Thromb Haemost 2018; 118(12): 2014-2017
DOI: 10.1055/s-0038-1676074
Invited Editorial Focus
Georg Thieme Verlag KG Stuttgart · New York

Optimizing Stroke and Bleeding Risk Assessment in Patients with Atrial Fibrillation: A Balance of Evidence, Practicality and Precision

Marco Proietti
1  Istituto di Ricerche Farmacologiche “Mario Negri” IRCCS, Milan, Italy
2  Department of Internal Medicine and Medical Specialties, Sapienza University of Rome, Rome, Italy
,
Nebojsa Mujovic
3  School of Medicine, University of Belgrade, Belgrade, Serbia
4  Cardiology Clinic, Clinical Centre of Serbia, University of Belgrade, Belgrade, Serbia
,
Tatjana S. Potpara
3  School of Medicine, University of Belgrade, Belgrade, Serbia
4  Cardiology Clinic, Clinical Centre of Serbia, University of Belgrade, Belgrade, Serbia
› Author Affiliations
Further Information

Address for correspondence

Tatjana S. Potpara, MD, PhD
Cardiology Clinic, Clinical Centre of Serbia
Visegradska 26, 11000 Belgrade
Serbia   

Publication History

31 October 2018

31 October 2018

Publication Date:
20 November 2018 (eFirst)

 

Optimal stroke prevention using oral anticoagulant (OAC) therapy is essential for the management of patients with atrial fibrillation (AF) who are at increased risk of stroke, but decision making on thromboprophylaxis requires evaluation of individual patient's thromboembolic and bleeding risks.[1] Several risk assessment tools have been developed to streamline the evaluation of patients with AF but navigating through the labyrinth of options may be confusing for clinicians.[2]

Stroke Risk in AF: Evidence versus Practical Decision Making

In this issue of Thrombosis Haemostasis, Borre et al[3] reported a comprehensive systematic review and meta-analysis commissioned by the Patient-Centred Outcomes Research Institute to update a 2013 Agency for Healthcare Research and Quality review[4] of the comparative accuracy and impact on clinical decision making of several thromboembolic and bleeding risk assessment tools in patients with AF.

They addressed two key topics: (1) the comparative diagnostic accuracy and impact on clinical decision making of available clinical and imaging tools and associated risk factors for predicting thromboembolic risk; and (2) the comparative diagnostic accuracy and impact on clinical decision making of clinical tools and associated risk factors for predicting bleeding events. The authors conducted an extensive literature search (from 2000 to 2018) and evidence appraisal of published data for clinical stroke scores (CHADS2, CHA2DS2-VASc, Framingham and age, biomarkers and clinical history [ABC]-stroke), cardiac imaging predictors and clinical bleeding risk scores (HAS-BLED, HEMORR2HAGES, ATRIA, Bleeding Risk Index and ABC-bleeding).

Not surprisingly, the CHADS2 and CHA2DS2-VASc scores were the most commonly evaluated stroke risk assessment tools (29 and 24 studies, respectively). Overall, the CHADS2, CHA2DS2-VASc and ABC-stroke scores had the best evidence for prediction ability for stroke—however, notwithstanding the heterogeneity among studies included in the meta-analyses, the CHADS2, CHA2DS2-VASc, Framingham and ABC-stroke scores all had a modest prediction ability for thromboembolic events, with c-statistic values ranging from 0.63 (Framingham) to 0.69 (continuous CHADS2 score).[3] This is in line with previous meta-analyses[5] [6] [7] [8] showing broadly similar performance of various clinical risk factor-based stroke scores (including the CHA2DS2-VASc) in prediction of thromboembolic events. However, the difference of the CHA2DS2-VASc in comparison to other stroke scores is its relative accuracy in identifying AF patients at truly low risk of stroke (i.e. CHA2DS2-VASc of 0 in men, or 1 in women) who do not need any anti-thrombotic therapy.[8] [9] [10]

Recognizing that all clinical scores only have modest predictive value for high-risk patients that sustain events and that current risk scores are designed to be simple and reductionist, the use of a simplified, clinical risk factor-based approach to the management of AF-related stroke risk has been acknowledged by recent international AF guidelines.[1] [11] The default should be to ‘offer stroke prevention, unless the patient is “low risk”’. A simple message is needed, as guideline-adherent treatment has been associated with improved outcomes in multiple AF cohorts.[12] [13] [14] [15] [16]

The presence of even a single CHA2DS2-VASc risk factor is associated with an excess in stroke and mortality,[17] [18] and OAC use is associated with positive net clinical benefit in comparison to aspirin (which is harmful) or no therapy.[19] Variations in reported event rates for patients' single-risk-factor AF cohorts have caused debate[20] but may reflect different study settings (hospital vs. community), not all risk factors carrying equal risk, methodological errors (with some papers ‘conditioning on the future’ by excluding patients who had ever started OAC even during follow-up) and the inclusion of females (who have 1 point, yet low risk) in reports of event rates with CHA2DS2-VASc score of 1.[16] [21] Indeed, female sex is rather a risk modifier than an independent stroke risk factor and becomes relevant only in the presence of other CHA2DS2-VASc factors[22]; however, ignoring female sex (i.e. using the CHA2DS2-VA score, without ‘Sc’[23]) could under-estimate stroke risk in females with AF.[24]

As mentioned, clinical risk factor-based scores generally have a modest prediction ability for the outcome event of interest. Adding various biomarkers (‘biological markers’) to clinical scores—that is, blood biomarkers (e.g. brain natriuretic peptide, cardiac troponin, creatinine, etc.), cardiac or cerebral imaging (echocardiographic, magnetic resonance, etc.) or electrocardiographic indices in sinus rhythm (e.g., the P wave axis[25])—improve the score's prediction ability, at least statistically, but such complexity may be impractical or unfeasible outside a highly structured setting.[2]


#

Vascular Disease and Stroke Risk

Optimal acknowledgment of clinical information pertaining to specific CHA2DS2-VASc score component is crucial for proper stroke risk assessment in each patient. For example, the ‘V’ (vascular disease) traditionally includes validated factors such as previous myocardial infarction, complex aortic plaque or peripheral artery disease.

In this same issue of Thrombosis Haemostasis, Steensig et al[26] extend their recent report on significant association between the presence of angiographically documented coronary artery disease (CAD) and subsequent thromboembolic events[27] showing that the extent of CAD (i.e. 1-, 2-, 3-vessel or diffuse) did not add additional risk prediction information regarding ischaemic stroke, transient ischaemic attack or systemic embolism among 12,690 AF patients undergoing coronary angiography (with CAD diagnosed in 59.4%) over a 3-year follow-up. Hence, angiographically documented CAD should be included in the ‘V’ component of the CHA2DS2-VASc score. This would translate to an indication for OAC use in 3% of low-risk patients according to the European AF guidelines, or in 6% per the U.S. guidelines. Non-invasively diagnosed CAD using contrast angiography or stress testing would probably have the same significance, albeit not investigated in the present study.


#

Bleeding Risk Assessment

Bleeding risk assessment is a sensitive part of risk evaluation in AF patients that is sometimes misinterpreted or even abused.[28] All international AF guidelines recommend bleeding risk assessment but beyond listing the bleeding risk factors, practical guidance on clinical decision making on OAC use in high-risk AF patients at increased risk of bleeding is less well specified.[1]

In their systematic review and evidence appraisal, Borre et al[3] also explored 38 studies that reported on bleeding risk in AF patients. Overall, they found moderate strength of evidence on increased bleeding risk in AF patients with chronic kidney disease, and that the HAS-BLED score provided the best prediction of bleeding events among the investigated bleeding risk assessment tools. Indeed, the HAS-BLED score has been extensively validated in various setting, including AF patients on OAC (either vitamin K antagonists [VKA] or non-VKA OACs), aspirin or no anti-thrombotic therapy, with respect to various bleeding outcomes (i.e. major bleeding or intracranial haemorrhage).[28]

Bleeding risk assessment using the HAS-BLED score is not meant to preclude the use of OAC, but to identify high-risk patients in whom modifiable bleeding risk factors such as uncontrolled hypertension (H), labile international normalized ratios in patient taking VKAs (L), concomitant use of non-steroidal anti-inflammatory drugs, anti-platelet drugs or excessive alcohol consumption (D) can be addressed or modified, and who need clinical follow-up earlier rather than later (e.g. 4 weeks rather than 4–6 months). Importantly, using a formal bleeding risk assessment tool (i.e. the HAS-BLED score) had significantly better predictive value for bleeding risk assessment compared with the approach based only on modifiable bleeding risk factors.[29] [30] [31]


#

The Dynamic Nature of Stroke and Bleeding Risk

Many studies of risk factors and biomarkers have investigated something at study entry (‘baseline’) and ascertained outcomes many years later (sometimes 5 or 10 years). The patient's clinical risk profile changes over time and this change has been shown to have better prediction ability for the respective risk than simply relying on the baseline score values.[32] [33] Hence, neither thromboembolic nor bleeding risks are static and must be re-assessed regularly. This is part of a comprehensive and holistic approach to the management of patients with AF[34] (see [Fig. 1]).

Zoom Image
Fig. 1 The Atrial fibrillation Better Care (ABC) pathway. NOAC, non-vitamin K antagonist oral anticoagulant; OAC, oral anticoagulant; TTR, time in therapeutic range; VKA, vitamin K antagonist.

#
#

Conflict of Interest

None.


Address for correspondence

Tatjana S. Potpara, MD, PhD
Cardiology Clinic, Clinical Centre of Serbia
Visegradska 26, 11000 Belgrade
Serbia   


Zoom Image
Fig. 1 The Atrial fibrillation Better Care (ABC) pathway. NOAC, non-vitamin K antagonist oral anticoagulant; OAC, oral anticoagulant; TTR, time in therapeutic range; VKA, vitamin K antagonist.