New tools for stroke prediction in atrial fibrillation

SNOWMASS, COLO. – High-sensitivity troponin T and brain natriuretic peptide levels are better predictors of stroke risk in patients with atrial fibrillation than the CHA₂DS₂-VASc score that will replace the CHADS₂ score in the forthcoming revised American College of Cardiology/American Heart Association guidelines.

Recent evidence indicates the biomarkers may be novel tools for improved stroke prediction in atrial fibrillation (AF), with prognostic value above and beyond that provided by the CHA2DS2-VASc scores.

These findings raise important unanswered questions about the relationship between AF and stroke. Conventional wisdom has held that left atrial thrombus is the cause of most strokes in patients with AF. But it’s not that simple, Dr. Bernard J. Gersh asserted at the Annual Cardiovascular Conference at Snowmass.

"What are we measuring with these biomarkers? This is what we really don’t understand. What has high-sensitivity troponin T got to do with left atrial thrombus?" asked Dr. Gersh, professor of medicine at the Mayo Clinic in Rochester, Minn.

It seems increasingly clear that it’s not just the atrial arrhythmia that’s important in stroke risk, it’s also the company AF keeps. In a substantial but still uncertain proportion of patients, AF is a marker of vascular disease burden expressed through atrial and vascular endothelial dysfunction, vascular inflammation, left atrial dilatation and fibrosis, and a hypercoagulable state, the cardiologist continued.

He was a coinvestigator on a couple of recent groundbreaking studies that show the prognostic power of biomarkers in predicting both stroke risk and cardiac death in AF patients.

In one report, the investigators looked at baseline high-sensitivity troponin T (hsTnT) levels, clinical risk factors for stroke, and CHA₂DS₂-VASc scores in 12,892 patients with AF who were randomized to apixaban or warfarin in the prospective, double-blind ARISTOTLE (Apixaban for the Prevention of Stroke in Subjects With Atrial Fibrillation) trial (N. Engl. J. Med. 2011;365:981-92). During a median 1.9 years of follow-up, patients in the highest quartile for baseline hsTnT had roughly a twofold greater risk of stroke or systemic embolism than did those in the lowest quartile.

Moreover, patients with a low-risk CHA₂DS₂-VASc score of 0 or 1 but in the top quartile for hsTnT, with a level in excess of 13 ng/L, had a very substantial stroke rate of 2.7% per year despite anticoagulation with apixaban (Eliquis) or warfarin. The relationship was even stronger for cardiac death, where subjects with a low-risk CHA₂DS₂-VASc score who were in the top quartile for hsTnT had a 6% annual risk. A higher baseline hsTnT was also independently associated with sharply increased risk of major bleeding in a multivariate regression analysis (J. Am. Coll. Cardiol. 2014;63:52-61).

In another recently published study, he and his international coworkers showed in ARISTOTLE participants with baseline N-terminal pro-B-type natriuretic peptide (NT-proBNP) levels that this biomarker also improved stroke prediction in AF, providing added value to CHA₂DS₂-VASc scores. Subjects in the top quartile for baseline NT-proBNP had an adjusted 2.35-fold greater risk of stroke or systemic embolism than those in the lowest quartile, irrespective of CHA₂DS₂-VASc score. They also had a 2.5-fold greater risk of cardiac death (J. Am. Coll. Cardiol. 2013;61:2274-84).

A study Dr. Gersh highlighted as "extremely interesting" involved the use of brain natriuretic peptide (BNP) as a marker to rule out delayed AF in stroke patients. The study, led by investigators at the University Hospital Center of Nice (France) and known as TARGET-AF, included 300 consecutive acute stroke patients with no history of AF and no AF on their baseline ECG. During a median 6.8 days of in-hospital Holter monitoring, 17% of the stroke patients developed newly diagnosed AF.

The strongest predictor of delayed AF was baseline plasma BNP. It outperformed the CHA₂DS₂-VASc score and all the other parameters examined, including anterior circulation location of the stroke, P-wave initial force, gender, National Institutes of Health Stroke Scale score, age, left atrial dilatation, and Score for the Targeting of AF (STAF) score. A BNP level greater than 131 pg/mL had a 98.1% sensitivity, 71.4% specificity, and 99.4% negative predictive value for delayed AF.

"Our data indicate that a BNP level of 131 pg/mL or less might rule out delayed AF in stroke survivors and could be included in algorithms for AF detection," the French investigators concluded (J. Stroke Cerebrovasc. Dis. 2013;22:e103-10).

There is plenty of direct evidence from transesophageal echocardiography studies and other sources that a substantial proportion of thromboemboli are directly the result of AF. However, indirect evidence points to additional causal factors. For example, there is a high incidence of thromboembolic events in AF patients without left atrial appendage thrombus. Plus, in natural history studies patients with AF without additional risk factors have a low incidence of stroke. And CHADS₂ and CHA₂DS₂-VASc scores predict vascular events but don’t correlate with left atrial appendage thrombus, Dr. Gersh noted.

He said the CHA₂DS₂-VASc score is clearly an improvement over CHADS2, and its adoption in the forthcoming ACC/AHA guidelines is to be welcomed. The CHA₂DS₂-VASc score increases the number of patients considered at significant risk of stroke and therefore warranting anticoagulation. For example, in a large Danish registry of nearly 48,000 AF patients with a CHADS₂ score of 0-1 not on anticoagulation, patients with a CHADS₂ score of 1 but a CHA₂DS₂-VASc score of 2 had twice the stroke risk of patients with a CHA₂DS₂-VASc of 1 (Thromb. Haemost. 2012;107:1172-9).

That being said, neither risk score is all that impressive. The C-statistic, a measure of a test’s predictive power, is 0.56 for CHADS₂ and it was 0.62 for CHA₂DS₂-VASc in the ARISTOTLE analysis. To put those figures in perspective, a coin toss has a C-statistic of 0.50.

"The individual predictive values are not good. We use CHADS₂ and CHA₂DS₂-VASc in practice and in the guidelines, but we should not pretend they are highly predictive. We need new risk stratification schemes," according to Dr. Gersh.

He reported serving as an adviser to Boston Scientific and St. Jude Medical.

[email protected]

SNOWMASS, COLO. – High-sensitivity troponin T and brain natriuretic peptide levels are better predictors of stroke risk in patients with atrial fibrillation than the CHA₂DS₂-VASc score that will replace the CHADS₂ score in the forthcoming revised American College of Cardiology/American Heart Association guidelines.

Recent evidence indicates the biomarkers may be novel tools for improved stroke prediction in atrial fibrillation (AF), with prognostic value above and beyond that provided by the CHA2DS2-VASc scores.

These findings raise important unanswered questions about the relationship between AF and stroke. Conventional wisdom has held that left atrial thrombus is the cause of most strokes in patients with AF. But it’s not that simple, Dr. Bernard J. Gersh asserted at the Annual Cardiovascular Conference at Snowmass.

"What are we measuring with these biomarkers? This is what we really don’t understand. What has high-sensitivity troponin T got to do with left atrial thrombus?" asked Dr. Gersh, professor of medicine at the Mayo Clinic in Rochester, Minn.

It seems increasingly clear that it’s not just the atrial arrhythmia that’s important in stroke risk, it’s also the company AF keeps. In a substantial but still uncertain proportion of patients, AF is a marker of vascular disease burden expressed through atrial and vascular endothelial dysfunction, vascular inflammation, left atrial dilatation and fibrosis, and a hypercoagulable state, the cardiologist continued.

He was a coinvestigator on a couple of recent groundbreaking studies that show the prognostic power of biomarkers in predicting both stroke risk and cardiac death in AF patients.

In one report, the investigators looked at baseline high-sensitivity troponin T (hsTnT) levels, clinical risk factors for stroke, and CHA₂DS₂-VASc scores in 12,892 patients with AF who were randomized to apixaban or warfarin in the prospective, double-blind ARISTOTLE (Apixaban for the Prevention of Stroke in Subjects With Atrial Fibrillation) trial (N. Engl. J. Med. 2011;365:981-92). During a median 1.9 years of follow-up, patients in the highest quartile for baseline hsTnT had roughly a twofold greater risk of stroke or systemic embolism than did those in the lowest quartile.

Moreover, patients with a low-risk CHA₂DS₂-VASc score of 0 or 1 but in the top quartile for hsTnT, with a level in excess of 13 ng/L, had a very substantial stroke rate of 2.7% per year despite anticoagulation with apixaban (Eliquis) or warfarin. The relationship was even stronger for cardiac death, where subjects with a low-risk CHA₂DS₂-VASc score who were in the top quartile for hsTnT had a 6% annual risk. A higher baseline hsTnT was also independently associated with sharply increased risk of major bleeding in a multivariate regression analysis (J. Am. Coll. Cardiol. 2014;63:52-61).

In another recently published study, he and his international coworkers showed in ARISTOTLE participants with baseline N-terminal pro-B-type natriuretic peptide (NT-proBNP) levels that this biomarker also improved stroke prediction in AF, providing added value to CHA₂DS₂-VASc scores. Subjects in the top quartile for baseline NT-proBNP had an adjusted 2.35-fold greater risk of stroke or systemic embolism than those in the lowest quartile, irrespective of CHA₂DS₂-VASc score. They also had a 2.5-fold greater risk of cardiac death (J. Am. Coll. Cardiol. 2013;61:2274-84).

A study Dr. Gersh highlighted as "extremely interesting" involved the use of brain natriuretic peptide (BNP) as a marker to rule out delayed AF in stroke patients. The study, led by investigators at the University Hospital Center of Nice (France) and known as TARGET-AF, included 300 consecutive acute stroke patients with no history of AF and no AF on their baseline ECG. During a median 6.8 days of in-hospital Holter monitoring, 17% of the stroke patients developed newly diagnosed AF.

The strongest predictor of delayed AF was baseline plasma BNP. It outperformed the CHA₂DS₂-VASc score and all the other parameters examined, including anterior circulation location of the stroke, P-wave initial force, gender, National Institutes of Health Stroke Scale score, age, left atrial dilatation, and Score for the Targeting of AF (STAF) score. A BNP level greater than 131 pg/mL had a 98.1% sensitivity, 71.4% specificity, and 99.4% negative predictive value for delayed AF.

"Our data indicate that a BNP level of 131 pg/mL or less might rule out delayed AF in stroke survivors and could be included in algorithms for AF detection," the French investigators concluded (J. Stroke Cerebrovasc. Dis. 2013;22:e103-10).

There is plenty of direct evidence from transesophageal echocardiography studies and other sources that a substantial proportion of thromboemboli are directly the result of AF. However, indirect evidence points to additional causal factors. For example, there is a high incidence of thromboembolic events in AF patients without left atrial appendage thrombus. Plus, in natural history studies patients with AF without additional risk factors have a low incidence of stroke. And CHADS₂ and CHA₂DS₂-VASc scores predict vascular events but don’t correlate with left atrial appendage thrombus, Dr. Gersh noted.

He said the CHA₂DS₂-VASc score is clearly an improvement over CHADS2, and its adoption in the forthcoming ACC/AHA guidelines is to be welcomed. The CHA₂DS₂-VASc score increases the number of patients considered at significant risk of stroke and therefore warranting anticoagulation. For example, in a large Danish registry of nearly 48,000 AF patients with a CHADS₂ score of 0-1 not on anticoagulation, patients with a CHADS₂ score of 1 but a CHA₂DS₂-VASc score of 2 had twice the stroke risk of patients with a CHA₂DS₂-VASc of 1 (Thromb. Haemost. 2012;107:1172-9).

That being said, neither risk score is all that impressive. The C-statistic, a measure of a test’s predictive power, is 0.56 for CHADS₂ and it was 0.62 for CHA₂DS₂-VASc in the ARISTOTLE analysis. To put those figures in perspective, a coin toss has a C-statistic of 0.50.

"The individual predictive values are not good. We use CHADS₂ and CHA₂DS₂-VASc in practice and in the guidelines, but we should not pretend they are highly predictive. We need new risk stratification schemes," according to Dr. Gersh.

He reported serving as an adviser to Boston Scientific and St. Jude Medical.

[email protected]

SNOWMASS, COLO. – High-sensitivity troponin T and brain natriuretic peptide levels are better predictors of stroke risk in patients with atrial fibrillation than the CHA₂DS₂-VASc score that will replace the CHADS₂ score in the forthcoming revised American College of Cardiology/American Heart Association guidelines.

Recent evidence indicates the biomarkers may be novel tools for improved stroke prediction in atrial fibrillation (AF), with prognostic value above and beyond that provided by the CHA2DS2-VASc scores.

These findings raise important unanswered questions about the relationship between AF and stroke. Conventional wisdom has held that left atrial thrombus is the cause of most strokes in patients with AF. But it’s not that simple, Dr. Bernard J. Gersh asserted at the Annual Cardiovascular Conference at Snowmass.

"What are we measuring with these biomarkers? This is what we really don’t understand. What has high-sensitivity troponin T got to do with left atrial thrombus?" asked Dr. Gersh, professor of medicine at the Mayo Clinic in Rochester, Minn.

It seems increasingly clear that it’s not just the atrial arrhythmia that’s important in stroke risk, it’s also the company AF keeps. In a substantial but still uncertain proportion of patients, AF is a marker of vascular disease burden expressed through atrial and vascular endothelial dysfunction, vascular inflammation, left atrial dilatation and fibrosis, and a hypercoagulable state, the cardiologist continued.

He was a coinvestigator on a couple of recent groundbreaking studies that show the prognostic power of biomarkers in predicting both stroke risk and cardiac death in AF patients.

In one report, the investigators looked at baseline high-sensitivity troponin T (hsTnT) levels, clinical risk factors for stroke, and CHA₂DS₂-VASc scores in 12,892 patients with AF who were randomized to apixaban or warfarin in the prospective, double-blind ARISTOTLE (Apixaban for the Prevention of Stroke in Subjects With Atrial Fibrillation) trial (N. Engl. J. Med. 2011;365:981-92). During a median 1.9 years of follow-up, patients in the highest quartile for baseline hsTnT had roughly a twofold greater risk of stroke or systemic embolism than did those in the lowest quartile.

Moreover, patients with a low-risk CHA₂DS₂-VASc score of 0 or 1 but in the top quartile for hsTnT, with a level in excess of 13 ng/L, had a very substantial stroke rate of 2.7% per year despite anticoagulation with apixaban (Eliquis) or warfarin. The relationship was even stronger for cardiac death, where subjects with a low-risk CHA₂DS₂-VASc score who were in the top quartile for hsTnT had a 6% annual risk. A higher baseline hsTnT was also independently associated with sharply increased risk of major bleeding in a multivariate regression analysis (J. Am. Coll. Cardiol. 2014;63:52-61).

In another recently published study, he and his international coworkers showed in ARISTOTLE participants with baseline N-terminal pro-B-type natriuretic peptide (NT-proBNP) levels that this biomarker also improved stroke prediction in AF, providing added value to CHA₂DS₂-VASc scores. Subjects in the top quartile for baseline NT-proBNP had an adjusted 2.35-fold greater risk of stroke or systemic embolism than those in the lowest quartile, irrespective of CHA₂DS₂-VASc score. They also had a 2.5-fold greater risk of cardiac death (J. Am. Coll. Cardiol. 2013;61:2274-84).

A study Dr. Gersh highlighted as "extremely interesting" involved the use of brain natriuretic peptide (BNP) as a marker to rule out delayed AF in stroke patients. The study, led by investigators at the University Hospital Center of Nice (France) and known as TARGET-AF, included 300 consecutive acute stroke patients with no history of AF and no AF on their baseline ECG. During a median 6.8 days of in-hospital Holter monitoring, 17% of the stroke patients developed newly diagnosed AF.

The strongest predictor of delayed AF was baseline plasma BNP. It outperformed the CHA₂DS₂-VASc score and all the other parameters examined, including anterior circulation location of the stroke, P-wave initial force, gender, National Institutes of Health Stroke Scale score, age, left atrial dilatation, and Score for the Targeting of AF (STAF) score. A BNP level greater than 131 pg/mL had a 98.1% sensitivity, 71.4% specificity, and 99.4% negative predictive value for delayed AF.

"Our data indicate that a BNP level of 131 pg/mL or less might rule out delayed AF in stroke survivors and could be included in algorithms for AF detection," the French investigators concluded (J. Stroke Cerebrovasc. Dis. 2013;22:e103-10).

There is plenty of direct evidence from transesophageal echocardiography studies and other sources that a substantial proportion of thromboemboli are directly the result of AF. However, indirect evidence points to additional causal factors. For example, there is a high incidence of thromboembolic events in AF patients without left atrial appendage thrombus. Plus, in natural history studies patients with AF without additional risk factors have a low incidence of stroke. And CHADS₂ and CHA₂DS₂-VASc scores predict vascular events but don’t correlate with left atrial appendage thrombus, Dr. Gersh noted.

He said the CHA₂DS₂-VASc score is clearly an improvement over CHADS2, and its adoption in the forthcoming ACC/AHA guidelines is to be welcomed. The CHA₂DS₂-VASc score increases the number of patients considered at significant risk of stroke and therefore warranting anticoagulation. For example, in a large Danish registry of nearly 48,000 AF patients with a CHADS₂ score of 0-1 not on anticoagulation, patients with a CHADS₂ score of 1 but a CHA₂DS₂-VASc score of 2 had twice the stroke risk of patients with a CHA₂DS₂-VASc of 1 (Thromb. Haemost. 2012;107:1172-9).

That being said, neither risk score is all that impressive. The C-statistic, a measure of a test’s predictive power, is 0.56 for CHADS₂ and it was 0.62 for CHA₂DS₂-VASc in the ARISTOTLE analysis. To put those figures in perspective, a coin toss has a C-statistic of 0.50.

"The individual predictive values are not good. We use CHADS₂ and CHA₂DS₂-VASc in practice and in the guidelines, but we should not pretend they are highly predictive. We need new risk stratification schemes," according to Dr. Gersh.

He reported serving as an adviser to Boston Scientific and St. Jude Medical.

[email protected]