http://orcid.org/0000-0001-5944-6814
Abstract
Introduction: Severe warfarin overanticoagulation is a risk factor for bleeding, but there is little information on its manifestations, prognosis and factors affecting the outcome. We describe the manifestations and clinical outcomes of severe warfarin overanticoagulation in a large group of patients with atrial fibrillation (AF).
Material and methods: All international normalized ratio (INR) samples (n = 961,431) in the Turku University Hospital region between 2003 and 2015 were screened. A total of 412 AF patients with INR ≥9 were compared to 405 patients with stable warfarin anticoagulation for AF. Electronic patient records were manually reviewed to collect comprehensive data.
Results: Of the 412 patients with INR ≥9, bleeding was the primary manifestation in 105 (25.5%). Non-bleeding symptoms were recorded in 165 (40.0%) patients and 142 (34.5%) had no symptoms. A total of 17 (16.2%) patients with a bleed and 67 (21.8%) without bleeding died within 30 days after the event. Intracranial haemorrhage strongly predicted death within 30 days. Other significant predictors were non-bleeding symptoms, active malignancies, recent bleed, history of myocardial infarction, older age, renal dysfunction and a recent treatment episode.
Conclusions: Bleeds are not the major determinant of the poor prognosis in severe overanticoagulation, as coincidental INR ≥9 findings also associate with high mortality.
Only a quarter of AF patients with INR ≥9 suffered a bleeding event and the clinical manifestation of INR ≥9 had a significant impact on patient outcome.
The 30-day mortality rate in patients with INR ≥9 was high ranging from 9.2 to 32.7%.
Several significant predictors of 30-day mortality after INR ≥9 were identified.
KEY MESSAGES
Introduction
Anticoagulants are the most common reason for emergency department visits due to adverse drug effects [Citation1]. International normalized ratio (INR) fluctuations are common during long-term warfarin anticoagulation and high INR values associate with major bleeding complications as well as increased mortality [Citation2–6]. However, causes of death in large atrial fibrillation (AF) trials are mostly related to factors other than bleeding [Citation7]. In our previous report, we identified several risk factors for very high INR values (≥9) in patients with AF [Citation8]. Nevertheless, little is known about the clinical manifestations of excessive warfarin anticoagulation (EWA) or the outcome of these patients, especially in patients without bleeding complications. The aim of our study was to investigate the clinical manifestations and outcome of severe warfarin overanticoagulation (INR ≥9), to identify risk factors for short-term mortality and to study the relationship between clinical manifestation and patient outcomes. Studies focusing on severe overanticoagulation are scarce with limited number of patients and thus the present EWA study provides novel and clinically relevant information on this phenomenon [Citation5,Citation6].
Material and methods
This study is a pre-specified analysis of the Excessive Warfarin Anticoagulation Study (the EWA Study, ClinicalTrials.gov identifier: NCT02761941) which belongs to a series of study protocols assessing anticoagulation-related complications in the treatment of AF [Citation8–10]. In this study, we investigated a patient group of 412 patients with severe overanticoagulation, defined as INR ≥9, during warfarin treatment in patients with AF (the EWA Group). Patients were identified with computer searches from Turku University Hospital and Turku City Hospital laboratory database of 961,431 INR samples between 2003 and 2015. From this database, we also identified a Control group including all patients (n = 405) with AF on long-term (at least 730 days), regularly controlled (maximum INR test interval 60 days) warfarin anticoagulation without INR elevations >4. The event date used for data collection in the Control group was the date with the highest INR value (between 2.7 and 4.0) in each patient. Detailed patient selection criteria and the patient characteristics are described in our previous report on the EWA study [Citation8]. After initial screening process, all individual electronic patient records were reviewed using a standardized protocol to collect specific information on clinical manifestations and patient management during the high INR event. The short- and long-term outcomes were also recorded, as well as a 90-day follow-up on thromboembolic and bleeding complications. Mortality data was collected from the Official Statistics of Finland managed by the National Statistical Service in Finland.
Definitions
Recent antibiotic or antifungal therapy included antibiotic therapy during the preceding 14 days of the index event date or antifungal therapy in the preceding 7 days. Alcohol abuse was defined as an alcohol-related diagnosis or a hospital/health center visit due to excessive alcohol use. Recent medical treatment was defined as an outpatient clinic visit with or without hospitalization during the preceding 30 days. Patients living independently at home without requiring outside help in everyday life were classified as independent. Estimated glomerular filtration rate (eGFR) was calculated using CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) equation to classify renal function according to KDIGO (Kidney Disease Improving Global Outcomes) clinical practice guidelines [Citation11]. The definition of intracranial bleed included intracerebral, subarachnoid or subdural bleeding event. Recent bleeding event included all bleeds in the preceding 30 days that were reported by the patient or led to either hospitalization or outpatient clinic visit. All bleeding events during severe overanticoagulation were classified using the International Society on Thrombosis and Haemostasis/Scientific and Standardization Committee (ISTH/SSC) bleeding score [Citation12]. Patients suffering from other than bleeding-related symptoms (e.g. shortness of breath, fever or nausea) were assigned to the Other symptoms-group.
Statistical analysis
Continuous variables were reported as mean ± standard deviation if they were normally distributed, and as median [inter-quartile range (IQR)] if they were skewed unless stated otherwise. Categorical variables were described with absolute and relative (percentage) frequencies. Chi-squared test and Fisher’s exact test were used for categorical variables as appropriate. For all variables with more than 0.5% missing data, the exact number of patients with missing data is marked in the table. A binary stepwise logistic regression analysis (backward Wald) was performed to identify independent predictors of 30-day mortality and bleeding complications in patients with very high INR values. Variables strongly correlated with the dependent variable by univariate analyses (p < .1) were entered in the model as covariates.
We included prior myocardial infarction and age over 75 years in the multivariate model and excluded coronary artery disease and not-independent-living due to significant intercorrelations. All tests were two-sided and statistical significance was set at 5%. Kaplan–Meier survival curves were constructed to display the time-to-event relationship for the occurrence of death. This article was written following STROBE guidelines for the reporting of observational studies [Citation13]. Statistical analyses were performed with SPSS software (version 23.0, SPSS Inc., Chicago, IL) and SAS software (version 9.2, SAS Institute, Inc., Cary, NC).
The study protocol was approved by the Medical Ethics Committee of the Hospital District of Southwest Finland. Informed consent was not required, because of the register-based nature of the study. The study complies with the Declaration of Helsinki as revised in 2002.
Results
Severe EWA was observed in 412 AF patients (the EWA Group) comprising 3.0% of all 13,618 patients living in the Turku University Hospital region with at least one INR value of ≥2 as an indirect marker of warfarin anticoagulation. The previous INR value before the event was 2.72 ± 0.37 in the Control Group and 3.04 ± 1.44 in the EWA Group (p < .001).
Manifestations
Symptoms other than bleeding were recorded in 165 (40.0%) patients, whereas severe overanticoagulation was a coincidental finding without any symptoms in 142 (34.5%) patients. Bleeding was the primary clinical manifestation of EWA in 105 (25.5%) patients. Only four patients (1.0%) in the Control group suffered a bleeding event on the index event date. Patient characteristics according to EWA manifestation type are presented in .
Table 1. Patient characteristics.
A total of 112 bleeding events were observed in 105 patients during severe overanticoagulation. Major bleeds (ISTH 3–4) comprised 49.1% of all bleeds. A third of all bleeds (33.0%) were from the gastrointestinal tract, half of which (n = 19) were severe (ISTH 4). Cutaneous bleeds, urinary tract bleeds or nosebleeds comprised 41.1% of all bleeds, the majority of them being classified as minor (ISTH 1–2). Only five patients suffered an intracranial bleeding event. Bleeding types, frequencies and severity classifications are listed in . Independent predictors of bleeding complications associated with EWA in a multivariate analysis were use of tramadol [OR: 3.63; 95% confidence interval (CI) = 1.28–10.3, p = .015] and a bleeding event within 30 days before EWA episode (OR: 5.87; 95% CI =1.93–17.9, p = .002), while a history of malignancy was of borderline significance (OR: 1.67; 95% CI = 0.99–2.81, p = .053).
Table 2. Observed number and classification of bleeding complications during severe overanticoagulation.
In EWA episodes without symptoms, previous INR was >3 in 29.6% of the patients [n = 42, mean INR 2.91 (95% CI =2.69–3.14)] and the median interval between the previous sample and EWA was 15 (IQR 19) days. Previous INR was >3 in 24.8% [n = 26, mean INR 2.67 (95% CI =2.46–2.89)] of the patients with bleeding symptoms, and in 38.8% [n = 64, mean INR 3.38 (95% CI =3.13–3.62)] of patients with non-bleeding symptoms, while the INR sample intervals were 16 (IQR 26) and 9 (IQR 15) days respectively.
Treatment
Warfarin was temporarily or permanently discontinued in all patients with severe overanticoagulation. Vitamin K (oral, intramuscular or intravenous) was administered in 353 (85.7%) patients with an EWA episode, without statistically significant difference between Bleeding- (88.6%, n = 93), Other symptoms- (85.5%, n = 141) and No-symptoms groups (83.8, n = 119) (p = .568). Other medications [prothrombin complex concentrate, solvent detergent (S/D) treated human plasma or tranexamic acid] were used to counteract overanticoagulation in 41.9% (n = 44) of patients with a bleeding event, 5.9% (n = 4) without symptoms and 29.4% (n = 20) in patients with symptoms other than bleeding. Five (4.8%) patients with bleeding complications did not receive any drugs to reverse warfarin effect.
Outcomes
The 30-day mortality rate was 20.4% (n = 84) in the EWA Group and 0.2% (n = 1) in the Control Group. A total of 17 (16.2%) patients with bleeding as the primary symptom and 67 (21.8%) without a bleed died within 30 days of severe overanticoagulation event. Bleeding as the clinical manifestation of the EWA episode did not have a statistically significant effect on short-term outcome in the multivariate analysis. Nearly a third (n = 54, 32.7%) of patients presenting with symptoms other than bleeding died within 30 days after EWA. Coincidental symptomless finding of INR ≥9 was associated with the best 30-day outcome, as there were 13 deaths (9.2%) within 1 month of EWA. At 12 months, the mortality rates were 40.5% (n = 167) in the EWA Group and 1.7% (n = 7) in the Control group. The survival advantage of the Control group over EWA group persisted through long-term follow up and at 4 years, 20 patients (4.9%) had died in the Control group and 257 (62.4%) in the EWA group. The cumulative survival analysis for both groups, as well as for each manifestation type of EWA is presented in .
Figure 1. Kaplan–Meier survival analysis according to manifestation of EWA. Other symptoms, patients suffering other than bleeding-related symptoms.
Independent predictors of 30-day mortality in patients with EWA are presented in and the results of the univariate analysis in Supplementary Table S1. Intracranial haemorrhage was the only bleed to independently predict short-term mortality, as no other bleeding type [even the most severe (ISTH class 4)] predicted death in 30 days after EWA. Only 8 (1.9%) patients suffered an ischaemic stroke and one patient (0.2%) had a peripheral embolism in the 90 days following the EWA episode. There was no association between the use of vitamin K or other drugs used for warfarin reversal, and thromboembolic complications, short-term mortality or duration of hospital stay.
Table 3. Multivariate predictors of 30-day mortality in patients with excessive warfarin anticoagulation.
Discussion
To our knowledge, the current observational study is the largest analysis on severe warfarin overanticoagulation. Our observational study shows that severe overanticoagulation during warfarin treatment associates with high mortality in patients with AF, even when the high INR value is detected in routine controls without any symptoms. We identified several clinical risk factors for 30-day mortality as well as for bleeding complications during EWA. Our results suggest that even though very high INR predisposes to severe bleeds, the majority of deaths are explained by other factors than bleeding complications—with the obvious exception of intracranial bleeds [Citation5]. Long-term follow-up revealed a persisting difference in mortality rate between the groups.
The observed high mortality of patients experiencing non-bleeding-related symptoms during the EWA event may be explained with the poor prognosis of the underlying critical illnesses which may cause INR elevations even without warfarin [Citation14]. Therefore, it was not a surprise that these patients had a four-fold increase in the risk of death within 30 days and the mortality rate was very high (35.1%). Even though the poor outcomes of these critically ill patients explain the high overall mortality associated with EWA, it was unexpected that bleedings—except for intracranial haemorrhages—did not predict short-term mortality. Our findings in these patients with extreme anticoagulation intensity are in line with a recent meta-analysis by Gómez-Outes et al. [Citation7] reporting that only 6% of deaths in modern anticoagulation studies in AF patients are caused by bleedings. Also in accordance with previous studies, intracranial bleeding event during EWA was the strongest predictor of 30-day mortality in our study [Citation4,Citation15].
Among all clinical manifestation types of EWA, subjects with coincidental asymptomatic finding of INR value ≥9 had the best prognosis. This is expected, since these EWA episodes were detected in routine controls before symptomatic complications had occurred and the overanticoagulation was less frequently related to acute illness. Nevertheless, to put the better prognosis of this patient group in perspective, the 30-day mortality rate (9.2%) after symptomless INR ≥9 episode was much higher than for example after acute ST-elevation myocardial infarction (.9–3.1%) in contemporary clinical trials [Citation16,Citation17]. The poor prognosis of all EWA manifestation types highlights the importance of recognizing the previously reported clinical risk factors for severe overanticoagulation [Citation8].
As we have reported previously, hospital treatment episodes or emergency department/health care center visits independently predict severe overanticoagulation in the following 30 days [Citation8]. Our current results show that these treatments episodes are also independent predictors of 30-day mortality in this patient group. The medical condition itself in combination with possible operations and medications (such as antibiotics or antifungals) used for managing the condition all probably contribute to the worse outcome of these patients. This finding highlights the importance of frequent and systematic INR controls during and following these treatment episodes.
The predictors of overanticoagulation describe the characteristic of patients in each subgroup () and reflect the underlying mechanism of overanticoagulation [Citation8]. Comorbidities such as renal failure, heart failure or treatment of an active disease process prior to the overanticoagulation (antibiotics or antifungals, recent medical treatment) were logically most frequent in the “Other symptoms” group. Alcohol consumption and active malignancies were most common in patients with bleeding symptoms and mechanical heart valves and active smoking in asymptomatic patients.
Interestingly, a bleeding event during EWA did not have an effect on 30-day mortality, but a prior bleeding event (in the preceding 30 days) increased the risk 3.4-fold. This increased risk of patients with recent bleeding events is probably related to underlying comorbidities and acute diseases, which may predispose to bleeding events even during lower anticoagulation intensity. Furthermore, the consequential treatment interventions may also have an effect on prognosis when INR level later reaches ≥9. This finding emphasizes the importance of anticoagulation stability in patients who are prone to bleedings during warfarin treatment.
As expected, active malignancies independently predicted short-term mortality after EWA, since cancer associates with poor prognosis also in patients without warfarin. As active malignancies have also been shown to predict severe overanticoagulation, the risks and benefits have to be weighted carefully when considering warfarin for these patients [Citation8]. Moreover, patients with a history of myocardial infarction and AF are known to be at increased risk of cardiovascular death [Citation7,Citation18]. When these patients with cardiovascular diseases experienced an EWA episode, the risk of 30-day mortality was high (2.9-fold) compared to patients without a history of infarction.
Older age (≥75 years) was expectedly an independent predictor of short-term mortality. Different acute conditions in elderly patients—such as dehydration or infections—may cause only subtle and non-specific symptoms that may delay diagnosis and treatment interventions [Citation19]. Besides being a well-established risk factor for bleeding and all-cause mortality, renal impairment is also a risk factor for severe overanticoagulation in AF patients [Citation8,Citation20,Citation21]. Considering this background, it was a logical finding that there was a two-fold increase in the risk of death after EWA in patients with severe renal impairment (eGFR <30).
The causes of death are often multifactorial even in patients with a bleeding event and are subject to interpretation for several reasons. Determining the causalities and associations of different comorbidities in the process is a subjective interpretation often even after autopsy. For instance, active bleeding is a prothrombotic condition, which can modify the clinical presentation of the disease process in the final stages of life, thus obscuring the significance of overanticoagulation and bleeding.
Recent bleeding event is a well-established risk factor for subsequent bleeding in anticoagulated patients reflecting their general susceptibility to bleeding [Citation22,Citation23]. As the intensity of anticoagulation increases, the threshold for haemorrhages is lowered and bleedings should manifest more easily in patients with a history of bleeding [Citation24]. In line with this reasoning, recent bleeding event was the strongest (5.9) risk factor of bleeding complication during EWA episode. Even though the rate of major bleeding events in our study was higher than in contemporary AF trials due to the extreme anticoagulation intensity, the majority of patients did not experience bleeding during the EWA episode [Citation4,Citation7,Citation25–28]. Of note, INR elevations seem to develop rapidly prior to the bleeding event and thus it may be challenging to prevent bleeding complications by more frequent INR controls to detect overanticoagulation before bleedings become evident [Citation29]. Furthermore, majority of bleeding events in anticoagulated patients occur during therapeutic anticoagulation intensity from occult bleeding sites revealed already by low intensity anticoagulation [Citation4,Citation30]. This may also explain the finding that recent initiation of warfarin treatment was not an independent risk factor for bleeding in our study, as most bleedings in this scenario probably occur before INR reaches the level ≥9 [Citation31,Citation32]. Moreover, extreme warfarin anticoagulation intensity did not have an effect on the bleeding sites since bleeding types were similar to those in several earlier anticoagulation studies [Citation3,Citation30,Citation31,Citation33–36].
It would be straightforward to think that the standard treatment of administering vitamin K would benefit these patients. In accordance with previous reports, the administration of vitamin K or other counteracting drugs had no effect on occurrence of complications (death, thromboembolism or length of hospital stay) [Citation5,Citation37]. It should be noted, however, that most of the patients received vitamin K and it was impossible to analyze reasons for not using vitamin K in the retrospective study setting.
Strengths and limitations
The strengths of this study include exceptionally good coverage of all INR samples collected in the southwestern Finland (475,580 inhabitants). Turku University Hospital laboratory service provider (TYKSLAB) performs all the INR measurements regardless of the place of residence in the district. All the major complications related to EWA episodes are treated in the same hospitals where our study data was also gathered. Statistics of Finland is the official governmental agency, whose purpose is to store reliable and up-to-date data on the death rate and causes of death in Finland. In addition, the vast majority of people living in the southwestern Finland tend not to move to elsewhere.
This study has all the inherent limitations of the retrospective study. As the data was collected by study personnel retrospectively, we had to rely on the data and diagnoses recorded by physicians treating the patients. Treatment decisions were at the treating physicians’ discretion and factors not covered by the case report form may affect the results. Warfarin dosage and level of adherence to medication were not available. Despite these limitations, findings from our “real world” study offer solid observations on how these frail patients were treated and what affected their outcome in short and long term.
Conclusions
Severe overanticoagulation is a dangerous complication of warfarin treatment for AF and it appears that bleeding is not the major determinant for the poor prognosis. Some risk factors for excessive anticoagulation are modifiable, emphasizing the importance of identifying high-risk patients and considering alternative treatment strategies in an attempt to improve treatment safety.
Samuli_et_al._Supplementary_Material.docx
Download MS Word (14.9 KB)Acknowledgements
We thank our study coordinator Tuija Vasankari, RN, for her crucial input in study management and Ari Törmä, M.Sc., for providing laboratory data for the study, and Henri Sallinen, BM, Marianne Mäkäräinen, BM, and Melina Issakoff, BM, for the help in the collection of the data. We also thank Ville Langen, MD, for helping with the figure.
Disclosure statement
S.J.: received research grants from the Finnish Foundation for Cardiovascular Research; Clinical Research Fund (EVO) of Turku University Hospital, Turku, Finland, Finnish Cardiac Society. I.N.: None. T.O.K.: Lectures for Bayer, Boehringer Ingelheim, Bristol-Myers-Squibb-Pfizer, Medicines Company, MSD, Astra Zeneca and St Jude Medical, received research grants from the Finnish Medical Foundation, the Finnish Foundation for Cardiovascular Research; Clinical Research Fund (EVO) of Turku University Hospital, Turku, Finland, Finnish Cardiac Society, and an unrestricted grant from Bristol-Myers-Squibb-Pfizer. Member of advisory board for Boehringer-Ingelheim; MSD. R.V.: None. A.V.: None. K.E.J.A.: research grants from the Finnish Foundation for Cardiovascular Research, Clinical Research Fund (EVO) of Turku University Hospital, Turku, Finland; Lectures for Bayer, Cardiome and Boehringer Ingelheim, Member in the advisory boards for Bayer, Astra Zeneca, Bristol-Myers-Squibb-Pfizer.
Additional information
Funding
References
- Shehab N, Lovegrove MC, Geller AI, et al. US emergency department visits for outpatient adverse drug events, 2013–2014. JAMA. 2016;316:2115–2125.
- Hylek EM, Go AS, Chang Y, et al. Effect of intensity of oral anticoagulation on stroke severity and mortality in atrial fibrillation. N Engl J Med. 2003;349:1019–1026.
- Hylek EM, Evans-Molina C, Shea C, et al. Major hemorrhage and tolerability of warfarin in the first year of therapy among elderly patients with atrial fibrillation. Circulation. 2007;115:2689–2696.
- Palareti G, Leali N, Coccheri S, et al. Bleeding complications of oral anticoagulant treatment: an inception-cohort, prospective collaborative study (ISCOAT). Italian Study on Complications of Oral Anticoagulant Therapy. Lancet. 1996;348:423–428.
- Koo S, Kucher N, Nguyen PL, et al. The effect of excessive anticoagulation on mortality and morbidity in hospitalized patients with anticoagulant-related major hemorrhage. Arch Intern Med. 2004;164:1557–1560.
- Hylek EM, Chang YC, Skates SJ, et al. Prospective study of the outcomes of ambulatory patients with excessive warfarin anticoagulation. Arch Intern Med. 2000;160:1612–1617.
- Gomez-Outes A, Lagunar-Ruiz J, Terleira-Fernandez AI, et al. Causes of death in anticoagulated patients with atrial fibrillation. J Am Coll Cardiol. 2016;68:2508–2521.
- Jaakkola S, Nuotio I, Kiviniemi TO, et al. Incidence and predictors of excessive warfarin anticoagulation in patients with atrial fibrillation-the EWA study. PLoS One. 2017;12:e0175975.
- Airaksinen KE, Gronberg T, Nuotio I, et al. Thromboembolic complications after cardioversion of acute atrial fibrillation: the FinCV (Finnish CardioVersion) study. J Am Coll Cardiol. 2013;62:1187–1192.
- Palomaki A, Mustonen P, Hartikainen JE, et al. Strokes after cardioversion of atrial fibrillation-The FibStroke study. Int J Cardiol. 2016;203:269–273.
- Chapter 1: definition and classification of CKD. Kidney Int Suppl. 2011;3:19–62.
- Rodeghiero F, Tosetto A, Abshire T, et al. ISTH/SSC bleeding assessment tool: a standardized questionnaire and a proposal for a new bleeding score for inherited bleeding disorders. J Thromb Haemost. 2010;8:2063–2065.
- von Elm E, Altman DG, Egger M, et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. BMJ. 2007;335:806–808.
- Oden A, Fahlen M. Oral anticoagulation and risk of death: a medical record linkage study. BMJ. 2002;325:1073–1075.
- Hart RG, Boop BS, Anderson DC. Oral anticoagulants and intracranial hemorrhage. Facts and hypotheses. Stroke. 1995;26:1471–1477.
- Stone GW, Witzenbichler B, Guagliumi G, et al. Bivalirudin during primary PCI in acute myocardial infarction. N Engl J Med. 2008;358:2218–2230.
- Stone GW, White HD, Ohman EM, et al. Bivalirudin in patients with acute coronary syndromes undergoing percutaneous coronary intervention: a subgroup analysis from the Acute Catheterization and Urgent Intervention Triage strategy (ACUITY) trial. Lancet. 2007;369:907–919.
- Norgaard ML, Andersen SS, Schramm TK, et al. Changes in short- and long-term cardiovascular risk of incident diabetes and incident myocardial infarction–a nationwide study. Diabetologia. 2010;53:1612–1619.
- High KP, Bradley SF, Gravenstein S, et al. Clinical practice guideline for the evaluation of fever and infection in older adult residents of long-term care facilities: 2008 update by the Infectious Diseases Society of America. Clin Infect Dis. 2009;48:149–171.
- Boriani G, Laroche C, Diemberger I, et al. Glomerular filtration rate in patients with atrial fibrillation and 1-year outcomes. Sci Rep. 2016;6:30271.
- Pavord S, Myers B. Bleeding and thrombotic complications of kidney disease. Blood Rev. 2011;25:271–278.
- Fang MC, Go AS, Chang Y, et al. A new risk scheme to predict warfarin-associated hemorrhage: the ATRIA (Anticoagulation and Risk Factors in Atrial Fibrillation) study. J Am Coll Cardiol. 2011;58:395–401.
- Pisters R, Lane DA, Nieuwlaat R, et al. A novel user-friendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation: the Euro Heart Survey. Chest. 2010;138:1093–1100.
- Grzymala-Lubanski B, Svensson PJ, Renlund H, et al. Warfarin treatment quality and prognosis in patients with mechanical heart valve prosthesis. Heart. 2017;103:198–203.
- Ezekowitz MD, Bridgers SL, James KE, et al. Warfarin in the prevention of stroke associated with nonrheumatic atrial fibrillation. Veterans Affairs Stroke Prevention in Nonrheumatic Atrial Fibrillation Investigators. N Engl J Med. 1992;327:1406–1412.
- Connolly SJ, Laupacis A, Gent M, et al. Canadian Atrial Fibrillation Anticoagulation (CAFA) study. J Am Coll Cardiol. 1991;18:349–355.
- 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.
- Schulman S, Beyth RJ, Kearon C, et al. Hemorrhagic complications of anticoagulant and thrombolytic treatment: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition). Chest. 2008;133(Suppl. 6):257s–298s.
- Kucher N, Connolly S, Beckman JA, et al. International normalized ratio increase before warfarin-associated hemorrhage: brief and subtle. Arch Intern Med. 2004;164:2176–2179.
- 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.
- Landefeld CS, Beyth RJ. Anticoagulant-related bleeding: clinical epidemiology, prediction, and prevention. Am J Med. 1993;95:315–328.
- Copland M, Walker ID, Tait RC. Oral anticoagulation and hemorrhagic complications in an elderly population with atrial fibrillation. Arch Intern Med. 2001;161:2125–2128.
- Granger CB, Alexander JH, McMurray JJ, et al. Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med. 2011;365:981–992.
- Giugliano RP, Ruff CT, Braunwald E, et al. Edoxaban versus warfarin in patients with atrial fibrillation. N Engl J Med. 2013;369:2093–2104.
- Connolly SJ, Ezekowitz MD, Yusuf S, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med. 2009;361:1139–1151.
- Torn M, Bollen WL, van der Meer FJ, et al. Risks of oral anticoagulant therapy with increasing age. Arch Intern Med. 2005;165:1527–1532.
- Crowther MA, Ageno W, Garcia D, et al. Oral vitamin K versus placebo to correct excessive anticoagulation in patients receiving warfarin: a randomized trial. Ann Intern Med. 2009;150:293–300.