Preview Mode Links will not work in preview mode

Circulation on the Run


Feb 25, 2019

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 Carolyn Lam, associate editor from the National Heart Center and Duke National University of Singapore.

Dr Greg Hundley:             And I'm Greg Hundley, associate editor and director of the Pauley Heart Center from VCU Health in Richmond, Virginia.

Dr Carolyn Lam:                So, Greg, are we any closer to the holy grail of safe ED discharge for acute heart failure based on a risk score? Well, we're going to be discussing that coming right up after Greg and I share about the papers that we'd like to discuss today. Lovely issue, isn't it?

Dr Greg Hundley:             Yup, and time to get your coffee and bring it up. My first paper, Carolyn, is from Michael Chu from London Health Sciences Center, and is really investigating the surgical management of thoracic aortic disease, and looking at the impact of gender or sex related differences. Sex related differences have not been thoroughly studied. This group looked at a total of 1653 patients, 30% were women, who underwent thoracic aortic surgery with hypothermic circulatory arrest between the years of 2002 and 2017 across Canada in 10 institutions.

                                                Well, women underwent less aortic root reconstruction, including aortic root replacement, Ross procedures, or valve sparing root operations. But, even with less invasive, the women experienced higher rates of mortality, 11% versus 7%, stroke, and that composite of the thoracic surgeons' adverse events. On multi variable analysis, female sex or women was an independent predictor of overall mortality, stroke, and those comorbidities.

Dr Carolyn Lam:                Greg, you know how much I love these papers, so I'm going to repeat that. You're saying the women received less ominous procedures and yet had worse outcomes, and this was independent of the clinical covariances, right?

Dr Greg Hundley:             Absolutely. Putting all this together, women underwent thoracic aortic surgery a little bit older, and with larger index aortic aneurysm size than men. Intraoperatively, women undergo fewer concomitant procedures, such as the aortic root repairs, and things that you just mentioned. But nevertheless, women experience significantly worse outcomes identified as an independent predictor of mortality, stroke, and that composite endpoint for mortality, morbidity, after multi variable analysis.

                                                What should we think about this? Well, sex specific considerations are important when considering thoracic aortic surgery and future research should focus on the development of a personalized approach to thoracic aortic surgery with respect to gender. For example, utilization of maybe lower size thresholds for women for aortic aneurysms should be considered, and for earlier interventions, and improved outcomes.

                                                Carolyn, tell me about one of your papers.

Dr Carolyn Lam:                All right, so I chose a paper looking at the stroke outcomes in the COMPASS trial. Now, let's remind everybody that the COMPASS trial, where patients with stable coronary artery disease or peripheral artery disease, and randomly assigned to receive aspirin 100 milligrams daily, rivaroxaban five milligrams twice daily, or the combination of rivaroxaban 2.5 milligrams twice daily plus aspirin. Patients requiring anticoagulation with a stroke within a month had a previous lacunar stroke or intracerebral hemorrhage were excluded.

                                                Now, in the current paper, and this is from Dr Sharma from Population Health Research Institute, and their colleagues, basically they looked at a detailed analysis of the stroke by type, predictors, and anti-thrombotic effects in the key subgroups. They found that the combination of low dose rivaroxaban and aspirin prevented stroke and disabling stroke better than aspirin in patients without atrial fibrillation and with stable vascular disease, and without an increasing risk of hemorrhagic stroke; which is really important. This effect was consistent across subgroups of baseline risk, and particularly marked in those with a history of previous stroke.

Dr Greg Hundley:             Carolyn, what about that rivaroxaban five milligrams twice daily alone?

Dr Carolyn Lam:                There was no significant difference in the occurrence of stroke in the rivaroxaban alone group compared with aspirin. But all of this simply says perhaps low dose rivaroxaban and aspirin may be a really important new anti-thrombotic option for primary and secondary stroke prevention in patients with clinical stable atherosclerosis.

Dr Greg Hundley:             Very interesting. I'm going to follow your lead and go into another sort of anticoagulant-related topic on iliofemoral deep vein thrombosis. This paper is by Suresh Vedantham from the Washington University of St. Louis.

                                                Let's talk about just what is the definition? This is a DVT that involves the iliac and/or the common femoral vein with or without involvement of additional veins. It basically obstructs the outflow of the veins. These patients are phenotypically distinct from patients with cath or femoral popliteal DVT because that totally obstructs flow, and they have more frequent recurrence of venous thromboembolic events, and more frequent post-thrombotic syndrome. Well, that's a horrible condition because of that obstruction, it leads to calf muscle dysfunction, edema, subcutaneous fibrosis, tissue hypoxia, and ulceration.

Dr Carolyn Lam:                Great background. What did this study show?

Dr Greg Hundley:             This is a sub-study of the ATTRACT trial. The ATTRACT trial basically is looking at anticoagulation plus perhaps mechanical intervention, or direct catheter directed thrombolysis therapy versus just anticoagulation alone. This sub-study is 391 patients with acute DVT involving just the iliac or the common femoral veins, and following these individuals for 24 months to compare short and long-term outcomes.

                                                What did the study show? Well, this interventional group did have a reduction in leg pain and swelling, and improvement in quality of life related to that lower extremity. But, no overall difference in overall quality of life, and very importantly, no difference in the occurrence of this post thrombotic syndrome.

Dr Carolyn Lam:                That's kind of disappointing. I understand that the ATTRACT study is not the first to look at this, though. That was in an editorial discussing this. Could you tell us about that?

Dr Greg Hundley:             Yeah, Carolyn. Jay Giri from University of Pennsylvania just had an incredible editorial. I think if you have an opportunity, listeners, to take a look at that, I highly recommend it. He reminded us of the CaVenT trial, which is basically performed as an open label randomized control trial of 209 patients across 20 hospitals in Norway.

                                                What was different in the CaVenT trial is that at 24 months of follow up, the intervention with thrombolysis and systemic anticoagulation improved iliofemoral patency. It reduced the incidence of this post thrombotic syndrome. In ATTRACT, in this sub-study, it was intravenous thrombolysis, systemic anticoagulation, and mechanical intervention on the vein versus in the other study from Norway, CaVenT, just the inter vein thrombolysis and the systemic anticoagulation.

                                                What Dr Giri points out is that maybe something related to intervention in that vein when you're stripping out thrombus, et cetera, are we damaging the veins in the vessel that prevents reflux, et cetera?

                                                I think really moving forward, you're going to have to personalize this decision in individual patients until we have more data on this subject.

Dr Carolyn Lam:                Great learning. I learned a lot from this next paper, too, because I actually chose a basic science paper. This is a paper that uncovers a new fine tuning factor that modulates myocardial infarction induced inflammation. That is a small GTPase called RhoE.

                                                In this study, Drs Chang from Texas A&M University College of Medicine, and Song from Fuwai Hospital in Beijing used three genetic mouse model lines. Those are the global knockout, the cardiomyocyte specific RhoE heterozygous mouse, and the cardiomyocyte specific RhoE over expression mouse. With this combination, they showed that RhoE deficiency causes excessive inflammatory response in infarct animal heart, resulting in enlarged heart, decreased contractility, and increased mortality. The mechanism is that RhoE binds to P65 and P50, which impedes their dimerization and blocks these two proteins from nuclear translocation. Now, over expression of cardiac RhoE inhibits NF-κB, restrains post MI inflammation, and improves cardiac function and survival.

                                                Importantly as you always say, Greg, there is human data. They found that the expression of RhoE was elevated in the infarct patient heart and that patients with a higher expression of RhoE exhibited a better prognosis and better cardiac function recovery.

Dr Greg Hundley:             Carolyn, tell me a little bit about the clinical significance of this.

Dr Carolyn Lam:                You just wanted to ask me a tough question. I can see it on your face. Basically, I think this is really exciting because RhoE may serve as a new potential biomarker for the assessment of myocardial infarction in patients, and manipulation of RhoE could be a potential therapeutic approach for MI. There.

Dr Greg Hundley:             Very good.

Dr Carolyn Lam:                That's all the time we have for our little discussion here. Now, let's go onto the feature paper. ...

                                                Over 80% of emergency department patients with acute heart failure are admitted to the hospital. Now, contrast this with the fact that over 80% of all emergency department visits result in discharge. So, why is that many other emergency department based cardiovascular disease processes like for acute coronary syndrome have evolved from high rates of admission to timely and safe discharge whereas decision making in acute heart failure has not experienced a similar evolution. Do we need perhaps a better acute heart failure prognostic score that's validated?

                                                Well, guess what? We're going to talk about this right now in our feature discussion, and a beautiful feature paper that we're so proud to have the corresponding author, Dr Douglas Lee from University of Toronto right here to discuss; along with the managing editor, Dr Justin Ezekowitz, who's associate editor from University of Alberta, and the editorialist, Dr Sean Collins from Vanderbilt University Medical Center. Welcome everyone, and Doug, please, could you just start by telling us about this great paper?

Dr Douglas Lee:                 We validated, and it's a tool, decision making tool, for acute heart failure patients in the emergency department. We, in this study, wanted to prospectively validate a decision making prognostic tool called the Emergency Heart Failure Mortality Risk Grade, or EHFMRG for short, to see how well it performed in the real world busy emergency department hospital setting.

                                                We studied just under 2,000 patients who came to emergency departments at multiple centers, and asked physicians to rate their prognostic estimation of what's going to happen to that patient in the next seven days. We compared that with the EHFMRG model, which predicts outcomes of seven days and 30 days. We were very careful to ask physicians to provide their prognostic estimates. This is their intuitive guesstimation of the risk of the patient before calculating the score because we didn't want the physicians to be influenced by the score.

                                                What we found was that when we looked at how well physicians' estimates performed, they actually performed quite well. The c-statistic for physician estimated risk was around .7, which is a reasonable discrimination. However, the physicians' estimates were not as good as the EHFMRG risk score, which had a C greater than .8. The mathematical model seemed to do better in terms of predicting what's going to happen to the patient than physicians' estimates.

                                                Interestingly, when we combined the physicians' estimates with the EHFMRG risk score, the c-statistic improved by another 1%, so there's some additive value of having both factors combined.

                                                The other interesting finding was that patients in the lowest risk groups had 0% mortality at seven days, and 0% mortality at 30 days. We may be able to identify, using the score, patients who have a very low risk of events in that seven to 30 day period after emergency department presentation.

Dr Carolyn Lam:                Thanks so much, Doug. I have to tell you, I am a fan of the EHFMRG score. In fact, we're trying to study how well it performs in our local situation even here in Singapore.

                                                Justin, you've been thinking a lot about this. I would love for you to share the reactions that we got when we discussed this among the editors.

Dr Justin Ezekowitz:        We had a lot of good discussion about this from a number of different aspects. First, it's an in-practice assessment, a physician-based risk assessment, as we survey hundreds of physicians in the ER, which is a busy environment, and get these types of information. That's a very unique piece of this study where, in addition to the just under 2,000 patients and collecting the other data in a robust way, this really does have a potential to contribute to the literature.

                                                A lot of the discussion was about how data rich this is, and that this is an area where unlike acute cardiovascular disease where there are good risk assessment tools and other therapies, it's a really need of a scoring system that was well validated, can be replicated, and both in clinical practice as well as in selective cohorts. Doug, my congrats to your and the other parts of the team that's helped put this together.

                                                One of the questions that came up when we were discussing it was the risk textiles and buckets were very important for people to think about the very low risk, as you mentioned, 0% all the way up through much higher percents for seven day mortality, but how discrepant the risk was of the physicians versus the mathematical model; and a very good reminder of the inaccuracy of sometimes our assessments of risk in practice, especially in acute care.

                                                I wonder if you could comment on what your fence was from the physicians who participated in the study, and then the data of those, the most striking findings of that piece about where physicians make judgements on risk in for that seven-day mortality. Just any comments you may have?

Dr Douglas Lee:                 We didn't know what to expect because there haven’t been many studies of this type before. What we found in our study was that physicians tended to overestimate the risk of lower risk patients. They thought bad things would happen to healthier patients, just to put it very simply. Physicians also underestimated the risk of the highest risk patients. They thought that the highest risk patients would do well.

                                                We were surprised about that finding, but also, we were not surprised in the fact that it seems to explain some of our earlier findings that in our earlier work, we found that low risk patients are hospitalized, and we think it's probably that physicians are admitting those patients because they want to ensure that they're making a safe decision; and no harm will fall in the patient. Maybe physicians are erring on the side of admitting those patients, even though they know they're a little bit low risk.

                                                At the other extreme, physicians underestimated risk in the highest risk patients. We think it might explain the observation that we made previously that sometimes high risk patients are discharged home, and they die at home after discharge. That may be because patients who look well to physicians, I think there's great value in the clinical experience of a seasoned physician looking at a patient and knowing that, that patient is sick or not sick. But in certain cases, patients may look relatively well, but their numbers would indicate that they're actually higher risk. I think it's that group where we found they're higher risk, but physicians thought that they were healthier than they were. It seems physicians' estimations really have great value, but it seems that they can be improved.

Dr Carolyn Lam:                Sean, you discussed this beautifully in your editorial. Share with us your thoughts, and especially thoughts on the question you posed: are we any closer to the holy grail of safe emergency department discharge based on acute heart failure risk rules?

Dr Sean Collins:                 Doug, kudos to you. Nearly 2,000 patients, nine different hospitals, prospective data collection, as Justin said. I don't think this can be overstated. From a data cleaning perspective, this is truly a labor of love, and to get this done, congratulations to you and your team.

                                                I think the most interesting part of this is this exact disconnect of patients look well who are high risk, and patients may look a little bit unwell who may be low risk, ironically. That's where a risk tool is much needed, as Carolyn said in her introduction to sort of change the dynamic of 80 to 90% of patients are admitted to the hospital. If we even chipped away at 10 to 15% to able to be discharged, it would be a huge win for partly for management for an emergency department perspective.

                                                I think that the importantly, the next steps will be now looking at implementing this in some sort of a randomized manner, somewhat like what you did with asking physicians gestalt about what their level of risk is, but really finding out how does a physician gestalt when it comes to nuance and heart failure. A relative amount of congestion, even when the tool says the patient may be low risk, can they go home? I think that will be the crucial next step to find out how much does this augment and/or detract from physician decision making? We have a long way to go, as Carolyn said. It's just the complete opposite at almost every other disease process, including chest pain, from a discharge perspective. Even a little bit improvement would be great, so I'm looking forward to seeing the next steps, and I'm wondering what your thoughts are about the next steps, Doug.

Dr Douglas Lee:                 There's actually great value in physicians' clinical judgment. It's been, I think relatively understudied. I'm hopeful that future studies where decision tools or prognostic tools are validated, we can see more potentially, more comparisons with clinicians because we don't have a real great understanding, I think, of how doctors think, especially in an acute setting. More research in this area, I think would be really helpful, especially as we ... As more and more clinical decision tools being published, it would be great to see how well they hold up against good clinician judgment.

                                                In terms of next steps and implementation, when we talk to our emergency colleagues, they have brought up an issue about it's great that patients are low risk, and that we could potentially discharge them from hospital; but where is the receptor to take that patient and to care for that patient once they've left the hospital? Are they going to get good care once they leave the hospital? Are there structures in place?

                                                We're now embarking on testing this in the clinical trial where we will be comparing two strategies. The first strategy will be using the risk score at a hospital-wide level, and then discharging home patients who are in the lower risk categories, and having them follow up, and receive their care in a rapid ambulatory follow up clinic within two to three days after discharge from the emergency. This will be compared to the control, which is not using the risk score, and having usual follow up care. This trial is called the Comparison of Outcomes and Access to Heart Failure Trial, or the COAHFT trial. It is currently ongoing.

Dr Sean Collins:                 Great point, Doug. As Carolyn suggested with chest pain and heart failure as the interesting dichotomy is that unlike chest pain, when we safely rule somebody out and send them home, we're sort of done with that acute episode. Heart failure, it doesn't end. We've found that they're safe enough to go home, but now they need great collaboration and outpatient support with their heart failure provider, which may be as equally heavy lift as externally validating the EHFMRG score. You bring up a great point, which is we need to have outpatient follow up and collaboration for this to be successful. Thanks.

Dr Carolyn Lam:                Awesome comments, guys. Could I switch tracks a bit and maybe just ask Justin to round up by sharing? Circulation, we get a lot of papers about risk scores and so on. There's a bit of fatigue, I think, about scores in all kinds of things. Now, could you maybe tell us, Justin, what makes us look at a paper twice, and in fact, feature this one with a good editorial? I mean it's clearly very clinically applicable. Could you share some thoughts there?

Dr Justin Ezekowitz:        Yes, that's a great point. The things that make a risk score like this kind of elevated into kind of a circulation level of manuscript is A) the data quality has to be excellent. There has to be lots of completeness of data, but also capture of elements that we think are quite important. Two, the data science about how it's analyzed and put together, and interpreted, it has to be to the bar that we feel would be robust, and be able ... if somebody could repeat it and replicate it without an obvious challenge to the quality.

                                                The third, I think is the clinical applicability. It's okay to write a data model and come up with all these great risk scores, but if they haven't been thought through about how either a patient will be seeing this, or clinicians behave, or the environment that it has to be deployed in that, that isn't necessarily going to be something that is going to be implemented. Then, the question is: why would somebody do the study in the first place?

                                                Now, it's okay if somebody's forward thinking and saying, 'Look, EMRs are coming, or other EHRs around, so this could be implemented if there was enough impetuous and it's a good enough quality.' That's actually okay, but in the reverse where if you try to implement a model that is too complex, and it's in a hand-off to the environment, it just won't work. We just want to make sure people have thought that next knowledge translation and dissemination approach through.

                                                The final part is things that have a very local impact are, that are very unique to the environment they're in, such as it only would work in your hometown or your own country because of some environment, that's okay. But under that, the much more global focus that, that is, it could be picked up and trans located to any major city, providence, state, or country, because vis vises are global. Those things have a much greater impact because the circulation leadership is global. The patients are global. The clinicians who care for them are also global. People are all looking for very similar situations and can adapt to their own environments.

Dr Carolyn Lam:                Awesome, Justin. I don't think any of us could have said it better. Those are the reasons that we're so grateful that you publish with us, Doug. Thank you so much, Sean, for your excellent editorial, too.

                                                Thank you, listeners, for joining us today. You've been listening to Circulation on the Run. Don't forget to tune in again next week.

                                                This program is copyright American Heart Association 2019.