Chest Email Content Delivery
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Guest Access | Sign In via User Name/Password
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Shireman, T. I.
Right arrow Articles by Ellerbeck, E. F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Shireman, T. I.
Right arrow Articles by Ellerbeck, E. F.
(Chest. 2006;130:1390-1396.)
© 2006 American College of Chest Physicians

Development of a Contemporary Bleeding Risk Model for Elderly Warfarin Recipients*

Theresa I. Shireman, PhD, RPh; Jonathan D. Mahnken, PhD; Patricia A. Howard, PharmD; Timothy F. Kresowik, MD; Qingjiang Hou, MS and Edward F. Ellerbeck, MD, MPH

* From the Departments of Preventive Medicine & Public Health (Drs. Shireman, Mahnken, Ellerbeck and Mr. Hou) and Pharmacy Practice (Dr. Howard), University of Kansas Medical Center, Kansas City, KS; and the Department of Surgery (Dr. Kresowik), University of Iowa College of Medicine, Iowa City, IA.

Correspondence to: Theresa I. Shireman, PhD, RPh, Preventive Medicine & Public Health, University of Kansas Medical Center, 3901 Rainbow Blvd, Mail Stop 1008, Kansas City, KS 66160; e-mail: tshireman{at}kumc.edu

Abstract

Background and purpose: Develop and validate a contemporary bleeding risk model to guide the clinical use of warfarin in the elderly atrial fibrillation (AF) population.

Methods: Chart-abstracted data from the National Registry of Atrial Fibrillation was combined with Medicare part A claims to identify major bleeding events requiring hospitalization. Using a split-sample technique, candidate variables that provided statistically stable relationships with major bleeding events were selected for model development. Three risk categories were created and validated. The new model was compared to existing bleeding risk models using c-statistics and Kaplan-Meier curves.

Results: Model development and validation was conducted on 26,345 AF patients who were > 65 years of age and had been discharged from the hospital while receiving warfarin therapy. The following eight variables were included in the final risk score model: age ≥ 70 years; gender; remote bleeding; recent (ie, during index hospitalization) bleeding; alcohol/drug abuse; diabetes; anemia; and antiplatelet use. Bleeding rates were 0.9%, 2.0%, and 5.4%, respectively, for the groups with low, moderate, and high risk, compared to the bleeding rates for groups with moderate risk (1.5% and 1.0%) and high risk (1.8% and 2.5%) from other models.

Conclusions: Using a nationally derived data set, we developed a model based on contemporary practice standards for determining major bleeding risk among AF patients receiving warfarin therapy. The larger sample size afforded the opportunity to incorporate additional risk factors. In addition, since the majority of our population was > 65 years of age, we had greater ability to stratify risk among the elderly.

Key Words: adverse effects • anticoagulants • atrial fibrillation • complications • drug therapy • epidemiology • hemorrhage • risk • therapeutic use • warfarin

Warfarin has been proven to be the most effective antithrombotic agent for the primary or secondary prevention of stroke in patients with atrial fibrillation (AF),1 yet it is associated with a risk of major bleeding events. Numerous studies2345 have documented the underutilization of warfarin therapy in eligible patients with AF, and physician surveys have shown that, while most believe warfarin is efficacious, fears of bleeding may prevent its utilization.6

Previously, two bleeding risk models have been developed and validated to assist clinicians in balancing the risks and benefits of antithrombotic therapy. The Outpatient Bleeding Risk Index (OBRI) considers age > 65 years, prior stroke, prior GI bleeding, and any of four comorbidities (eg, recent myocardial infarction [MI], anemia, diabetes, or renal insufficiency) in order to stratify patients into three risk groups.78910 Kuijer et al11 developed another model that uses age > 60 years, gender, and malignancy to stratify patients into three risk groups.

While these models have proven validity, there remains a need for a model that more fully addresses the contemporary use of warfarin therapy in elderly AF patients. Both the OBRI model and the Kuijer et al11 model stratified risk by age at 60 to 65 years, while the mean age of AF patients is 75 years.1 Since advanced age has consistently been associated with the underutilization of warfarin therapy,1312 a model that can address issues among the elderly is clearly needed. Second, these models did not include many other known factors that may be predictive of bleeding.781113141516171819 Given the extremely low incidence of the bleeding events, much larger sample sizes are needed to adequately model additional risk factors. Third, while there are an increasing number of medications that may alter bleeding risks when used concurrently with warfarin,2021 previous bleeding risk models222324 have not accounted for common drug interactions in AF patients. Finally, prior models13141516 were primarily developed prior to the use of lower intensity warfarin anticoagulation therapy. The goals of our study were to develop a contemporary bleeding risk model for users of warfarin to help predict a given elderly patient’s risk for major bleeding events, to examine the capability of this new model to predict bleeding events in a validation cohort, and to compare our new model with predictions based on previous bleeding risk models.

Materials and Methods

The study design was a retrospective cohort analysis using the National Registry of Atrial Fibrillation. Briefly, the Centers for Medicare & Medicaid Services authorized the collection of data on a national, random sample of patients discharged from the hospital between April 1998 and March 1999, and between July 2000 and June 2001 with a diagnosis of AF (International Classification of Diseases, ninth revision, clinical modification, code 427.31) as part of the National Stoke Project.45 We used the hospitalization selected in this random sample as the index hospitalization.

Patients who were at least 65 years of age were included in the present study if they had been discharged from the hospital while receiving warfarin therapy. The exclusion criteria included situations in which inadequate follow-up data or hospital discharge information would have been available, as follows: transfer to another acute care hospital; death during the hospitalization; discharged from the hospital against medical advice; and enrollment in a Medicare managed care organization during the 90-day follow-up period.

Data Sources
The following four sources of data were linked to carry out this study: (1) the initial Medicare part A claim identifying the index AF hospitalization; (2) the Medicare enrollment database to identify dates of death and managed care enrollment; (3) data abstracted from the medical record for the index AF hospitalization; and (4) 90-day follow-up Medicare part A claims. Data were obtained through collaboration with the Iowa Foundation for Medical Care and the Centers for Medicare & Medicaid Services. Data reliability for the chart-abstracted data ranged from 80 to 95%; the median interrater reliability was 90%.4

Study Variables
The primary outcome variable was hospitalization for a major acute bleeding event within 90 days of the hospital discharge date from the index AF hospital admission for either a GI hemorrhage (Diagnosis-Related Group code 174 or 175) or intracranial hemorrhage (International Classification of Diseases, ninth revision, clinical modification, codes 430 [subarachnoid], 431 [intracerebral], and 432 [other intracranial hemorrhage]).242526 We classified bleeding events by the presence/absence of a bleeding event and the time to the bleeding event (ie, the number of days between hospital discharge and readmission for a major bleeding event). Only the first bleeding event for a study subject was included in the analysis.

Predictor variables were determined from the chart-abstracted data. A host of risk factors for bleeding events have been identified in the literature.7811131415161718192021 In general, these risk factors can be grouped into demographics (age, gender, and nursing facility residence), current or remote bleeding event, alcohol/drug abuse, concomitant diseases (anemia, cancer, stroke, transient ischemic attacks, MI, hypertension, heart failure/cardiomyopathy, ischemic heart disease, diabetes, hepatic failure, or peptic ulcer disease), concomitant risks for injury (risk for falls, cognitive impairment, or surgery during index hospitalization), and concurrent medications that can influence hemostasis. Anemia was defined as a hematocrit of < 30% during the index hospitalization. Concurrent medications prescribed at hospital discharge included aspirin, ticlopidine, and clopidogrel, and agents from the following classes that have documented interactions: nonsteroidal antiinflammatory drugs (NSAIDs); dipyridamole; antibiotics; antiarrhythmic agents; histamine-2 antagonists; proton pump inhibitors; antifungal agents; pulmonary drugs; leukotriene antagonist agents; and lipid-lowering agents. We also tracked whether warfarin therapy was initiated during the index hospitalization or the subject had been admitted to the hospital while receiving warfarin therapy.

Statistical Analysis
We randomly split the sample for model development (75%) and validation (25%) using a random number generator. Bivariate relationships between the predictors and bleeding risk were examined using Kaplan-Meier curves, the log rank test, and Cox regression hazard ratios (HRs). Predictors with sufficient numbers and potential association indicated by at least one of the bivariate methods were considered for inclusion in the multivariable model.

Stepwise selection was used to determine the final Cox regression model (p < 0.15 to enter the study; p < 0.1 to remain in the study). Selected predictor variables were checked for collinearity. Risk scores were then created for each subject based on their adjusted HR of a bleeding event from the final model. All possible sets of cut points to determine a low, moderate, or high bleeding risk based on these scores were tested as predictors of a bleeding event using the development cohort. The cut points with the largest log-likelihood were selected as the optimal set of cut points for the contemporary bleeding risk model.

Risk scores were then generated for subjects in the validation sample and categorized based on the cut points as low, moderate, or high risk. The Pearson {chi}2 statistic was used to compare the proportion of bleeding events across risk categories. Unconditional logistic regression predicting bleeding events by risk category was used to perform pairwise comparisons of bleeding event proportions and to estimate the area under the receiver operating characteristic (ROC) curve.27 (The Bonferroni correction was used for these pairwise comparisons.28) Area under the ROC curve was estimated using the c-statistic.29 Kaplan-Meier survival curves were generated for each risk category, and the HRs of a bleeding event were compared using the log-rank test for trend.30

Using the validation sample, subjects were categorized into risk groups based on the model of Kuijer et al11 and the OBRI model. Modifications were made to best fit this risk model to the data available in our data set. The OBRI model combined all prior bleeding events, while we differentiated between bleeding events occurring during index hospitalization and remote bleeding events. Likewise, the OBRI included only recent MIs, whereas our data only contained a variable indicating any prior MI. Last, our data did not allow for us to examine the effect of renal insufficiency, thereby excluding one of the comorbidities from the OBRI model. Because all subjects were > 65 years old, there was no low-risk group in either the model of Kuijer et al11 or the OBRI model. The comparisons of our model to the Kuijer et al11 and OBRI models followed the same steps as our validation process. The University of Kansas Advisory Committee on Human Experimentation approved the research protocol.

Results

The cohort consisted of 76,177 unique observations. Reasons and counts for exclusion were made in the following order: failure to match with the Medicare denominator file; death before hospital discharge; discharged from the hospital against medical advice or discharged to an acute care hospital (n = 6,666); managed care enrollment (n = 3,215); no warfarin prescribed at the time of hospital discharge (n = 38,089); and age < 65 years (n = 1,862). There were 26,345 study subjects remaining for model development and validation. Cohort characteristics are found in Table 1 .


View this table:
[in this window]
[in a new window]

  		Table 1. Demographic, Clinical, and Outcome Descriptions of the AF Bleeding Risk Model Development and Validation Cohorts*

 
The following 18 variables were considered for inclusion in the multivariable model: age; gender; remote bleeding event; recent bleeding event; alcohol/drug abuse; stroke; hypertension; heart failure/cardiomyopathy; diabetes; anemia; peptic ulcer disease; cognitive impairment (ie, dementia, Alzheimer disease, or schizophrenia); risk for falls; discharged from the hospital while receiving an antiplatelet medication; discharged from the hospital while receiving a NSAID; discharged from the hospital while receiving an antibiotic; discharged from the hospital while receiving another medication with potential bleeding effects; and newly prescribed warfarin. After the stepwise selection process, eight variables remained in the model (Table 2 ). The resulting equation combining the eight risk factors to compute an individual’s risk score, as follows:

Formula
where Xj equals 1 when the specific characteristic is present and 0 otherwise. Based on the results of the log-likelihood scores, bleeding risk scores of ≤ 1.07, > 1.07 but < 2.19, and ≥ 2.19, respectively, were classified as low, moderate, and high risk.


View this table:
[in this window]
[in a new window]

  		Table 2. Final Variables Included in the Bleeding Risk Model Development*

 
In the validation sample, 35 of the low-risk individuals (0.9%), 48 of the moderate-risk individuals (2.0%), and 12 of the high-risk individuals (5.4%) had major bleeding events (p < 0.0001). Looking further at the pairwise comparisons, we found a statistically significant difference between low-risk and high-risk subjects (p < 0.0001), between the low-risk moderate-risk subjects (p < 0.001), and between moderate-risk and high-risk subjects (p < 0.01). The area under the ROC curve was 0.632, a level between "no discrimination" and "acceptable discrimination."27 Kaplan-Meier curves for the three risk categories are shown in Figure 1 , top left, A.


Figure 1
View larger version (18K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. Kaplan-Meier curves for each bleeding risk model. Top left, A: current study. Top right, B: model of Kuijer et al.11 Bottom, C: OBRI model.78910

 
In the application of the Kuijer et al11 model to the validation sample, 1.5% of the intermediate-risk individuals experienced bleeding events, and 1.8% of those individuals who were at high risk experienced bleeding events (p = 0.74); the area under the ROC curve was 0.503, representing "no discrimination."27 The Kaplan-Meier curves (Fig 1, top right, B) for these two risk categories showed no significant difference between the high-risk and intermediate-risk groups (p = 0.75).

The OBRI model78910 performed better than the Kuijer et al11 model, with 1.0% of the intermediate-risk individuals and 2.5% of the high-risk individuals having experienced bleeding events (p < 0.0001). The area under the ROC curve was 0.613, representing a level between "no discrimination" and "acceptable discrimination."27 Kaplan-Meier curves (Fig 1, bottom, C) also showed the expected trends in the risk of bleeding events (p < 0.0001).

Discussion

We were able to identify eight variables that were significant predictors of major bleeding events. The resulting model combining these risk factors performed well in differentiating among low-risk, moderate-risk, and high-risk groups, and likewise performed well when compared with two previous bleeding risk models. Our contemporary model improved the discrimination between subjects at risk for a bleeding event based on areas under the ROC curve and had three categories that demonstrated a trend toward increasing bleeding risk with increasing risk category. The bleeding event rate was greatest in the high-risk group from our contemporary model. In addition, the bleeding rate of the high-risk group was more than twice that of the moderate-risk group in our model, and more than twice that of the high-risk categories from the other models.

Our entire study population was > 65 years of age, and 43% of subjects were ≥ 80 years of age, which is a substantially higher rate than that reported in either of the previous models. For instance, in the second OBRI model,8 only 6% of persons in the derivation and validation cohorts were > 80 years of age. Advanced age has consistently been associated with the underutilization of warfarin therapy, despite the fact that benefits are potentially greater since the risk of stroke and death increase with age.1312 Women have inconsistently been found to have higher bleeding rates.11131819

A clinically important advantage of our model over the two existing models is the identification of additional risk factors and our ability to separately quantify the effects of each risk factor. The added bleeding risk associated with concomitant antiplatelet therapy is particularly important given its prevalent use in treating comorbidities such as coronary heart disease. Approximately 20% of AF patients received concomitant warfarin-antiplatelet therapy,2224 and antiplatelet therapy increases major bleeding rates in warfarin users, as noted in a previous analysis of these data and other work.2431

In examining the weights for the eight risk factors, alcohol/drug abuse and anemia were particularly important risk factors. While the OBRI model78910 includes anemia as one of several important comorbidities, it did not include alcohol/drug abuse. Chronic alcohol/drug abuse is likely to cause liver impairment, and although it is considered to be a contraindication to warfarin therapy, its presence was noted in nearly 2% of the study cohort.

The finding that bleeding risk was increased in diabetic patients is significant given the epidemiologic association between AF and diabetes,3233 and the increasing prevalence of this comorbidity. Our findings confirm recent findings34 showing increased bleeding risk in AF patients with diabetes. As with the Kuijer et al model11 and other studies,12131819 we found increased bleeding rates among women, although gender was not included in the OBRI model.78910

Limitations to our approach included the lack of ability to track compliance with, continuation of, or initiation of drug therapy after hospitalization and the inability to observe clinical laboratory values. We were left with the assumption that patients continued to receive the therapies they had been prescribed at hospital discharge throughout the study period, which is an intention-to-treat approach. We limited the follow-up time to 90 days, although this time frame should be adequate since studies1519353637 have shown that major bleeding events most often occur early in the course of warfarin therapy. However, we found no significant difference in the number of bleeding events between those who had been admitted to the hospital while receiving warfarin therapy and new users.

International normalized ratio (INR) values at hospital discharge or during follow-up were not documented in our data set. While INRs are used to monitor warfarin levels and to adjust therapy in order to avoid bleeding events, additional patient characteristics must also be considered. The availability of INR values may have improved the strength of the predictive model; however, the ability to stratify risk indicates that there was significant variation in bleeding events based on other patient-specific factors. These variables can assist clinicians who want to estimate the bleeding risk prior to drug initiation and INR monitoring.

Finally, we only examined risk among persons receiving warfarin therapy, and it is quite likely that additional AF patients who were not discharged from the hospital while warfarin therapy would have benefited from antithrombotic therapy. The Stroke Patient Outcomes Research Team concluded that improved antithrombotic therapy in AF patients could prevent an additional 40,000 strokes annually and save an estimated $600 million.38 Nonetheless, clinicians must weigh the potential benefits against each individual patient’s risk for major bleeding. For example, although diabetes and age were independent risk factors for bleeding, they were also independent risk factors for stroke. The 5.4% risk of major bleeding events that was seen in our group with the highest risk of stroke must be weighed against the > 7% annual stroke risk faced by the highest risk patients with AF.39 At the patient level, this tradeoff can be difficult for clinicians to estimate.40

Summary

Using a nationally derived data set, we have developed a model based on contemporary practice standards for determining the risk of major bleeding events among AF patients receiving warfarin therapy. Risk scores can be computed for individuals considering the following eight separately weighted factors: age; gender; recent bleeding event; remote bleeding event; alcohol/drug abuse; diabetes; anemia; and concomitant antiplatelet therapy. Validation of this model in other study populations is warranted.

Acknowledgements

The authors acknowledge the assistance of the Iowa Foundation for Medical Care and the Centers for Medicare & Medicaid Services in providing the data that made this research possible.

Footnotes

Abbreviations: AF = atrial fibrillation; HR = hazard ratio; INR = international normalized ratio; MI = myocardial infarction; NSAID = nonsteroidal antiinflammatory drug; OBRI = Outpatient Bleeding Risk Index; ROC = receiver operating characteristic

The American Heart Association provided funding for this project through grant-in-aid No. 0450071Z.

The conclusions presented are solely those of the authors and do not represent those of the Iowa Foundation for Medical Care or the Centers for Medicare & Medicaid Services.

The authors have reported to the ACCP that no significant conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.

Received for publication June 8, 2006. Accepted for publication June 13, 2006.

References

  1. . Atrial Fibrillation Investigators. (1994) Risk factors for stroke and efficacy of antithrombotic therapy in atrial fibrillation: analysis of pooled data from five randomized controlled trials. Arch Intern Med 154,1449-1457[Abstract]
  2. Stafford, RS, Singer, DE Recent national patterns of warfarin use in atrial fibrillation. Circulation 1998;97,1231-1233[Abstract/Free Full Text]
  3. Gage, BF, Boechler, M, Doggette, AL, et al Adverse outcomes and predictors of underuse of antithrombotic therapy in Medicare beneficiaries with chronic atrial fibrillation. Stroke 2000;31,822-827[Abstract/Free Full Text]
  4. Jencks, SF, Cuerdon, T, Burwen, DR, et al Quality of medical care delivered to Medicare beneficiaries: a profile at state and national levels. JAMA 2000;284,1670-1676[Abstract/Free Full Text]
  5. Jencks, SF, Huff, ED, Cuerdon, T Change in the quality of care delivered to Medicare beneficiaries, 1998–1999 to 2000–2001. JAMA 2003;289,305-312[Abstract/Free Full Text]
  6. McCrory, DC, Matchar, DB, Samsa, G, et al Physician attitudes about anticoagulation for nonvalvular atrial fibrillation in the elderly. Arch Intern Med 1995;155,277-281[Abstract]
  7. Landefeld, CS, Goldman, L Major bleeding in outpatients treated with warfarin: incidence and prediction by factors known at the start of outpatient therapy. Am J Med 1989;87,144-152[ISI][Medline]
  8. Beyth, RJ, Quinn, LM, Landefeld, CS Prospective evaluation of an index for predicting the risk of major bleeding in outpatients treated with warfarin. Am J Med 1998;105,91-99[CrossRef][ISI][Medline]
  9. Wells, PS, Forgie, MA, Simms, M, et al The Outpatient Bleeding Risk Index: validation of a tool for predicting bleeding rates in patients treated for deep venous thrombosis and pulmonary embolism. Arch Intern Med 2003;163,917-920[Abstract/Free Full Text]
  10. Aspinall, SL, DeSanzo, BE, Trilli, LE, et al Bleeding risk index in an anticoagulation clinic: assessment by indication and implications for care. J Gen Intern Med 2005;20,1008-1013[CrossRef][ISI][Medline]
  11. Kuijer, PMM, Hutten, BA, Prins, MH, et al Prediction of the risk of bleeding during anticoagulant treatment for venous thromboembolism Arch Intern Med 1999;159,457-460[Abstract/Free Full Text]
  12. Howard, PA, Ellerbeck, EF, Engelman, KK, et al Warfarin for stroke prevention in octogenarians with atrial fibrillation. Am J Geriatr Cardiol 2001;10,139-144[Medline]
  13. van der Meer, FJM, Rosendaal, FR, Vandenbroucke, JP, et al Bleeding complications in oral anticoagulant therapy: an analysis of risk factors. Arch Intern Med 1993;153,1557-1562[Abstract]
  14. Hylek, EM, Singer, DE Risk factors for intracranial hemorrhage in outpatients taking warfarin. Ann Intern Med 1994;120,897-902[Abstract/Free Full Text]
  15. Fihn, SD, McDonell, M, Martin, D, et al Risk factors for complications of chronic anticoagulation. Ann Intern Med 1993;118,511-520[Abstract/Free Full Text]
  16. The Stroke Prevention in Atrial Fibrillation Investigators.. Bleeding during antithrombotic therapy in patients with atrial fibrillation. Arch Intern Med 1996;156,409-416[Abstract]
  17. Fihn, SD, Callahan, CM, Martin, DC, et al The risk for and severity of bleeding complications in elderly patients treated with warfarin. Ann Intern Med 1996;124,970-979[Abstract/Free Full Text]
  18. Wehinger, C, Stollberger, C, Langer, T, et al Evaluation of risk factors for stroke/embolism and of complications due to anticoagulant therapy in atrial fibrillation. Stroke 2001;32,2246-2252[Abstract/Free Full Text]
  19. Pengo, V, Legnani, C, Noventa, F, et al Oral anticoagulant therapy in patients with nonrheumatic atrial fibrillation and risk of bleeding: a multicenter inception cohort study. Thromb Haemost 2001;85,418-422[ISI][Medline]
  20. Howard, PA, Dunn, MI Update on drug interactions: warfarin and anti-infective agents. J Respir Dis 1998;19,1037-1046
  21. Freedman, MD, Olatidoye, AG Clinically significant drug interactions with the oral anticoagulants. Drug Saf 1994;10,381-394[ISI][Medline]
  22. Howard, PA, Ellerbeck, EF, Engelman, KK, et al The nature and frequency of potential warfarin drug interactions that increase the risk of bleeding in patients with atrial fibrillation. Pharmacoepidemiol Drug Saf 2002;11,569-576[CrossRef][ISI][Medline]
  23. Quilliam, BJ, Lapane, KL, Eaton, CB, et al Effects of antiplatelet and anticoagulant agents on risk of hospitalization for bleeding among a population of elderly nursing home stroke survivors. Stroke 2001;32,2299-2304[Abstract/Free Full Text]
  24. Shireman, TI, Howard, PA, Kresowik, TF, et al Combined anticoagulant-antiplatelet use and major bleeding events in elderly atrial fibrillation patients. Stroke 2004;35,2362-2367[Abstract/Free Full Text]
  25. Fisher, ES, Wennberg, JE, Stukel, TA, et al Hospital readmission rates for cohorts of Medicare beneficiaries in Boston and New Haven. N Engl J Med 1994;331,989-995[Abstract/Free Full Text]
  26. Tirschwell, DL, Longstreth, WT Validating administrative data in stroke research. Stroke 2002;33,2465-2470[Abstract/Free Full Text]
  27. Hosmer, DW, Lemeshow, S Applied logistic regression 2nd ed. 2000 John Wiley & Sons. New York, NY:
  28. Rao, PV Statistical research methods in the life sciences. 1998 Brooks/Cole Publishing Company. Pacific Grove, CA:
  29. SAS Institute Inc.. SAS/STAT 9.1 user’s guide. 2004 SAS Institute Inc. Cary, NC:
  30. Klein, JP, Moeschberger, ML Survival analysis: techniques for censored and truncated data. 1997 Springer-Verlag New York. New York, NY:
  31. Gasse, C, Hollowell, J, Meier, CR, et al Drug interactions and risk of acute bleeding leading to hospitalisation or death in patients with chronic atrial fibrillation treated with warfarin. Thromb Haemost 2005;94,537-543[ISI][Medline]
  32. Lip, GY, Varughese, GL Diabetes mellitus and atrial fibrillation: perspectives on epidemiological and pathophysiological links. Int J Cardiol 2005;105,319-321[CrossRef][ISI][Medline]
  33. Ostgren, CJ, Merlo, J, Rastam, L, et al Atrial fibrillation and its association with type 2 diabetes and hypertension in a Swedish community. Diabetes Obes Metab 2004;6,367-374[CrossRef][ISI][Medline]
  34. DiMarco, JP, Flaker, G, Waldo, AL, et al Factors affecting bleeding risk during anticoagulant therapy in patients with atrial fibrillation: observations from the Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM) study. Am Heart J 2005;149,650-656[CrossRef][ISI][Medline]
  35. McMahan, DA, Smith, DM, Carey, MA, et al Risk of major hemorrhage for outpatients treated with warfarin. J Gen Intern Med 1998;13,311-316[CrossRef][ISI][Medline]
  36. White, RH, Beyth, RJ, Zhou, H, et al Major bleeding after hospitalization for deep-venous thrombosis. Am J Med 1999;107,414-424[CrossRef][ISI][Medline]
  37. Steffensen, FH, Kristensen, K, Ejlersen, E, et al Major haemorrhagic complications during oral anticoagulant therapy in a Danish population-based cohort. J Intern Med 1997;242,497-503[ISI][Medline]
  38. Matchar, DB, Samsa, GP Secondary and tertiary prevention of stroke: patient outcomes research team; excerpts, seventh progress report. 1995 Center for Health Policy Research and Education, Duke University. Durham, NC:
  39. Albers, GW, Dalen, JE, Laupacis, A, et al Antithrombotic therapy in atrial fibrillation. Chest 2001;119(suppl),194S-206S[Free Full Text]
  40. Lip, GYH, Zarifis, J, Watson, RDS, et al Physician variation in the management of patients with atrial fibrillation. Heart 1996;75,200-205[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 ChestHome page
D. A. Lane and G. Y. H. Lip
Maintaining Therapeutic Anticoagulation: The Importance of Keeping "Within Range"
Chest, May 1, 2007; 131(5): 1277 - 1279.
[Full Text] [PDF]

Home page
 ChestHome page
P. Kakar, D. Lane, and G. Y. H. Lip
Bleeding risk stratification models in deciding on anticoagulation in patients with atrial fibrillation: a useful complement to stroke risk stratification schema.
Chest, November 1, 2006; 130(5): 1296 - 1299.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Shireman, T. I.
Right arrow Articles by Ellerbeck, E. F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Shireman, T. I.
Right arrow Articles by Ellerbeck, E. F.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS