Should HAS-BLED scoring be revised for better risk estimation in patients with intracerebral hemorrhage?

Abstract

Many efforts have been made to develop decision-support tools and bleeding prediction schemes to start or resume anticoagulation after intracerebral hemorrhage, related with anticoagulation use or not, such as CHA₂DS₂-VASc or HAS-BLED scoring. HAS-BLED is a validated scoring system to predict the risk of major bleeding in a patient with atrial fibrillation; some current scientific guidelines suggest its use in ‘risk–benefit’ reasoning when deciding whether to start long-term oral anticoagulation. Here the authors present a patient with atrial fibrillation and intracerebral hemorrhage, and aim to discuss the use of HAS-BLED, suggesting that some revisions may help better management of these patients for major bleeding risk.

Keywords::

• anticoagulation
• bleeding risk
• HAS-BLED scoring
• intracerebral hemorrhage
• revisions

There is an ongoing dilemma on starting or resuming anticoagulation after intracerebral hemorrhage (ICH). The decision of whether and when to start or resume anticoagulation after ICH – related with anticoagulation use or not – should be made on an individual basis after taking into consideration the patient’s risk factors for thromboembolism and his/her preferences after a thorough discussion of the risks versus benefits [1].

HAS-BLED is a validated scoring system to predict the risk of major bleeding in a patient with atrial fibrillation; some current scientific guidelines suggest its use in a ‘risk–benefit’ reasoning (where the cardioembolic risk is estimated using CHADS2 or CHA2DS2-VASc scoring) when deciding to start long-term oral anticoagulation [2,3]. HAS-BLED has been developed and validated in mixed real-world or trial populations in which patients had different bleeding risk factors and could or could not have experienced a prior ICH. Lip [3] has very recently published a comprehensive review and underlined the use of the validated CHA2DS2-VASc and HAS-BLED scores in bleeding risk assessment recently focused on making clinical decisions. However, it has been never specifically tested in a selected anticoagulant-naïve population with a prior ICH or in a selected population with a prior anticoagulant-related ICH. In addition, the predictive ability of HAS-BLED has been always measured looking at major bleedings; its performance in predicting ICH has been tested only once [4]. However, deciding whether to start anticoagulation in a patient with a prior ICH, and, even more, deciding whether to resume anticoagulation in a patient with a prior anticoagulant-related ICH are not uncommon clinical scenarios in which the use of HAS-BLED might not be fully satisfying. In this context, we aim to present a case report to discuss the use of HAS-BLED scoring in a patient with atrial fibrillation and ICH.

Case report

An 80-year-old male patient complaining of sudden-onset difficulty in speech and loss of muscle strength on his right arm and leg was admitted to the stroke unit with the suspicion of stroke. His neurological examination revealed mild motor aphasia and mild right-sided hemiparesis. The mini-mental state examination score was 26 over 30 points. His body mass index was 25.5 kg/m². His past medical history was unremarkable and he had no history of previous stroke or transient ischemic attack, hypertension, diabetes mellitus, cardiac disease, coagulopathy, smoking habits or alcohol intake. He was not on any drugs, and he denied the use of any illicit drugs. His detailed biochemical tests including renal and liver function tests were all within normal limits. Electrocardiography showed persistent atrial fibrillation; echocardiography was otherwise normal.

His cranial magnetic resonance images showed diffuse hyperintensity in bilateral periventricular and deep subcortical white matter, a resorbing chronic hematoma localized in the left basal ganglion and an acute lobar hematoma in the left temporoparietal cortico–subcortical region (Figure 1). Gradient-echo images showed small focal regions of microhemorrhages and suggested cerebral amyloid angiopathy. We considered the patient as high risk for recurrent ICH, and he was put on antiaggregant therapy but not given anticoagulation treatment. He is being followed in our outpatient clinic for more than 3 years now without any recurrent hemorrhagic or ischemic stroke.

Figure 1. Cranial magnetic resonance images showing diffuse hyperintensity in bilateral periventricular and deep subcortical white matter, a resorbing chronic hematoma localized in the left basal ganglion and an acute lobar hematoma in the left temporoparietal cortico–subcortical region.

Discussion

Based on this case example, we analyzed whether HAS-BLED was enough for risk estimation in making the decision to start anticoagulation. This patient has no hypertension (H = 0), normal renal or liver functions (A = 0), no history of ischemic stroke/transient ischemic attack (S = 0) and no labile INR levels (L = 0) or drug/alcohol history (D = 0). On the other hand, he has bleeding (B = 1) and he is older than 65 years (E = 1). In this context, his HAS-BLED score is calculated as 2 points. Moreover, as there is no other history of bleeding disorder in his past medical history, the hemorrhagic nature of stroke may be missed if not asked in detail and lead to lower scores. Under both circumstances, the decision to start oral anticoagulation therapy is possibly made by the physicians. Furthermore, CHA₂DS₂-VASc scoring is calculated as 5 points in this patient [2,5], which indicates an annual risk of stroke as 6.7% and supports the need for oral anticoagulation therapy in secondary prophylaxis.

On the other hand, we believe that some details would be incorporated better into these scoring systems, which may change the final decision of the physicians. First of all, type of stroke may separately be evaluated as ischemic or hemorrhagic stroke. Secondly, intracranial hemorrhage may further be grouped as deep versus lobar in location. On this basis, hemorrhagic stroke and lobar location would be scored differently. This is because oral anticoagulation therapy would result in about 31 fewer thromboembolic strokes per 1000 patients with deep hemorrhages at a cost of 19 additional ICH [6]. For 1000 patients with lobar hemorrhages, however, oral anticoagulation therapy would result in about 31 fewer thromboembolic strokes, but at a cost of 150 additional ICH during the first year of treatment [6]. Furthermore, the presence of white matter abnormalities in a patient with ICH may also be encountered, as hyperintense ischemic changes in white matter and leukoaraiosis are known as risk factors for increased risk of ICH related to oral anticoagulation therapy [7,8]. In addition to leukoaraiosis, the presence of microbleeds was also reported to be associated with ICH recurrences [9]. The presented patient also had diffuse white matter abnormalities as leukoaraiosis and focal regions of microhemorrhages suggestive of cerebral amyloid angiopathy, which should be taken into account while evaluating a patient with atrial fibrillation and a history of ICH to decide to start long-term oral anticoagulation.

On the basis of these data, we considered the presented patient as high risk for recurrent ICH, and he was put on antiaggregant therapy. He was well in the following 3 years without any recurrent hemorrhagic or ischemic stroke. However, it should also be mentioned that there is an ongoing debate around the efficacy and safety of aspirin, especially in terms of ICH, in elderly patients with atrial fibrillation [10]. This further emphasizes the need to better individualize the choice to anticoagulate or not in patients with prior ICH. In addition, resuming anticoagulation in a patient with a prior anticoagulant-related ICH is not an uncommon condition that physicians have great difficulty with, in which the use of HAS-BLED might also not be fully satisfying. As Lip has covered in his review, a more recent SAMe-TT2R2 scoring has also been proposed to identify those patients with atrial fibrillation likely to do well on warfarin or those likely to have poor anticoagulation control, aiding in decision-making of whether a non-vitamin K antagonist oral anticoagulant may be a better option or not [3]; as apparently lower rates of ICH were shown to be associated to these new drugs when compared with warfarin.

Clinicians studying cardio- and cerebrovascular disorders are usually confronted with difficult decisions regarding the risks and benefits of anticoagulation, which engenders the need for developing scoring systems with better risk stratification. We believe that some revisions in terms of stroke type, localization of hemorrhagic stroke and the presence of leukoaraiosis and cerebral microbleeds might help in better management of patients with ICH for major bleeding risk with the aid of the HAS-BLED scoring system.

Acknowledgements

We would like to thank M Marcucci (University of Milan, Italy) for her great and valuable contributions in revising our manuscript.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.