlled clinical trials. The quality-of-life data were obtained by interviewing 74 patients with AF. Costs were estimated from the literature review, phone survey, and Medicare reimbursement. In high-risk patients, defined as those with NVAF and two or more additional risk factors, warfarin saved $6200 in costs from stroke and TIA, which more than offset the $5500 in costs for warfarin monitoring and drug-induced hemorrhages. For a 65 year-old patient with NVAF and one risk factor for stroke, warfarin therapy cost $8000 per quality-adjusted life-year saved compared with aspirin. For a 65-year-old patient with lone NVAF (i.e., very low risk), warfarin cost $ per quality-adjusted life-year saved compared with aspirin. For patients who were not prescribed warfarin, aspirin was preferred over no therapy regardless of the presence of risk factors for both quality-adjusted survival and cost. In another cost study using a Markov decision analysis, Eckman et a1. 37 estimated that warfarin therapy for a young woman with AF would cost $1580 per quality-adjusted life-year gained compared with no therapy. This study did not compare warfarin with aspirin. Naglie and Detsky" calculated that the gain in quality-adjusted survival by prescribing warfarin versus aspirin was small, approximately 0.01 quality-adjusted life-year, but they used a twofold higher estimate for aspirin's efficacy, which is not supported by more recent data." These studies suggest that warfarin is cost-effective in high-risk patients with NVAF, provided the risks of hemorrhagic adverse effects is minimized. Therefore, achieving cost-effectiveness will no doubt depend on factors such as appropriate patient selection, physician expertise in antico- The AnnalsofPharmacotherapy 1997 October, Volume 31 ]]93

8 agulation, and the availability of effective systems for longterm follow-up and monitoring. Clinical Practice Guidelines AF is a common condition with a high potential for devastating medical and socioeconomic consequences. Because of the definitive findings of the randomized trials and because of the significant variations in clinical practice, efforts have been made to standardize care by developing practice guidelines. The American College of Chest Physicians fourth consensus conference on antithrombotic therapy recently published their revised guidelines." The panel recommends long-term warfarin therapy with an INR of for all patients over 65 years of age with AF and patients younger than 65 years of age with one or more of the following risk factors: previous TIA or stroke, hypertension, heart failure, diabetes, clinical coronary artery disease, mitral stenosis, prosthetic heart valves, or thyrotoxicosis. Those who decline warfarin or those with a contraindication should receive aspirin 325 mg daily. For patients younger than 65 years with no risk factors, either aspirin or no antithrombotic agent is recommended. For patients between 65 and 75 years of age with no additional risk factors, the decision to use warfarin or aspirin should be based on individual risk. These recommendations were based on the pooled findings of the five prospective, randomized NYAF trials.13 The American College of Physicians published guidelines for stroke prevention in July For patients with NYAF, warfarin is considered the antithrombotic drug of choice unless there are contraindications. Patients less than 60 years of age with no specific clinical or echocardiographic risk factors are considered low risk and do not benefit from warfarin therapy. To improve the benefit/risk ratio, the guidelines recommend closely monitoring the INR in a target range of Aspirin is recommended as an alternative for patients unwilling or unable to take warfarin. An algorithm for stroke prevention in AF was recently released by the Agency for Health Care Policy and Research." For patients with AF, the first step is to determine their risk for stroke. Low risk is defined as age less than 60 years with none of the following: previous TIA/stroke, hypertension, diabetes mellitus, heart failure, echocardiogram with left atrial enlargement, or global left ventricular dysfunction. Patients at low risk should receive aspirin unless contraindicated. High-risk patients should receive warfarin with a target INR of The INR should be monitored frequently during initiation of treatment and at least every 2 months thereafter. The practice guidelines are intended to educate practitioners, aid in decision making, and minimize the variability in care for patients with NYAF. Although the guidelines from the various agencies are fairly consistent, they have many of the limitations of the clinical trial evidence discussed earlier. Questions remain about the safety of warfarin in older patients (>75 y), who were not adequately represented in the trials. Due to a lack of definitive data, the clinical guidelines tend to put the decision for determining the benefit/risk ratio for the highest risk patients in the hands of the practitioner. While this approach may be appropriate, it reopens the door for greater variability in care. The guidelines also assume that compliance and appropriate levels of anticoagulation can be maintained longterm, which may not be possible in clinical practice. Finally, the guidelines do not recommend warfarin if there are contraindications; however, the decision as to what constitutes an appropriate contraindication is once again placed with the practitioner. For example, is warfarin absolutely contraindicated in all patients with Parkinson's disease or those with a history of falls? Until these issues are more fully addressed, it is likely that there will continue to be significant variations in care for patients with NYAF. Future Directions The efficacy of warfarin for high-risk individuals with NVAF and that of aspirin for low-risk individuals has been conclusively demonstrated in the published randomized trials. There remains a critical need for outcomes research to further assess the effectiveness of warfarin versus aspirin in everyday practice as well as the impact of implementing the clinical practice guidelines. These studies should focus on a number of primary outcomes, including the incidence of stroke and hemorrhagic complications, risk factors for bleeding, reasons for rehospitalizations, medication compliance, quality of life, and cost-effectiveness. Studies are also needed to evaluate the effect of warfarin drug interactions on patient outcomes. These interactions are often responsible for patients discontinuing warfarin or experiencing adverse effects, yet there is little research documenting the effect of such interactions on patient outcomes or healthcare costs. The patient populations for these studies must represent the real world and ideally include all patients except those with an absolute contraindication to warfarin or aspirin. Conclusions NYAF is a common arrhythmia that is associated with a substantial risk of stroke. The benefits and risks of warfarin therapy have been clearly documented in prospective, randomized, controlled clinical trials. The trials provide definitive evidence for the efficacy of warfarin in high-risk individuals. Questions remain, however, about the safety of warfarin, particularly in individuals over 75 years of age and those at risk for bleeding complications. Until these concerns are more fully addressed, it is likely that many physicians will continue to be reluctant to initiate warfarin therapy in older patients. Thus, despite the overwhelming evidence in favor of warfarin's efficacy, the effectiveness of this strategy will probably be considerably less in clinical practice. In view of the potentially devastating consequences of stroke, it is essential that these issues be resolved through ongoing outcomes research and education. ~ References I. Feinberg WM, Blackshear JL. Laupacia A. Prevalence, age distribution and gender of patientswith atrial fibrillation. Arch Intern Med 1995;155: The AnnalsofPharmacotherapy 1997 October, Volume 31

9 Stroke Prevention in Nonvalvular Atrial Fibrillation 2. Halperin JL, Hart RG. Atrial fibrillation and stroke: new ideas, persisting dilemmas (editorial). Stroke 1988;19: Albers GW. Atrial fibrillation and stroke: three new studies, three remaining questions. Arch Intern Med 1994; I 54: Gresham GE, Duncan PW, Stason WP, Adams HP, Adelman AM, Alexander DN, et al. Poststroke rehabilitation. Clinical Practice Guideline, No. 16, AHCPR Publication No Rockville, MD: US Department of Health and Human Services. Public Health Service, Agency for Health Care Policy and Research, May Dobkin B. The economic impact of stroke. Neurology 1995;45(suppl I): S6-S9. 6. The Boston Area Anticoagulation Trial for Atrial Fibrillation Investigators. The effect of low-dose warfarin on the risk of stroke in patients with nonrheumatic atrial fibrillation. N Engl J Med 1990;323: I50S-II. 7. Petersen P, Godfredsen J, Boysen G, Andersen ED, Andersen B. Placebo-controlled, randomized trial of warfarin and aspirin for prevention of thromboembolic complications in chronic atrial fibrillation. The Copenhagan AFASAK Study. Lancet 1989;1: The Stroke Prevention in Atrial Fibrillation Investigators. The Stroke Prevention in Atrial Fibrillation study: final results. Circulation 1991; 84:52i Connolly SJ, Laupacis A, Gent M, Roberts RS, Cairns JA, Joyner C, et al. Canadian Atrial Fibrillation Anticoagulation (CAFA) study. J Am Coil Cardioll99I;18: Ezekowitz MD, Bridgers SL, James KE, Carliner NH, Colling CL, Gornick CC, et ai., for the Veterans AffairsStroke Prevention in Nonrheumatic Atrial Fibrillation Investigators. Warfarin in the prevention of stroke associated with nonrheumatic atrial fibrillation. N Engl J Med 1992;327: II. The Stroke Prevention in Atrial Fibrillation Investigators. Warfarin versus aspirin for prevention of thromboembolism in atrial fibrillation: Stroke Prevention in Atrial Fibrillation II study. Lancet 1994;343: The Stroke Prevention in Atrial Fibrillation Investigators. Adjusted-dose warfarin versus low-intensity, fixed-dose warfarin plus aspirin for highrisk patients with atrial fibrillation: Stroke Prevention in Atrial Fibrillation ill randomized clinical trial. Lancet 1996;348: Atrial Fibrillation Investigators. Risk factors for stroke and efficacy of antithrombotic therapy in atrial fibrillation. Analysis of pooled data from five randomized controlled trials. Arch Intern Med 1994;154: Bussey HI, Force RW, Bianco TM, Leonard AD. Reliance on prothrombin time ratios causes significant errors in anticoagulation therapy. Arch Intern Med 1992; I52: IS. Antiplatelet Trialist's Collaboration. Collaborative overview of randomized trials of antiplatelet therapy. I. Prevention of death, MI and stroke by prolonged antiplatelet therapy in various categories of patients. BMJ 1994;308: Miller VT, Rothrock JF, Pearce LA, Feinberg WM, Hart RG, Anderson DC. Ischemic stroke in patients with atrial fibrillation: effect of aspirin according to stroke mechanism. Neurology 1993;43: Miller VT, Pearce LA, Feinberg WM, Rothrock JF, Anderson DC, Hart RG. Differential effect of aspirin versus warfarin on clinical stroke types in patients with atrial fibrillation. Neurology 1996;46: The Stroke Prevention in Atrial Fibrillation Investigators. Bleeding during anti-thrombotic therapy in patients with atrial fibrillation. Arch Intern Med 1996;156: Hylek EM, Singer DE. Risk factors for intracranial hemorrhage in outpatients taking warfarin. Ann Intern Med 1994;120: Connolly SJ, for the Atrial Fibrillation Investigators. Stroke Prevention in Atrial Fibrillation II study (letter). Lancet 1994;343: I Ansell JE. Imprecision of prothrombin time monitoring of oral anticoagulation. Am J Clin Pathol 1992;98: Morse EE, Panek S, Pisciotto P, Kiraly T, Bona R, Vose K, et al. Reemergence of the international normalized ratio for the standardization of prothrombin time. Ann Clin Lab Sci 1993;23: Gottlieb LK, Salem-Schatz S. Anticoagulation in atrial fibrillation. Does efficacy in clinical trials translate into effectiveness in practice? Arch Intern Med 1994;154: Leitner DA, McCart GM, Knoell K. Evaluation of anticoagulation for chronic skilled-nursing facility residents. Consult Pharm 1994;9: Lawson F, McAlister F, Ackman M, Ikuta R, Montague T. The utilization of anti-thrombotic prophylaxis for atrial fibrillation in a geriatric rehabilitation hospital. J Am Geriatr Soc 1996;44: Albers GW, Yim lm, Belew KM, Bittar N, Hattemer CR, Phillips BG, et al. Status of antithrombotic therapy for patients with atrial fibrillation in university hospitals. Arch Intern Med 1996;156: McCrory DC, Matchar DB, Samsa F, Sanders LL, Pritchett ELC. Physician attitudes about anticoagulation for nonvalvular atrial fibrillation in the elderly. Arch Intern Med 1995;155: Chang HJ, Bell JR, Deroo DB, Kirk JW, Wasson JH. Physician variation in anticoagulating patients with atrial fibrillation. Arch Intern Med 1990; 150: Kutner M, Nixon G, Silverstone F. Physicians' attitudes toward oral anticoagulants and anti platelet agents for stroke prevention in elderly patients with atrial fibrillation. Arch Intern Med 1991;151: Stafford RS, Singer DE. National patterns of warfarin use in atrial fibrillation. Arch Intern Med 1996;156: Brodsky MA, Chun JG, Podrid PJ, Douban S, Allen BJ, Cygan R. Regional attitudes of generalists, specialists and subspecialists about management of atrial fibrillation. Arch Intern Med 1996;I56: Hylek EM, Skates SJ, Sheehan MA, Singer DE. An analysis of the lowest effective intensity of prophylactic anticoagulation for patients with nonrheumatic atrial fibrillation. N Engl J Med 1996;335: Gustafsson C, Asplund K, Britton M, Norrving B, Olsson B, Marke LA. Cost effectiveness of stroke prevention in atrial fibrillation. Swedish national perspective. BMJ 1992;305: WolfPA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke 1991;22: Cairns JA, Conolly SJ. Nonrheumatic atrial fibrillation. Risk of stroke and role of antithrombotic therapy. Circulation 1991;84: Gage BF, Cardinali AB, Albers GW, Owens DK. Cost-effectiveness of warfarin and aspirin for prophylaxis of stroke in patients with atrial fibrillation. JAMA 1995;274: Eckman MH, Levine HJ, Pauker SG. Making decisions about antithromboric therapy in heart disease: decision analytic and cost-effectiveness issues. Chest 1995;108(suppI4):457S-70S. 38. Naglie G, Detsky AS. Treatment of chronic nonvalvular atrial fibrillation in the elderly: a decision analysis. Med Decis Making 1992;12: Barnett HJM, Eliasziw M, Meldrum HE. Drugs and surgery in the prevention of ischemic stroke. N Engl J Med 1995;332: Laupacis A, Albers G, Dalen J, Dunn M, Feinberg W, Jacobson A. Antithrombotic therapy in atrial fibrillation. Chest 1995;108(suppl):352S-9S. 41. American College of Physicians. Guidelines for medical treatment for stroke prevention. Ann Intern Med 1994; 121: Stroke prevention recommendations, AHCPR Publication No Rockville, MD: US Department of Health and Human Services. Public Health Service, Agency for Health Care Policy and Research, September EXTRACfO OB,JETIVO: Revisar los estudios clfnicos que evaluan el uso de warfarina como profilaxis primaria en fibrilaci6n atrial no-valvular (FANV) y discutir los beneficios y riesgos relativos de la terapia de warfarina versus aspirina. Revisar las gufas de practica clinica e identificar barreras potenciales para su implantaci6n en la practica clfnica, FUENTES DE INFORMACION: Se realiz6 una busqueda de la literatura utilizando MEDLINE para identificar estudios clfnicos de terapia antitromb6tica para FANV, gufas de practica clinica, estudios que evaluan las practicas y actitudes de los medicos, estudios de costoefectividad, y articulos de revisi6n pertinentes. Los terminos c1aves de la busqueda incluyeron fibrilaci6n atrial, apoplegfa, terapia antitromb6tica, warfarina, aspirina, y costo-efectividad. METODO DE EXTRACCION DE INFORMACION: Se seleccionaron para el analisis estudios prospectivos y aleatorios. Se revisaron las gufas de practica clfnica de paneles de expertos reconocidos. Se seleccionaron articulos de revisi6n comprensivos. SINTESIS: La FANV es una arritrnia cormin que se asocia con un riesgo sustancial de apoplegfa, Siete estudios clfnicos prospectivos y aleatorios han demostrado contundentemente la eficacia de warfarina para la prevenci6n de apoplegfa en los pacientes con FANV. Los mayores beneficios se obtienen en pacientes viejos y en pacientes con comorbilidad que aumente el riesgo de apoplegfa. Sin embargo, se debe pesar ei riesgo potencial de prevenir una apoplegfa contra el potencial de complicaciones de sangrado. Se ha demostrado que warfarina es costoefectiva en pacientes de alto riesgo siempre y cuando se minimizen las The AnnalsofPharmacotherapy 1997 October, Volume