Doi:10.1016/j.jamda.2005.08.002

To Anticoagulate or not to
Anticoagulate? A Common Dilemma for
the Provider: Physicians’ Opinion Poll
Based on a Case Study of an Older
Long-term Care Facility Resident With
Dementia and Atrial Fibrillation
T.S. Dharmarajan, MD, FACP, AGSF, Surendran Varma, MD, Shailaja Akkaladevi, MD, Anna S. Lebelt, MD, andEdward P. Norkus, PhD, FACN Objective: Anticoagulation therapy is an acceptable The most cited reasons for not providing anticoagula- strategy for the prevention of thromboembolic events tion were risk of falls (98%), dementia (40%), and in the presence of atrial fibrillation. However, this short life expectancy (32%). However, 92% of respon- strategy is controversial in older subjects particularly in dents felt that the patient was a candidate for short- the presence of dementia. We conducted an opinion term anticoagulation therapy. Interestingly, the poll regarding the decision to anticoagulate or not choices (yes, no, uncertain) to the questions were sim- among physicians in practice and in various levels of ilar for all physicians irrespective of their level of train- training (residents and fellows) that was based on a spe- ing or years in practice (or faculty) after training.
cific, yet not unusual, case scenario in the nursing home.
Conclusions: Although long-term anticoagulation for Setting: A university teaching hospital in the Bronx, NY.
thromboembolic events in atrial fibrillation is consid-ered beneficial, recent reports suggest that warfarin is Methods: A survey questionnaire was distributed to underused in older adults, especially in the long-term physicians to solicit opinions on the decision to antico- care setting. Our physician poll, based on a specific agulate based on an actual case from a LTCF and the case scenario, is consistent with this opinion as re- flected by both trainees and practicing physicians.
While there are absolute and relative contraindica- Results: One hundred seven completed surveys were tions to the use of long-term warfarin, decisions returned from 49 residents, 20 fellows, and 38 attend- should be individualized and based on risks, benefits, ing physicians. The majority (85%) felt that long-term and quality of life of the resident. (J Am Med Dir Assoc
anticoagulation therapy was not indicated in the case 2006; 7: 23–28)
patient. However, most (88%) felt they would providean antiplatelet agent, with the choice being 78% as- Keywords: Anticoagulation in long-term care; atrial pirin, 20% clopridogel, and 2% aspirin-dipyridamole.
fibrillation; physician opinion poll Anticoagulation therapy for atrial fibrillation has become oral warfarin is the traditional drug of choice. However, in an accepted practice in long-term management to prevent view of the risks associated with anticoagulation therapy, thromboembolic events in older For this purpose, many physicians have become cautious in their use of anti-coagulants, particularly in older patients. The question to Our Lady of Mercy Medical Center, Bronx, NY, and University Hospital of New anticoagulate or not is perhaps even more relevant in resi- York Medical College, Valhalla, NY (T.S.D., S.V., S.A., A.S.L., E.P.N.).
dents of long-term care facilities (LTCFs). This cautionary This study was internally funded as part of the Geriatric Medicine Fellowship approach prompted us to conduct a survey among physicians at various levels of training and experience (residents, fellows, Address correspondence to T.S. Dharmarajan, MD, 31 Pheasant Run, Scars- and attending physicians) at an inner city, university medical dale, NY 10583. E-mail: [email protected] center. Our survey questionnaire was based on a specific case Copyright 2006 American Medical Directors Association
history of a LTCF resident with dementia and atrial fibrilla- DOI: 10.1016/j.jamda.2005.08.002
would you choose (aspirin, clopidogrel, aspirin-dipyri-damole)? Survey Case
5. If you would not recommend long-term anticoagulation An 87-year-old white female with known Alzheimer’s dis- therapy in this patient, indicate the reason(s) for your ease and AF, plus a past history of surgery for left hip fracture, choice. You may list as many reasons as you wish.
was admitted to a LTCF following discharge from the hospi- (Examples: Dementia, risk of falls, short life expectancy, tal. Her echocardiogram revealed normal left ventricle (LV) drug interactions, need for frequent laboratory testing, size and function, with abnormal diastolic filling considered risk of legal action, plus any other reasons.) consistent with her age. Before her last hospitalization, she 6. If you feel that this patient is not a candidate for long- resided in the dementia unit of an assisted living facility.
term anticoagulation therapy, would you consider short- While in that facility, she used to ambulate using a walker term anticoagulant therapy (eg, after hip fracture) (yes, with assistance. Because of her cognitive impairment includ- ing severe short-term and long-term memory deficits, andimpaired judgment, she currently required assistance in most Data Analysis
areas of her activities of daily living (ADL). At times, she alsoexhibited behavioral problems and would become physically A single individual entered the information from all sur- veys into a PC-based spreadsheet format. The spreadsheet In the LTCF, she initially received nasogastric tube feed- then was transformed into a statistical dataset and analyzed ings because of poor food intake. Her diet eventually was using PC-based Statistical Analysis Software (STATA 8.0, changed to pureed food plus thickened liquids. During several months of this regimen, her oral intake improved. Despite Statistical analysis provided descriptive summaries that risperidone treatment, she continued to manifest occasional were expressed as actual numbers, percentages, or mean values episodes of abnormal behavior. On admission to the LTCF, (Ϯ standard deviation). Student t tests were used to detect she was placed on warfarin therapy for thromboembolic pro- significant differences between mean values of 2 independent, phylaxis and on donepezil. She was wheelchair bound and continuous variables while chi-square analyses or Fisher exact required assistance in all areas of ADL. She also failed to tests (test cell size dependent) were used to compare categor- cooperate for a mini mental state examination (MMSE).
ical variables. P values less than .05, in 2-tailed testing, were One night, 6 months after the initiation of warafin therapy, considered statistically significant.
she was found on the floor in her room by the nursing staff.
She had a swelling over her right forehead; her mental status appeared baseline. Her right forehead revealed a 4 ϫ 5 cm softtissue swelling, with surrounding areas of ecchymosis. Her Participants
blood pressure was normal, with no evidence of orthostasis.
One hundred seven surveys were completed and returned.
Heart rate was normal but irregular, consistent with AF. No Of these, 49 surveys were from residents in internal medicine, murmur or bruit was evident and focal neurological deficits 20 were from fellows in specialties of medicine (including were not apparent. Radiological examination was negative for geriatric medicine), and 38 were from attending physicians in fracture. Her international normalized ratio (INR), obtained the day prior to the incident, was 1.4 (normal, but neverthe-less could represent altered prothrombin activity). After the Answers to Question 1
fall, her educated son who also was her surrogate engaged thephysician in a long discussion regarding the pros, cons, and Thirteen percent of the respondents felt that the case study potential alternatives to anticoagulant therapy. After a pro- was a candidate for long-term anticoagulation therapy, 85% longed discussion, warfarin therapy was continued.
felt the patient was not a candidate for long-term anticoagu-lation, and 2% were uncertain. We determined no significant Survey Questions
differences in the pattern of response among residents, fellows,and attending physicians P ϭ .956), or when the The survey based on the described case was distributed to responses of trainees (residents and fellows combined) were physicians; the 6 questions and the possible choices for an- compared to those of the attending physicians (P ϭ .795).
1. Is this patient a candidate for long-term anticoagulation Answers to Question 2
Of the respondents who elected to provide long-term an- 2. If this patient is a candidate for long-term anticoagula- ticoagulation therapy, 8% stated that they would also provide tion therapy, would you also consider providing an an- antiplatelet therapy, 84% stated they would not, and 8% were uncertain. We observed significant differences in the pattern 3. If this patient is not a candidate for long-term anticoag- of response among residents, fellows, and attending physicians ulation therapy, would you nevertheless consider pro- (P ϭ .025). Residents were uniformly opposed to antiplatelet viding an antiplatelet agent (yes, no, uncertain)? therapy while fellows and attending physicians were more 4. If you were to provide an antiplatelet agent, what agent DISCUSSION
As the woman’s case scenario developed, concerns were resident
raised as to the whether or not we should initiate warfarin therapy for long-term thromboembolic prophylaxis for AF.
attending
Our discussions with her son (her surrogate) raised several additional concerns. Because of this, we decided to conduct an opinion poll, based on this woman’s specific case history, to determine the general consensus of physicians at our medical center regarding anticoagulation individualized to this resi- dent. One hundred seven physicians, in different stages of not treat
uncertain
training (residents and fellows) and experience (faculty phy-sicians, physicians in medicine and related specialties with Long-term anticoagulation therapy in AF. Responses to the practices in long-term care and the community) returned question 1: “Is this patient a candidate for long-term anticoagula- tion therapy in atrial fibrillation?” The results of our poll findings are interesting. By an almost 9 to 1 ratio, physicians elected not to provide long-termanticoagulation therapy for this woman In addi-tion, the pattern of physicians’ response (yes, no, uncertain) Answers to Question 3
to provide or not to provide long-term anticoagulation ther- Of the respondents who elected not to provide long-term apy was similar irrespective of the level of training or experi- anticoagulation therapy, 88% stated that they would provide ence (resident vs fellow vs.attending physician).
antiplatelet therapy, 9% responded that they would not, and Each physician provided a number of different reasons why 3% were uncertain. We observed similar patterns of response they would not provide warfarin therapy Risk of (Ϸ88% yes, Ϸ9% no, and Ϸ3% uncertain) among residents, falls (98% of respondents), dementia (40%), and short life fellows, and attending physicians (P ϭ .441).
expectancy were the most commonly cited reasons for notproviding long-term anticoagulation. Less commonly citedreasons were drug interactions involving warfarin (21%), the Answers to Question 4
need for frequent lab tests (14%), and fear of litigation (4%).
Among all respondents who stated that they would provide The vast majority of physicians who chose not to provide an antiplatelet agent, 78% chose aspirin, 20% chose clopri- anticoagulation were willing to provide an antiplatelet agent.
dogel, and 2% chose a combination of aspirin-dipyridamole.
Once again, the pattern of choice (yes, no, uncertain) for the We observed no significant differences in the choice of anti- use of antiplatelet agent was similar for physicians at all levels platelet agent among residents, fellows, and attending physi- of training and experience; approximately 88% would pro- cians (P ϭ .514). We also observed no significant difference vide, 9% would not, and 3% were uncertain. If an antiplatelet in the choice of antiplatelet agent whether or not the choice agent was provided, 78% of respondents chose aspirin, 20% was to provide long-term anticoagulation therapy to the pa- clopidogrel, and 2% a combination of aspirin-dipyridamole.
Again, the pattern of choice (to provide, not to provide, anduncertain) remained similar regardless of the level of training Answers to Question 5
Among the reasons why long-term anticoagulation therapy Interestingly, the overwhelming majority of physicians should not be provided, the risk of falls was cited by 98% of (92%) stated that they would consider short-term anticoagu- respondents, 40% cited dementia, 32% cited short life expect-ancy, 21% cited the possibility of drug interaction, 14% citedthe need for frequent lab testing, and 4% cited the potentialfor litigation We observed similar patterns for the reasons not to provide long-term anticoagulation therapy among residents, fellows, and attending physicians (P ϭ frequent labs
drug interaction
Answers to Question 6
short-life expectancy
dementia
Ninety-two percent of the respondents believed that the case study was a candidate for short-term anticoagulant ther- risk of falls
apy while 6% of respondents felt the patient was not a candidate for short-term anticoagulant therapy and 2% were Percentage of Responses to Question 5
uncertain. We observed no significant differences in the re-sponse patterns among residents, fellows, and attending phy- Reasons given for not recommending long-term anticoag- ulation therapy to patients with AF. lation therapy for this woman, if warranted, such as for the Warfarin: Benefits and Risks in Older Adults and
prevention of deep vein thrombosis following hip fracture.
Application to Our Specific Case
The fact that the majority of physicians (85%) also believed Warfarin has proven therapeutic benefit as an anticoagu- that long-term anticoagulation therapy was inappropriate in lant in Its role in the prevention and treatment of this particular case is consistent with several reports suggest- venous thromboembolism, nonvalvular and valvular AF, ing a tendency to underuse It is also interesting acute myocardial infarct (MI), valvular heart disease with that trainees and practicing physicians chose the same ap- history of systemic embolism, and prosthetic heart valves are proach of deciding against anticoagulation therapy even well recognized.It is unfortunate that older adults though both groups had vastly different levels of training and who may derive maximum benefits from anticoagulant ther- experience. A likely reason for this may be that, within any apy also tend to have many risk factors for complications given setting, trainees are influenced by exposure to their The prevalence of nonvalvular AF increases with Six After a prolonged discussion with the patient’s son, warfa- percent of men and 5% of women older than 65 years mani- rin therapy was continued for 2 additional months. During fested whereas 10% of people older than 80 years have these 2 months, she fell again and warfarin was discontinued.
The prevalence of AF in LTCF may be as high as The woman was well for 1 year after discontinuation of Both reports discuss warfarin underuse in LTCF, even in the warfarin therapy. During this year, additional falls did not absence of contraindications, and suggest the need for better occur nor were there any thromboembolic complications. She monitoring practices. The authors call for a more organized and eventually expired as a result of pneumonia and sepsis.
systematic approach to the choice of anticoagulation therapy in Prior to initiating anticoagulation, it is prudent to assess the risks and benefits of anticoagulation in the very old. A cost- The prevalence of stroke varies in patients with AF de- effective analysis revealed that the gain in quality adjusted life pending on other risk factors. High risk factors are prior stroke expectancy from anticoagulation for AF declines with advanc- and/or TIA or systemic embolus, history of hypertension, poor ing age. While a 23% increase in quality adjusted life years LV function, age older than 75, rheumatic mitral valve dis- occurs in a 65-year-old patient with hypertension, diabetes mel- ease, and prosthetic heart valve. Moderate risk factors are age litus, and previous transient ischemic attack (TIA) or stroke, the 65 to 75, diabetes mellitus, and coronary artery disease with increase in quality adjusted life years is only 4% for a 100-year- preserved LV function. Low risk factors include age younger old Baseline quality of life prior to initiating anti- than 65 with no clinical or echocardiographic evidence of coagulation is also relevant and “recommendations that all older persons with AF should be anticoagulated” are Another stroke-risk scheme, CHADS may be more Dementia and its associations are common in the LTC resident applicable to the profile of older adults who typically reside in and should be a consideration during discussions related to Using this scheme, 1 point is assigned to each the quality-of-life issues. One report, based on the Framingham following: congestive heart failure, hypertension, age older Heart Study, produced 5-year risk estimates for stroke or death than 75, and diabetes mellitus. Two points are assigned for a in patients with AF based on a patient’s risk factors at history of stroke or TIA and the higher the score the greater It indicated that risk estimates for stroke or death do not apply the risk. Using this scheme, the stroke rate per 100 patient- for patients on However, the data were based on years is 18.2 with a score of Based on either risk community older adults and not LTC residents, who are often assessment, our patient did not have a high risk for stroke.
older and present with more comorbidity compared with com- Pooled data from 5 randomized, controlled trials indicate munity patients. Thus, the Framingham risk estimates may not that warfarin therapy significantly reduces the stroke rate (by directly apply to our particular case scenario. A second recent 68%) in older patients with AF with hardly any increase in report suggests that the perception of underuse of anticoagulants The effect of aspirin (choice of most physicians in may need revision. In that report, the authors found 117 cases of our survey when anticoagulation was not considered) was AF (13%) in 934 LTC residents and observed that oral antico- Following a stroke or TIA, for secondary pre- agulation was prescribed 46% of the time, whereas aspirin or vention in AF, warfarin had substantial benefit while aspirin clopidogrel were provided 40% of the This report pro- vides data that are in contrast to a perceived trend of underusing Clinical practice guidelines by the American Geriatrics anticoagulation in AF. Our 87-year-old LTCF resident did not Society (AGS) recommend a target INR of 2.5 with an appear to have much to gain in terms of quality of life from acceptable range of 2.0 to The efficacy of warfarin significantly declines when the INR is below 2.0, while the The physician’s opinion regarding anticoagulation may not risk of complications due to bleeding increases with an INR necessarily reflect the patient’s viewpoint. A Canadian study above These recommendations do not differ for the conducted in tertiary and peripheral referral centers revealed This report recommends maintaining an INR be- that patients at high risk for AF valued prevention of stroke tween 2 and 3 even in elderly patients with AF.
more than avoidance of bleeding, compared with the physicians The benefits from aspirin or a combination of aspirin plus who treated Hence, patients and/or caregivers should be fixed dose warfarin are clearly inferior to adjusted dose warfa- encouraged to participate in the decision-making process.
Perhaps the reluctance to use long-term warfarin re- Known and Suspect Drug/Herb Interactions With Warfarin* (that potentiate the effects of warfarin)
(that diminish the effects of warfarin)
NSAIDs, non-steroidal anti-inflammatory drugs.
* Two recent reviews are recommended to the lates to the fear of bleeding, which is its most significant adverse in Phenytoin, steroids, ranitidine, and propylthiouracil, event. Other known risk factors for bleeding include advanced all used in the elderly, have variable effects. Congestive heart age (Ͼ 75), intensity of anticoagulation (INR Ͼ 4), treated failure (through impaired hepatic synthesis) and hyperthyroid- hypertension, cerebrovascular disease, serious heart disease, re- ism (due to hypermetabolic state) are associated with low con- nal insufficiency, and Concomitant aspirin use centration of clotting factors and increased warfarin sensitivity, increases this risk of bleeding most in the initial weeks of while hypothyroidism decreases the effects of warfarin-aspirin Occult pathologic lesions are often a We recognize that our study has limitations. The survey basis for bleeding even with therapeutic hence it is involved only 100ϩ physicians, all from one geographic area, prudent to exclude other causes of bleeding and not ascribe and physician responses possibly may have been influenced by bleeding to anticoagulation alone. Although past gastrointesti- similar practice patterns among trainees and mentors. Further, nal bleeding is described as a risk factor, peptic ulcer disease the survey was based on a single, specific case scenario. We alone (without past bleeding) is not a Prohibitive risks should remain cautious in generalizing our findings, as it is typically seen in the nursing home setting include serious non- possible that another case scenario might elicit a different re- compliance, active bleeding, and recent intracranial hemor- sponse from the same physicians. As suggested in the litera- rhage, while intermediate risks include age older than 80 years we believe that there is the need to conduct studies on and a history of Thus, the nursing home resident for our a larger scale and under different settings to identify current survey posed an intermediate risk for bleeding. A recent review physician thinking regarding the issue of long-term anticoagu- of MEDLINE lists bleeding diathesis, thrombocytopenia (below lation in older long-term residents. An approach with precau- 50,000/L), noncompliance, and uncontrolled hypertension tions for warfarin use in the nursing home is presented in (Ͼ160/90 mm Hg) as absolute contraindications (Grade C recommendation), while a predisposition to falls was not con-sidered a contraindication (Grade A CONCLUSION
Individuals older than 60 years appear more sensitive to the The literature suggests that long-term anticoagulation for anticoagulant effects of warfarin due to altered pharmacokinet- thromboembolic prophylaxis in AF may be beneficial. However, recent reports suggest that warfarin is underused for this indica- long-term care setting, may lead to potentially dangerous drug- tion particularly in older adults, and even more so in the long- drug interactions involving warfarin. Drugs interacting with the term care setting. Our physician poll, based on a specific case cytochrome P450 system in the liver may interfere with warfarin scenario, is consistent with this opinion at both the trainee and metabolism and alter INR. Many herbal remedies are well practicing physician levels. While absolute and relative contra- recognized to cause drug interactions with warfarin. Many more indications to the use of long-term warfarin exist, each case must herbal remedies are suspected to have a drug interaction with be assessed for anticoagulation, based on risks and benefits, as warfarin. A listing of the known and suspect herbs is presented also the consideration of quality of life of the individual.
Precautions for Warfarin Use in the Long-term Care 6. McCormick D, Gurwitz JH, Goldberg RJ, et al. Prevalence and quality of warfarin use for patients with atrial fibrillation in the long term caresetting. Arch Intern Med 2001:161:2458 –2463.
7. Lau E, Bungard TJ, Tsuyuki RT. Stroke prophylaxis in institutionalized elderly patients with atrial fibrillation. J Am Geriatr Soc 2004;52:428–433.
● Review all medications, prescribed and over the 8. Man-Son-Hing M, Laupacis A. Anticoagulant-related bleeding in older persons with atrial fibrillation. Arch Intern Med 2003;163:1580 –1586.
● Initial starting dose (usually 5 mg or less daily in the 9. Fang MC, Stafford RS, Ruskin JN, Singer DE. National trends in anti- arrhythmic and antithrombotic medication use in atrial fibrillation. Arch ● Monitor INR until stable at target level.
10. Gottlieb LK, Salem-Schatz S. Anticoagulation in atrial fibrillation. Does ● Continue to monitor INR monthly.
efficacy in clinical trials translate to effectiveness in practice? Arch Inter ● Additional monitoring with changes in medications, 11. Desbiens NA. Deciding on anticoagulating the oldest old with atrial ● Caution regarding use of dietary supplements and fibrillation: Insights from cost-effective analysis. J Am Geriatr Soc 2002; ● Minimize marked variations in diet (especially with 12. Wang TJ, Massaro JM, Levy D, et al. A risk score for predicting stroke or death for individuals with new-onset atrial fibrillation in the community: ● Minimize hazardous activities predisposing to falls.
The Framingham Heart Study. JAMA 2003;290:1049 –1056.
Management of warfarin therapy prior to surgery.
13. Latif AKA, Peng X, Messinger-Rapport BJ. Predictors of anticoagulation ● Warfarin to be withheld for 4 to 5 days prior to most prescription in nursing home residents with atrial fibrillation. J Am Med ● In high-risk situations (AF with history of stroke, 14. Devereaux PJ, Anderson DR, Gardner MJ, et al. Differences between mechanical valves), consider bridging therapy using perspectives of physicians and patients on anticoagulation in patients unfractionated heparin or low molecular weight with atrial fibrillation: Observational study. BMJ 2001;323:1–7.
heparin perioperatively (weigh patient safety, 15. Majerus PW, Tollefsen DM. Anticoagulant, thrombolytic and antiplate- let drugs. In: Hardman JG, Limbird LE, eds. Goodman and Gillman’s ● Dental procedures, arthrocentesis, cataract surgery, The Pharmacological Basis of Therapeutics, 10th ed. New York, NY: upper endoscopy, and colonoscopy without biopsy McGraw-Hill, 2001, pp. 1519 –1538.
16. Albers GW, Dalen JE, Laupacis A, et al. Antithrombotic therapy in Management of high INR values (with or without bleeding) atrial fibrillation. Chest 2001;119:194S–206S.
● Consider skipping 1 or 2 doses versus lowering the 17. Gage BF, Waterman AD, Shannon W, et al. Validation of clinical classification schemes for predicting stoke: Results from the National ● If risk of bleeding is high (INR over 5), in addition Registry of Atrial Fibrillation. JAMA 2001;285:2864 –2870.
18. Ezekowitz MD, Falk RH. The increasing need for anticoagulant therapy ● If high INR is associated with bleeding, in addition to to perevent stroke in patients with atrial fibrillation. Mayo Clin Proc withholding warfarin, consider use of intravenous vitamin K (10 mg slow IV infusion), fresh frozen 19. Furberg CD, Psaty BM, Manolio TA, et al. Prevalence of atrial fibrilla- plasma, or prothrombin complex concentrate.
tion in elderly subjects (the Cardiovascular Health Study). Am J Cardiol ● Contact physician to determine the need for 20. Krahn AD, Manfreda J, Tate RB, et al. The natural history of atrial fibrillation: Incidence, risk factors, and prognosis in the Manitoba Fol- AF, atrial fibrillation; INR, international normalized ratio.
low-up Study. Am J Med 1995;98:476– 484.
21. Latif AA, Messinger-Rapport BJ. Should nursing home residents with atrial fibrillation be anticoagulated? Cleve Clin J Med 2004;71:40 – 44.
22. EAFT (European Atrial Fibrillation Trial) Study Group. Secondary prevention in non-rheumatic atrial fibrillation after transient ischaemicattack or minor stroke. Lancet 1993;342:1255–1262.
23. Fang MC, Chang Y, Hylek EM, et al. Advanced age, anticoagulation REFERENCES
intensity, and risk for intracranial hemorrhage among patients taking 1. Gurwitz JH, Monette J, Rochon PA, et al. Atrial fibrillation and stroke warfarin for atrial fibrillation. Ann Intern Med 2004;141:745–752.
prevention with warfarin in the long term care setting. Arch Intern Med 24. Hart RG, Benavente O, McBride R, et al. Antithrombotic therapy to prevent stroke in patients with atrial fibrillation: A meta-analysis. Ann 2. American Geriatrics Society. Clinical practice guidelines. The use of oral anticoagulants (warfarin) in older people. J Am Geriatr Soc 2002;50: 25. Levine MN, Raskob G, Landefeld S, et al. Hemorrhagic complications of anticoagulant treatment. Chest 2001;119:108S–121S.
3. Atrial Fibrillation Investigators. Risk factors for stroke and efficacy of 26. De Met PAGM. Herbal remedies. N Engl J Med 2002;347:2046 –2056.
antithrombotic therapy in atrial fibrillation: Analysis of pooled data from 27. Izzo AA, Di Carlo G, Borrelli F, Ernst E. Cardiovascular pharmacother- five randomized controlled trials. Arch Intern Med 1994;154:1449–1457.
apy and herbal medicines: The risk of drug interaction. Int J Cardiol 4. Fuster V, Ryden LF, Asinger RW, et al. ACC/AHA/ESC guidelines for the management of patients with atrial fibrillation: Executive summary (a 28. Ganeval D, Fischer AM, Barre J, et al. Pharmacokinetics of warfarin in report of the American College of Cardiology/American Heart Associ- the nephrotic syndrome and effect on vitamin K-dependent clotting ation Task Force on Practice Guidelines and the European Society of factors. Clin Nephrol 1986;25:75– 80.
Cardiology Committee for Practice Guidelines and Policy Conferences).
29. Dunn AS, Turpie AG. Perioperative management of patients receiving J Am Coll Cardiol 2001;38:1231–1266.
oral anticoagulants: A systematic review. Arch Intern Med 2003;163: 5. Snow V, Weiss KB, LeFevre M, et al: Management of newly detected atrial fibrillation: A clinical practice guideline from the American Acad- 30. Garcia DA, Libby EN, Rich JS. Perioperative anticoagulation for patients emy of Family Physicians and the American College of Physicians. Ann with mechanical heart valves: A model comparing unfractionated and low molecular weight heparin. J Clin Outcomes Management 2005;12:25–31.

Source: http://www.fondazionemadrecabrini.org/Portals/63/Anticoagulate%20or%20not%20patient%20with%20dementia.pdf