The occurrence of AF is associated with important comorbidities that worsen quality of life and decrease survival rate. The need to take more medications, with higher costs to patients, more frequent blood sampling for laboratory tests, more frequent visits to the doctor’s office or hospitalizations, complications from the use of anticoagulants, increased fatigue with more frequent development of heart failure and functional limitations resulting from limb paralysis are factors that significantly worsen quality of life, not to mention in crease risk of death [8, 9].

However, the most worrisome AF data are on stroke. People with AF have a stroke rate approximately fivefold that of the general population, 15 % of all strokes occur in people with AF (25 % among elderly), and the risk of stroke in AF increases from 1.5 per year in the age group from 50 to 59 years to 23.5 % per year in the age group of 80–89 years [10, 11].The average risk of stroke in AF in the studies Atrial Fibrillation Investigators, Stroke Prevention and Atrial Fibrillation-SPAF and in the Framingham cohorts [12–14] ranged from 3.0 to 4.2 % per year, being higher in many elderly people. The risk factors identified for stroke in patients with AF are outlined in Table 1.

We should also keep in mind that AF is the main cause of embolic events of cardiac origin, occurring in nearly half those events, followed distantly by acute myocardial infarction (AMI) and left ventricular aneurysm [15].Furthermore, AF-related stroke is a more severe stroke, with a higher likelihood of causing significant motor sequelae and twice as likely to leave the patient permanently bedridden; therefore, it has a higher likelihood of death [10, 11].

However, we must remember that thromboembolic events occur in other areas in patients with FA, although stroke is frequently the most severe outcome. The incidence of stroke ranged from 48 to 82 %, the incidence of limb embolism ranged from 13 to 38 %, and visceral embolism occurred in 4–42 % of patients in several case series that analyzed arterial embolism in AF [16–20]

The attempt to characterize risk factors for stroke and systemic embolism in non-valvular AF in several studies [12–14] led to the identification of the following elements that comprise the CHADS₂ (congestive heart failure, hypertension, advanced age, diabetes, stroke) score in 2001: Prior stroke or transient ischemic attack (TIA), 2 points, and 1 point each for diabetes mellitus, heart failure, hypertension and advanced age (>75 years). Thus, the score ranges from 0 to 6 and is easy to use: 0 means low risk, 1 means intermediate risk, and 2 or higher means high risk for events. CHADS₂ [21] provides an adjusted risk of stroke ranging from 1.9 % in patients with risk 0–18.2 % for patients with score 6 and is recommended in several guidelines as a good indicator of risk of stroke and embolic events [22].

In 2010, Lip et al., based on data from the EuroSurvey, proposed a refinement of CHADS₂, the CHA₂DS₂-VASc [23], which ranges from 0 to 9 points. In that new score system, hypertension, diabetes and heart failure (or ejection fraction <40 %) continue to score 1 point, as in CHADS₂; stroke or TIA (or any other thromboembolic event) continues to score 2 points; age older than 75 years scores 2 points; age between 65 and 74 years of age 1 point; and the presence of vasculopathy in another area and female gender also score 1 point. That new score is more sensitive in the low-risk range and apparently enables us to better separate who effectively lacks risk factors and is not required to use anticoagulants from who has low risk but requires medications.

It is extremely important to accurately assess the risk of bleeding when intending to indicate the use of oral anticoagulants. That condition is particularly difficult in AF because many known risk factors for the occurrence of thromboembolic events are also risk factors for bleeding, including heart failure, advanced age, hypertension, prior stroke and concomitant use of antiplatelet drugs. The HAS-BLED [24] risk score was developed using scores from 0 to 9, and a score >3 indicates a high risk of bleeding (Table 2). That score serves as a counterpoint to the clinical decision of whether to use anticoagulants in a patient with AF and reminds us of the limitations we face in decision-making processes because a patient at high risk for stroke most likely will also be a patient at high-risk for bleeding. In those cases, the careful use of drugs at lower doses, or a slightly lower international normalized ratio (INR) in the case of warfarin, is most likely the most indicated.

Warfarin has been used for over 50 years and is undoubtedly a highly effective drug. It was initially used in humans only after thoroughly proving it caused severe bleeding and death in rats because of its anticoagulant effect. Due to the prothrombotic effect that occurs in some conditions and in AF with consequent severe systemic thromboembolic events, there was a demand for an effective anticoagulant, and hence, the opportunity to use warfarin emerged.

Several studies have shown a significant reduction of events in patients treated with warfarin versus placebo or antiplatelet drugs, and the meta-analysis by Hart et al. [25] showed a 61 % decrease in the risk of stroke, 31 % decrease in death rates, and 44 % decrease in stroke or acute myocardial infarction (AMI).

Warfarin exercises its anticoagulant effect by inhibiting the γ-carboxylation of vitamin K-dependent coagulation factors (factors II, VII, IX and X), and its effect may be measured by analyzing the result of the INR. Normal INR has an approximate value of 1, and we usually try to keep the INR between 2 and 3 because values higher than 3 significantly increase the risk of bleeding; conversely, INR values lower than 2 do not provide adequate protection against thromboembolic events. Thus, patients using warfarin must remain in a narrow INR range to prevent the risk of ischemic stroke (low INR, inadequate anticoagulation) or hemorrhagic stroke or other severe bleeding events (high INR, excessive anticoagulation). Thus, the individual use of warfarin has a severe risk of bleeding, and there is also the risk of its being ineffective if not within an optimal therapeutic range. Warfarin may also be used in patients with chronic renal failure or undergoing dialysis.

The narrow therapeutic range is one of the main limitations to the use of warfarin. We could further note that warfarin, combined with insulin, is one of the most commonly cited drugs that cause complications in emergency room visits in the US [26]. Other important limitations include interactions with food and drugs, the need to collect blood periodically for INR tests and important laboratorial variations in the INR analysis, which may cause serious problems of under or overdosing. There is some resistance from physicians in using the drug in very elderly individuals or those with increased risk of bleeding, which contributes to a high number of patients needing to take the medication who are ultimately excluded from the treatment. Some studies estimate that the number of high-risk patients who are not treated with warfarin may reach 50 %, regardless of whether it is a university or general hospital [27]. Furthermore, there is evidence that warfarin use does not increase with higher CHADS₂ scores, remaining at approximately 60 % of the optimal use, which is highly undesirable because it prevents many patients in extreme need from receiving this medication [28]. Data have been published confirming the care required when using warfarin in very elderly patients: Hylek et al. [29] reported 13 % severe hemorrhagic complications in people older than 80 years, with 26 % of patients discontinuing the treatment after using warfarin for one year. Those data emphasize the need to use warfarin with discretion and common sense, ensuring that patients are able to understand the caution required when using that type of drug and that patients will not stop taking the drug, use other drugs concomitantly or miss the collection of INR

In São Paulo, returning to the tele-ECG experiment mentioned above [7], we found that 70 % of AF cases diagnosed were classified as high-risk, although only 24 % of those cases were anticoagulated with warfarin and made no use of new drugs. All of the difficulties in using warfarin combined with the difficulty to schedule appointments in the Brazilian public health system, the consultations with different physicians each time, the fear that cardiologists have in prescribing a high-risk drug, especially to elderly patients, and comorbidities that may lead patients to take other medications, including anti-inflammatory drugs, make it extremely difficult to succeed in reaching a suitable level of regular warfarin use. Therefore, many strokes that could be prevented are not, even though the drug has an extremely low cost. Thus, similar to other undeveloped countries, Brazil has a lower rate of permanently adequate INR (meaning improper medication, non-standardized laboratories or external interference with the drug), which, added to the logistical difficulties of the Brazilian healthcare system, hinders even more the proper maintenance of the treatment. The information to be collected from the AF Registry of the Brazilian Society of Cardiology will be extremely important to better assess that scenario and confirm or refute the results found in São Paulo in the AF sample analyzed through tele-ECG.