Atrial fibrillation is a clinically encountered arrhythmia and encompasses lone atrial fibrillation to paroxysmal AF to chronic atrial fibrillation. Atrial fibrillation is mostly associated with heart failure, aging and diseases related to aging (Mathew et al. 2009). Atrial fibrillation has been associated with cardio embolic stroke, which can be neurologically devastating (del Conde and Halperin 2013). The major clinical feature of AF is a decrease in atrial contractility and increased atrial compliance, which leads to mechanical remodelling of the heart. This leads to stretching in the atrial myocardium and this atrial remodelling increase the atrial fibrosis and decreases the conduction velocity, which also shortens the refractory period in atria (Mathew et al. 2009). AF is also caused due to an irregular and increased conduction of impulses in atrial part of the heart. AF reduces the flow velocity of the left atria and causes delayed emptying from the atria which leads to thrombus formation. Strokes caused by AF have higher fatality when compared to other risk factors due to the formation of large thrombi in the atria which may occludes in the cerebral arteries causing infarction and stroke (Alberts et al. 2012).

The US practise guidelines recommended the use of CHADS2 score (assigns score for congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, transient ischemic attack or prior occurrence of stroke). This scheme assesses different risk factors and most important of those are the age and prior stroke. It assigns a score for each risk factor with ‘1’ except prior stroke which is assigned ‘2’ which makes up score ‘6’ in total (Poli et al. 2007). The patients with a score of 6 showed an increase of 2.5 to 2.7 folds in the hospitalization risk due to stroke (Naccarelli et al. 2012). Pre-admission CHADS2 score relates to the severity and stroke outcomes in patients with atrial fibrillation (Sato et al. 2011).

The European guidelines recommended the use of a new risk stratification scheme called CHA2DS2-VASc scheme, which was superior to the CHADS2 scheme as it also assesses additional risk factors that were previously underestimated. CHA2DS2-VASc score (assigns score for congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, transient ischemic attack or prior stroke – vascular disease, age 65-74 years, sex (female)) is used to assess all the possible risk factors, and scores them ‘1’ for all the factors except age≥75 years and prior stroke which are given ‘2’ where, the maximum score is ‘9’ (Chao et al. 2011).

This scheme assesses additional risk factors such as vascular diseases such as myocardial infarction, peripheral artery disease, and complex aortic plaque. The patients with a score of ‘9’ show an increase of 3.0 folds in the hospitalisation risk due to severity of the condition (Naccarelli et al. 2012). The CHA2DS2-VASc scheme is more advantageous when compared to CHADS2 scheme as it also assesses additional risk factors and is also useful in identifying patients who despite having a CHADS2 score of ‘0’, still have a high risk of stroke and can be prescribed anticoagulants. It is also used to identify the patients who will truly benefit from the antithrombotic agents and also who are at low risk of stroke with a CHA2DS2-VASc score equal to ‘0’ (del Conde and Halperin 2013). The score ‘2’ in both the schemes recommends the use of anti-platelet or anticoagulant therapy. The score ‘1’ which is risk factor specific can also be considered for anti-coagulant therapy according to the recent guidelines. In both scoring systems Age and prior stroke are considered as major risk factors (Wadke 2013).

Warfarin is a vitamin K antagonist, which has an inhibitory effect on the vitamin K cycle which in-turn inhibits the vitamin K dependent coagulation factors II, VII, IX and X. Warfarin inhibits the synthesis of vitamin K dependent coagulation factors (II, VII, IX, X, protein C and protein S) by inhibiting the γ-carboxylation of glutamate residues of these clotting factors, which ultimately inhibits them to bind with Calcium. The principle mechanism behind the anti-coagulant effect of warfarin is that it inhibits the reduction of oxidized vitamin K which is essential for the activation of clotting factors by carboxylation. This inhibits the formation of functionally active prothrombinase complex, thrombin and fibrin which, ultimately gives an anti-coagulant effect. The figure 1 shows us the coagulation pathway and inhibitory effect of warfarin (Riley et al. 2000). Warfarin is mainly used to prevent venous thromboembolism and systemic embolism in patients with atrial fibrillation or prosthetic heart valves (Eriksson and Wadelius 2012).

Warfarin is a highly plasma bound drug which is completely absorbed from the upper intestinal tract and reaches peak plasma level in one hour. It is 98-99% plasma bound and has a half-life of 40 hours. Warfarin is metabolised by microsomal hepatic enzymes and converted into hydroxylated inactive metabolites which, undergo conjugation with glucuronic acid and the conjugates are excreted in urine and faeces (Riley et al. 2000).

According to WHO guidelines, INR has been defined as, ‘For a given plasma or whole blood specimen from a patient on long-term oral anticoagulant therapy, a value calculated from the prothrombin-time ratio using a prothrombin-time system with a known ISI (international sensitivity index) according to the formula INR = (PT/MNPT)ISI.’ Determination of INR is mandatory for the controlled use of oral anti-coagulants (van den Besselaar et al. 2004). INR has been a great form of TDM (Therapeutic drug monitoring) in the patients who are on warfarin. The British Society of Haematology and the American college of Chest Physicians recommend a therapeutic range of INR as 2-3. It has been observed that an increase in INR more than 4.5 has shown an exponential increase in the risk of bleeding. The monitoring should start from second and third dose which should be continued until the INR reaches an optimal level and this frequency can be reduced when the stable dose administration is achieved (Hu et al. 2012).

European guidelines recommended a bleeding risk score which is called the HAS-BLED score (assigns score to hypertension, abnormal renal/liver function, stroke – bleeding history, labile international normalization ratio, elderly (Age>65 years), and drugs/alcohol concomitantly) which is calculated as 1 point for each factor with a maximum score of 7 (Kiviniemi et al. 2014). This score is used to assess the bleeding risk in atrial fibrillation patients. The patients with HAS-BLED score 0-2 have low risk of bleeding and with more than 3 have high risk of bleeding which should be treated by reversing the anticoagulant effect and by maintaining the INR by dose adjustments (Lip 2011).

It is a bleeding score, which is the only score that includes the genetic factors. HEMORR2HAGES scheme (assigns points for hepatic or renal disease, ethanol abuse, malignancy, older age (greater than 75), reduced platelet count or function, re-bleeding risk, uncontrolled hypertension, anaemia, genetic factor, excessive fall risk, and stroke) is used to assess different risk factors and scores one for all the risk factors except re-bleeding risk which is scored two points (del Conde and Halperin 2013). This is the only scoring system, which considers genetic factors (CYP2C9), excessive fall and stroke (Alberts et al. 2012).

Stroke severity can be measured using a severity scale, the National Institutes of Health Stroke Scale (NIHSS) that includes 15 item scales that include assessment of motor function and sensory function, level of consciousness and orientation. It ranges from 0 to 34 where if the patient score is below 10 they are likely to have a favourable outcome after 1 year when compared to patients with a score of more than 20. Patients with severe stroke and with a score of more than 22 have poor prognosis (Institutes et al. 2009). NIHSS shows a high accuracy of prediction of ADL (activities of daily living) dependency by the patients post-stroke (Kwakkel et al. 2010).

The modified Rankin scale has been devised to assess the disability after stroke. It is a measure of functional independence, which incorporates the WHO components such as body function, activities and participations. A patient using adaptive but still does not need any assistance is considered independent and a patient who takes aid of another person is considered dependent (Kasner 2006). This scale ranges from 0 to 6 where 0 indicates no symptoms; and 1-2 indicates mild disability; 3 shows moderate disability; and 4 to 5 show severe disability and 6 is death. This shows favourable out-come with a score of 0-1 and unfavourable score of 2-6. Table 4 shows Modified Rankin Scale in detail (Institutes et al. 2009). It has been shown that modified rankin scale can be used to study the quality of life in stroke survivors (Singhpoo et al. 2012).

Glasgow coma scale has been a simple measure of conscious level and this scale has been widely accepted by Neurological units in most of the English speaking countries. It has been a cumulative measure of arousal, awareness, and activity, which have been termed as eye opening response, best verbal response and best motor response. The maximum score calculated is 15 and this is further categorised in as minor (GCS ≥ 13), moderate (GCS 9–12), and severe injury (GCS<9). The components of GCS are better described inTable 5 (Barlow 2012) (Matis and Birbilis 2008).

BI is a scale, which consists of 10 items that measure the function in terms of activities such as, daily living and mobility. The scale ranges from 0 to 100 the higher the score, more the patient is independent. The score 100 indicates that he is independent in feeding, dressing, getting in and out of bed, bathing, can walk at least one block; and can ascend and descend stairs without help (Katz for the Association of Rheumat 2003). A study investigating the accuracy of Barthel Index revealed good discriminative properties at 6 months post-stroke (Kwakkel et al. 2011).

Six simple variable model has been developed in Oxfordshire Community Stroke Project (OSCP) to predict the survival free of dependency (modified Rankin scale <3) (Dennis 2008). SSV model is useful in measuring the dependency and death in stroke patients. This model includes six different variables such as age, verbal component of Glasgow coma scale (GCS), and arm power, ability to walk, pre-stroke living condition and pre-stroke dependency (Institutes et al. 2009). SSV model predicts the outcome of stroke patients with cerebral infracts with a great precision (Li et al. 2012).