Jan 16, 2017
Dr. Carolyn Lam: Welcome to Circulation on the Run, your weekly podcast summary and backstage pass to the journal and its editors. I'm Dr. Dr. Carolyn Lam, associate editor from the National Heart Center and Duke National University of Singapore.
In today's episode, we are discussing very important new data regarding stroke risk stratification in patients with atrial fibrillation. First though, let me give you the highlights of this week's journal.
The first paper provides mechanistic evidence that endothelial-derived microparticles may play a key role in the development of endothelial dysfunction following acute coronary syndrome. In this paper from first author, Dr. Abbas, co-corresponding authors, Dr. Toti and Morel from the University of Strasbourg in France, authors expose core sign coronary artery endothelial cells to microparticles shed from senescent cells, or circulating microparticles from patients with acute coronary syndrome.
They showed that exposure to these microparticles induced increase senescence-associated beta-galactosidase activity, oxidative stress, and early phosphorylation of MAP kinases and AKT, and upregulation of p53, p21 and p16. Depletion of endothelial-derived microparticles from acute coronary syndrome patients reduced the induction of senescence.
On the other hand, pro-senescent microparticles promoted endothelial cell thrombogenicity. These microparticles exhibited angiotensin-converting enzyme activity and upregulated AT1 receptors and ACE in endothelial cells. Losartan and AT1 receptor antagonist and inhibitors of either MAT kinases or PI3-kinase prevented the microparticle-induced endothelial senescence.
In summary, these findings indicate that endothelial-derived microparticles from acute coronary syndrome patients induce premature endothelial senescence and thrombogenicity suggesting that targeting endothil-derived microparticles and their bioactivity may be a promising therapeutic strategy to limit the development of endothelial dysfunction post-acute coronary syndrome.
The next study is the first large and prospective study showing that NT-proBNP is associated with cardiovascular events in patients with adult congenital heart disease independent of multiple clinical and echocardiographic variables.
This is a study from first author, Dr. Bekan; and corresponding author, Dr. Roos-Hesselink and colleagues from the Erasmus University Medical Center in Rotterdam, the Netherlands. The author studied 595 clinically stable patients with adult congenital heart disease who attended the outpatient clinic between 2011 and 2013.
All patients underwent clinical assessment, electrocardiography, echocardiography and biomarker measurement, including NT-proBNP, high-sensitivity troponin T and growth differentiation factor 15. Patients were prospectively followed over a median of 42 months for the occurrence of cardiovascular events including death, heart failure, hospitalization, arrhythmia, thromboembolic events and reintervention.
They found that of the three evaluated biomarkers, NT-proBNP was most strongly associated with cardiovascular events. Importantly, patients with a low-risk of death and heart failure could be accurately identified with a high negative predictive value.
In patients with elevated NT-proBNP, elevations of high sensitivity troponin T and growth differentiation factor 15 identified those patients at highest risk of cardiovascular events.
In summary, these biomarkers may play an important role in the monitoring and management of patients with adult congenital heart disease.
The next study describes heart failure stages among older adults in the community. Dr. Shah and colleagues from the Brigham and Women's Hospital in Boston Massachusets classified more than 6,000 participants in the atherosclerosis risk and community study into heart failure stages. These were stage A; asymptomatic individuals with heart failure risk factors, but no cardiac structural or functional abnormalities. Stage B; asymptomatic individuals with structural abnormalities such as left ventricle hypertrophy, dilation, dysfunction, or valve disease. Stage C1; clinical heart failure without prior hospitalization. Stage C2; clinical heart failure with prior hospitalization.
They found that only 5% of examined participants were free of heart failure risk factors or structural heart disease. 52% were categorized as stage A, 30% stage B, 7% stage C1, and 6% stage C2. Worst heart failure stage was associated with a greater risk of incident heart failure hospitalization or death at a median follow up of 608 days.
Left ventricular ejection fraction was preserved in 77% of stage C1 and 65% of stage C2 respectively. In corporation of longitudinal strain measurements and diastolic dysfunction into the stage B definition, reclassified 14% of the sample from stage A to B.
Abnormal LV structure, systolic function, whether based on ejection fraction of longitudinal strain, and diastolic dysfunction, were each independently and additively associated with the risk of incident heart failure hospitalization or death in stage A and B participants.
The authors concluded that the majority of older adults in the community are at risk of heart failure, appreciably more compared to previous reports in younger community-based samples. The study also highlighted the burden of heart failure with preserved ejection fraction in the elderly and provided evidence that left ventricular diastolic function and longitudinal strain provide incremental prognostic value beyond conventional measures of LV structure and ejection fraction in identifying patient at risk of heart failure hospitalization or death.
The next study sheds light on the association of the LPA gene, ethnicity and cardiovascular events. First author, Dr. Lee; corresponding author Dr. Tsimikas and colleagues from University of California San Diego studied 1,792 black, 1,030 white, and 597 Hispanic subjects all enrolled in the Dallas Heart Study. They measured LPA snips, apolipoprotein A isoforms, LP(a) and oxidized phospholipids on apolipoprotein B100.
These individuals were also followed for a median of 9.5 years for major adverse cardiovascular events. The authors found that the prevalence of LPA snips and apolipoprotein A isoforms were very different across ethnic groups. LPA snips that were associated with elevated LP(a) in whites were associated with low LP(a) in Hispanics mainly due to differences in apoliproprotein A isoforms size.
After multi-variable adjustment, LP(a) and oxidized phospholipids on apolipoprotein B were both predictors of major adverse cardiovascular events. Conversely, LPA snips and apolipoprotein A isoforms did not add predictive value to models and did not show clinical utility in this study.
These data suggests that much of LP(s) mediated major adverse coronary events is driven by oxidized phospholipids. Importantly, elevated LP(a) and oxidized phospholipids on apolipoprotein B must be recognized as important predictors of major adverse cardiovascular events across racial groups.
The final study addresses the question of the optimal antithrombotic regimen for longterm management of patients with symptomatic peripheral artery disease, or PAD, with a history of limb revascularization. To answer this question, Dr. Jones and colleagues from Duke Clinical Research Institute looked at the EUCLID trial, or examining use of ticagrelor in PAD trial, which randomized patients with PAD to treatment with ticagrelor 90 milligrams twice daily, or clopidogrel 75 milligrams daily.
As a reminder, patients in EUCLID were enrolled based on a normal ankle-brachial index of less .8, or a prior lower extremity revascularization. The current paper really focus on the subset of 7,875 patients who were enrolled based on a prior lower extremity revascularization criterion.
The authors found that after adjustment for baseline characteristics, patients enrolled based on prior revascularization for PAD had higher higher rates of myocardial infarction and acute limb ischemia with similar composite rates of cardiovascular death, myocardial infarction and stroke when compared with patients enrolled based on the ankle brachial index criterion.
Overall, there were no significant differences between ticagrelor and clopidogrel for the reduction of cardiovascular or acute limb events.
Those were your highlights. Now, for our featured discussion.
On today's podcast, we are discussing the very, very important issue of stroke risk in patients with atrial fibrillation. Most of us use the international guidelines for anticoagulation in atrial fibrillation that mostly suggest that we use the CHADS VASc scoring system to determine the stroke risk in a particular patient and then determine whether or not this patient meets the threshold for anticoagulation.
This assumes that the CHADS VASc score corresponds to a fixed stroke rate. Today, in our journal, we have very, very interesting results from a paper with corresponding author, Dr. Daniel Singer who really suggest that we may need to rethink that. Dr. Daniel Singer joins us today from Massachusets General Hospital.
Dr. Daniel Singer: Thank you for having me.
Dr. Carolyn Lam: Great. Today, we also have Dr. Sana Al-Khatib who's the associate editor from Duke University who managed this paper. Welcome Sana.
Dr. Sana Al-Khatib : Thank you Dr. Carolyn, I'm happy to be here.
Dr. Carolyn Lam: Daniel, could we start by you letting us know what you sought to do in your study and what you found?
Dr. Daniel Singer: We all know that anticoagulants are extraordinarily effective at preventing stroke in patients with atrial fibrillation, but they also raise the risk of bleeding, and sometimes that bleeding could be quite serious and even fatal. As a result, for that past 10, 15 years, we have used a risk-based approach to the decision about whether to start a patient on anticoagulation, and that risk is the stroke risk that a patient faces if they weren't taking anticoagulants. Then we figured that anticoagulants will reduce it by about two-thirds.
There are formal decision analysis and then a more informal sense that a patient has to face an anticoagulated risk of stroke of about 2%, some people might say 1% to 2% before anticoagulation results in an expected net clinical benefit that the effect in reducing ischemic stroke will exceed the risk of increasing bleeding.
While the CHADs VASc score has been widely accepted as the basis for estimating that risk, it became apparent to us as we looked across the studies that were underlying that assumption, that the risk that were associated with various CHADs VASc scores were extremely variable. Many of these risks actually were less than that 1% or 2% threshold for anticoagulation.
What I mean is that the stroke risk associated with CHADs VASc score of one, or two, which is the basis for the guideline threshold for anticoagulation actually corresponded to risk less than 1% in many of these very large studies. We have conducted a systematic review just to be sure that we were capturing the overall evidence base for this, and that's what we report in our paper.
Dr. Carolyn Lam: Perhaps you could start by letting us know exactly how far off are we in our stroke risk estimation.
Dr. Daniel Singer: We looked at 34 studies that were quite large and then we zeroed in on the largest one. If you looked at the rate for stroke overall, they varied enormously in terms of the overall stroke rate. Then when we focused down on CHADs VASc score of 1, or 2, we found that the majority of these studies, actually, for CHADs VASc 1, was less than 1% per year. For CHADs VASc 2 score was in the majority these studies less than 2% per year.
Both of those stroke risks have raised us the question where are these patients could gain in that clinical benefit from being anti-coagulated, because those stroke risk, if they were reduced by two-thirds, would really be a very small reduction in risk and yet they'd still face the bleeding risk.
Among the most interesting findings actually is that we found that a Swedish national database and the large Danish national database came up with threefold difference in their estimate of stroke rates. The Swedish database produced lower risk, and the Danish database produced substantially higher risk.
If you think about it, there are probably no two countries in the world that are more similar in terms of gene, social environmental, medical care systems, and that raises the specific question of, "Is it underlying rates that vary across different cohorts and different geographies, or is it a different in methodology?"
We think a lot of the differences are due to methodologic difference, and that we need to standardize these differences together, better handle on what the real stroke rate is among patients with these low CHADs VASc scores.
Dr. Carolyn Lam: The variability that you pointed on your paper is really striking, but another possibility, do you think, is that maybe stroke risk isn't static.
Dr. Daniel Singer: Yeah. If that's the case, we face a great difficulty in developing predictions rules of what the stroke risk could be. I think most people feel it's the function of their age, and whether they've had a prior stroke, and whether they have the comorbidity, hypertension, and diabetes, and so on, that are incorporated into the various stroke risk scores, in particular, CHADs VASc.
We tend to think that that's pretty fixed until you get older or until you accumulated another comorbidity. I think the striking difference is that, one, that we actually anticipated in the beginning, was that the stroke rates in people with atrial fibrillation were also coming down. The stroke rates in general have been dramatically decreasing for decades now.
One issue is whether that applies as well to atrial fibrillation associated stroke. There is a suggestion of that, but the variability across the cohorts is so great that you can't pick up a strong signal in terms of calendar time. Although I suspect that there is a strong calendar effect. Exactly why that is, we could speculate. I suspect a lot of it is control of blood pressure, but that's speculation.
Dr. Carolyn Lam: Daniel, congratulations again for that fascinating and really very sobering findings.
Sana, you managed this paper. It's very important paper. In fact, important enough that you invited an editorial. Could you please share some of your thoughts?
Dr. Sana Al-Khatib : Oh, yeah. Absolutely. First, I'd like to start by congratulating Daniel and his team on conducting this really important study. I enjoyed reading it and managing it. Definitely, congratulations.
A couple of thoughts that I have. I completely agree with this really important finding, that there is a lot of variability in the rates of stroke that come from different patient populations and databases. As you pointed out Daniel, I think this is indeed largely due to differences in methodology in terms of how the information was selected, how certain things were defined.
I agree with you there. You called for standardization of this, and I wonder if you have any thoughts about how we can go about doing that. I also want to bring up some of the newer studies now that are showing some significance in terms of biomarkers. Is that really adding significantly to the predictive ability of risk prediction models? I wanted to get your thoughts on that as well.
Dr. Daniel Singer: Let me address your last question, which is simply you state that the CHADs VASc score, the CHAD score and so on, are based on very simple clinical features, and it would be unusual for them to be highly predictive. In fact, they're only mediocrely predictive, and the addition of biomarkers high-sensitivity troponin proBNP, now, people have suggested the imaging biomarkers like magnetic resonance to asses fibrosis in the left atrium. These are all very, very promising in terms of getting better models.
The problem is to do that on a very scale such that we can get precise and well-calibrated predictions. We've found when we're analyzing to pair risk scores, we found that the most important issue is the underlying risk, so that, yes, you can get a great model, but if you have high variability in the underlying rate, you can have a problem specifying an individual with a stroke risk.
We have to standardize and improve the quality of bringing people into these cohorts, and of interrogating the cohorts and databases and making sure that we have the same approach to assessing outcomes.
This could probably be best done in very big scientific prospective registry studies, but it's tough to get all that information. There are some registry studies now ongoing, the ORBIT registries, the GARFIELD registries that may help us a lot with specifying stroke risk, but they don't have the biomarkers embedded in them. I'm hopeful that with better message, and large studies, and incorporating biomarkers, that we'll really get down to very accurate and generalizable stroke risk.
I think the CHADs VASc and similar simple stroke risk scores will be in the rear-view mirror.
Dr. Sana Al-Khatib : That's great. Can I ask one other question, because I completely agree with you looking at your numbers and the data that you presented, is that when you look, especially at the CHADs VASc score 1 patient, the risk seems to be pretty low.
As you very well know, the guideline documents don't really ... At least, for the American AHA/ACC guideline document, they don't really verbalize very definitively the need to anticoagulate patients with a CHADs VASc score of 1.
If you look at the numbers related to a CHADs VASc score of 2, I'm not sure that I completely agree that the risk is very low. Certainly, there was 33% of the studies reported stroke rates of greater than 2% per year. I think maybe different people have different thresholds. While I completely agree with you on the CHADs VASc score of 1 patients, I find that the findings on patients with a CHADs VASc score of 2 a bit more concerning.
In fact, if anything, I would want based even on your data, not on the guidelines to offer anticoagulation to patients with CHADs VASc score of 2. What would you say to that?
Dr. Daniel Singer: I'm looking at our table that has this, and a lot of the CHADs VASc 2 scores are under 2%, but they're in mid 1%. In the North American cohorts in particular, the rates tend to be lower. That said, I think the heart of the problem here is that we have focused on the threshold for anticoagulation. I think there's an argument to be made that you lay out the risks and benefits to the patients and engage them in a decision, particularly with regards to these lower CHADs VASc scores.
At least you make a lot of, perhaps, even more emphasis on being sure that the higher CHADs VASc scores, that anticoagulation is the net benefits of anticoagulation are made very clear to the patient, and that we don't have large fractions of patients who can take anticoagulants not taking them.
We know from the pinnacle registry and other registries, that even at high CHADs VASc scores, we have 40% plus of atrial fibrillation patients who are not getting anticoagulants. I think that's where we have a lot more assurance that the net benefit is positive and that we can make a different both in terms of a patient in front of us, and in terms of the overall public health aspects of atrial fibrillation and stroke.
Dr. Sana Al-Khatib : I do believe that this is really important, but it is also important to keep in mind that with the novel novel oral anticoagulants, I think the whole landscape has changed. Not only do patients have different options to consider, but certainly, the risk of bleeding, which is the other part of this equation, has gone down significantly with the novel agents.
I think as we engage in shared decision making with patients, I think it is really important to highlight these really very remarkable features about the agents that have really changed the care of patients with atrial fibrillation.
One thing to add to this whole topic is, really, all the new advances that we're seeing in this field that has been really life-changing for us and for our patients.
Dr. Carolyn Lam: Indeed Sana. I was about to bring up the bleeding risk part, the flip side of the coin as well. Also, the point that most of my patients with atrial fibrillation, they really strongly value the avoidance of stroke even more than avoidance of bleeding. Someone, that needs to be taken into consideration as well.
Daniel, I'd love to give you the last words. You mentioned that you like to highlight, maybe, some more of the implications of your findings.
Dr. Daniel Singer: I guess I would say there's a scientific implication, which is what we've ben discussing, which is the importance of trying to get these rates down correctly and accurately, and maybe we have to get people together to say how they're doing these studies.
The second is, for the individual patient, that we should engage them in this discussion. Maybe patients who are perfectly willing to a novel anticoagulant and CHADs VASc score of zero. That would come out of a discussion with the patient. That our emphasis at this point since we're a little unsure about the threshold level, our emphasis both at the individual patient level, and then from the public heath perspective should be on the higher CHADs VASc scores where we know that we can expect a net clinical benefit from the vast majority of patients with AF.
I agree with Dr. Al-Khatib, that the novel anticoagulants post an important advantage in the sense not so much in their overall bleeding, but particularly in terms of their intercranial bleeding, which is the lethal bleeding we most want to avoid.
Dr. Carolyn Lam: Thank you both for joining us. Thank you listeners for joining us. Don't forget to tune in next week.