Chapter 5 - Who is at risk of AF?

Key points

  • For many people AF is silent, with no symptoms to prompt either concern or medical consultation
  • AF is often not detected until the development of a serious complication such as stroke or heart failure
  • Routine opportunistic pulse checks represent an effective and cost-effective method of improving the detection of AF in patients at risk of stroke
  • Many methods are used by doctors to determine stroke risk for an individual patient, and hence the ideal treatment to prevent stroke
  • Patients in the UK may currently be receiving inconsistent advice and therapy, due to a lack of consensus on AF risk stratification

Atrial fibrillation is often present without symptoms

AF patients, because of their elevated risk both of having a stroke and having a severe stroke, represent an ideal focus for stroke prevention. However, we have also discussed in earlier chapters that a substantial proportion of AF patients are currently undiagnosed or inappropriately treated. This chapter reviews detection, diagnosis and decisions on who and how to treat for the prevention of stroke.

Although AF may be recognised by symptoms such as palpitations or dizziness it is commonly asymptomatic, when the patient is unaware of anything that might indicate that they have AF. This is commonly called silent AF. Consequently, many people have undetected AF, and might have had it for some time.8 Studies using heart monitors that record heart rhythms 24 hours a day have shown that it is common for a patient to have periods of both symptomatic and asymptomatic AF.8 It is often the case that a diagnosis of AF is only made following a serious complication such as stroke or heart failure.64

Detection and diagnosis of atrial fibrillation

Effective AF detection and diagnosis strategies are essential before treatment and the prevention of deadly and debilitating strokes become possible. Steps to improve the detection of AF have already been made; guidelines from the National Institute of Health and clinical Excellence (NICE) recommend that an ECG is required for all patients with an irregular pulse whether symptomatic or not.64 ECG stands for electrocardiogram and is the method doctors use to capture the electrical signature of heart rhythm problems. However, for this recommendation to be effective, patients first need to have their pulse checked for irregularity. Modern electric pulse meters have all but taken over this job from doctors and nurses. Unfortunately, extremely few modern electrical pulse check machines are capable of detecting an irregular rhythm. Consequently, one of the most valuable detection methods for AF has largely been eradicated from the NHS by the advance of technology.

Many patients have other risk factors for stroke, such as high blood pressure, diabetes and heart disease. For these conditions they will already be receiving frequent check ups. The introduction of opportunistic screening by making routine manual pulse checks during these check ups would appear to be a prudent, simple and low cost method to increase the detection of AF and prevent strokes.

Opportunistic pulse checks are a low cost option

A UK study, involving almost 15,000 patients, compared opportunistic screening for AF, as described above, with another, systematic, method. Compared to routine clinical care, opportunistic screening identified approximately 50% more cases of AF. Systematic screening identified approximately 70% more case than routine clinical care, but was associated with a high cost per patient, nearly five times the cost of the opportunistic approach.139 The opportunistic screening was associated with a cost of only £363 per patient. However, only a tiny fraction of this was for the inclusion of a manual pulse check at routine appointments. The majority of the cost was accounted for by the subsequent ECG and interpretation;139 steps already required by the NICE Guidelines upon detection of an irregular pulse.

The study highlighted the important role of a simple, routine pulse check, in helping to improve detection of AF. The policy implications arising from the results of this study are that an opportunistic approach using pulse- taking followed by ECG is probably the most cost-effective option for any screening programme implemented through primary care.139

Despite these results, today there is no policy, formal target or recommendation from the Department of Health that recommends or mandates the routine manual pulse checks for any group of patients at GP surgeries.

The role of ambulatory monitoring

Episodes of AF can be both short and infrequent which adds to the difficulty of detection and diagnosis.  For some patients, for example those who have reported instances of their own pulse being irregular, there is potential value in the use of heart rhythm monitors that record the pulse constantly for up to several days at a time.  A patient would need to wear the device for between 12 and 72 hours but the constant monitoring would greatly increase the likelihood of a positive diagnosis.  When prolonged monitoring is used in ischaemic stroke patients, AF is detected in one in 20 cases, providing powerful evidence for the routine adoption of such monitoring in these stroke patients.

Making treatment decisions based upon risk of stroke

Once an AF patient is identified, it becomes necessary to make a treatment decision that best addresses the needs of the patients; lowering risk of stroke as much as possible while not adding further unnecessary risks to the patient’s health. The treatment options to prevent stroke will be covered in more detail in the following chapter but in short; patients at low risk of stroke are often candidates for treatment with the antiplatelet drug aspirin.116 For patients at moderate and high risk of stroke, aspirin has been shown to offer only modest efficacy despite having a similar risk for major bleeding as warfarin,65 104 that is therefore recommended for these patients. To determine who receives which of these two drugs, much attention has been given to this risk factors for stroke in AF patients. Factors reported to increase further the risk of stroke in patients with AF include:8 66

  • Being female
  • Being elderly
  • Having previously had a stroke or TIA (mini stroke)
  • Having high blood pressure
  • Having heart failure or valvular heart disease
  • Having diabetes
  • Having vascular disease

Furthermore, the factors contribute in different ways and by different amounts to the risks of an AF patient having a stroke.

For example, a history of stroke or TIA is the strongest independent predictor of stroke in patients with AF, increasing the risk of another stroke approximately threefold.8 Increasing age also has a marked effect on the risk of stroke: among patients with AF, the incidence of stroke is approximately sevenfold higher in patients in their 80s compared with those in their 40s.67 High blood pressure increases the risk of stroke approximately threefold in patients with AF.68  Vascular disease also independently increases the risk of stroke and death in AF patients.81 194 197  Specifically, the development of coronary artery disease in AF patients has been found to be predictive of the formation of stroke-causing clots.203

Although stroke and AF are both more prevalent in men than in women,69 70 71 the literature shows that death rate from stroke is increased fourfold in women with AF compared with twofold in men with AF.105 However, it should be noted that not all studies have demonstrated such a significant difference between the genders.105 72

Given that every patient is different, it is necessary for doctors to work out individual AF patient risk of stroke from all the above factors. This is called risk stratification. Many approaches to evaluating individual stroke risk in AF patients have been developed. Using these, doctors can quickly assess the patient’s risk and make the necessary treatment decisions.

Approaches to risk stratification

Several different methods are used by UK doctors when seeking to determine the risk of stroke among AF patients. These all provide risk scores based upon the presence of risk factors including those reviewed above.73 74  Risk calculation schemes vary by the risk factors they incorporate and in their methods of scoring and risk evaluation. Schemes in current are, usually known by their acronyms, include: AFI, SPAF, ACCP, Framingham, CHADS2 and CHA2DS2VASc.  Furthering potential for the adoption of different methods among doctors, NICE developed its own new scheme instead of adopting an existing option.

NHS Improvement’s GRASP-AF tool

NHS Improvement has sought to assist GPs with the adoption of risk stratification among AF patients through the introduction of the GRASP-AF tool.  GRASP-AF is made available free to all GPs, allowing them to analyse patient records swiftly and easily for those at risk of stroke because of AF. The tool automatically calculates CHA2DS2VASc and CHADS2 scores for all patients, identifying all those in need of anticoagulation.  Below are listed the automatic functions of the GRASP-AF tool:

  • Identify AF patients
  • Search for risk factors
  • Calculate CHA2DS2VASc and CHADS2 scores
  • Search for current medication status
  • Search for reasons for not treating with an anticoagulant
  • Alert GPs to those patients at risk of stroke but not receiving anticoagulation medication

Considerable advances could be made in the identification and appropriate treatment of AF patients if GPs made routine use of this free and powerful tool.  More information on GRASP-AF can be found in the next chapter and at

The role and limitations of CHADS2

CHADS2 is a simple algorithm that is an evolution of the AFI and SPAF risk schemes.  It is based on clinical trial results and works on a simple points system:

CHADS2 and how it's calculated

C Congestive heart failure: 1 point
H Hypertension (high blood pressure): 1 point
A Age greater than 75 years: 1 point
D Diabetes: 1 point
S2 Stroke or TIA: 2 point

From these risk factors, the acronym CHADS is derived. The ‘2’ is then added denoting that Stroke and TIA attract double the risk score of the other factors. The points for a particular patient are then added up allowing doctors to assign the patient to low, moderate or high risk and prescribe the appropriate treatment.

An authoritative comparison of CHADS2 and 11 other schemes concluded that there were substantial, clinically-relevant differences between them. Most were found to have only a modest ability to predict stroke and that the proportion of patients assigned to low, moderate and high risk categories varied widely between the schemes.

For example, while CHADS2 has been found to be a more accurate predictor of stroke, in patients not on anticoagulation treatment, than the older AFI117 and SPAF118 schemes (from which it was derived), it does not take account of several common stroke risk factors.  Consequently, CHADS2 categorises many patients as at moderate risk despite evidence that many of them would derive benefit from taking an anticoagulant instead of aspirin.116

The development and validation of CHA2DS2VASc

Because of these limitations, a further development was made called CHA2DS2VASc, to complement the CHADS2 scheme. CHA2DS2VASc extends CHADS2 by considering additional common risk factors and focusing on a continuum of risk as opposed to arbitrary levels of low, moderate and high risk. The risk factors and scores for CHA2DS2VASc are listed below.

CHA2DS2VASc and how it's calculated

C Congestive heart failure 1 point
H Hypertension (high blood pressure) 1 point
A2 Age greater than 74 years 2 point
D Diabetes 1 point
S2 Stroke or TIA 2 point
V Vascular disease 1 point
A Age 65-74 1 point
Sc Sex (i.e. female 1 point

CHA2DS2VASc was validated in an analysis from the Euro Heart Survey79 and in several other studies.75 76 77

The most recently published international consensus guidelines, from the European Society of Cardiology, have adopted the CHA2DS2VASc scheme and endorses use of the scheme to categorise the stroke risk among patients with a CHADS2 score lower than two, to ensure the most appropriate treatment decision is made.116

Subsequent to the publication of the ESC guidelines, CHA2DS2VASc has received further support from new data illustrating that the scheme performs better than CHADS2 in predicting patients at high risk of stroke. The new data also showed that patients categorised as low risk by CHA2DS2VASc were indeed at truly low risk for stroke, unlike similarly categorised patient under other schemes.81

Having different schemes, while confusing, becomes inevitable as scientific understanding increases and evidence builds for one method versus another. For example, at least until the next revision of the NICE guidelines on the management of AF, many British doctors will be in the challenging position of having different risk assessment and treatment advice being endorsed by different authorities. As a result, patients in the UK may receive inconsistent advice and therapy, depending on local preferences.

Calculating the risk of treatment-related bleeding

Despite increasing pressure from advanced stroke risk calculation schemes and recommendations of new guidelines, many physicians hesitate to use an anticoagulant because of perceived risks of the patient suffering from dangerous excessive bleeding. This perception is frequently unjustified. To address this concern with bleeding risk, the authors of the ESC consensus guidelines included an additional risk calculation scheme for bleeding among AF patients.

Using data on risk factors for major bleeding from the Euro Heart Survey as well as from a systematic review of the literature, a simple bleeding risk scheme, HAS-BLED, was derived for patients with AF:80

HAS-BLED and how it is calculated

H Hypertension 1 point
A Abnormal renal/liver function 1 point each
S Stroke 1 point
B Bleeding history/predisposition 1 point
L Labile INR (unstable warfarin impact) 1 point
E Elderly (eg, >65 years) 1 point
D Drugs/alcohol 1 point for certain drugs plus 1 for alcohol excess – max 2

HAS-BLED aims to provide a simple way to evaluate risk of bleeding in AF patients so that effective treatment decisions can be made.  HAS-BLED has also been developed to encourage doctors to consider reducing the risk of bleeding by focusing on correctable risks such as high blood pressure and poor warfarin control.  The HAS-BLED score has been validated in large independent groups of AF patients80 200 201 and has also been recommended in recent Canadian treatment guidelines,198

The new ESC guidelines state that it would seem reasonable to use the HAS-BLED score to assess bleeding risk in AF patients on the basis that a score of ≥3 indicates ‘high risk’.116  In the light of growing evidence, the more recent Canadian guidelines recommend that bleeding risk is assessed using the HAS-BLED score.198