Christopher O’Connor, MD: Hello. Welcome to this program titled "Biomarker-Guided Heart Failure Therapy: Is It Cost- Effective?" I am Chris O’Connor from the Duke Clinical Research Institute and the Duke Heart Center in Durham, North Carolina. I am joined by Jim Januzzi, my good colleague from the Massachusetts General Hospital in Boston. Jim, first of all, congratulations. Mass General was ranked the number 1 hospital in the world.

James L. Januzzi, MD: In the United States, actually, but thank you. Yes, you guys did not do too badly yourself.

Dr O’Connor: Thank you. Heart failure is a tremendous burden on the cost of care in America today. As a heart center director, I know this very well. My administrators and the leaders of the healthcare system are always on me about the cost and burden of heart failure admissions and readmissions. Tell me what you think the challenges are in the management of acute decompensated heart failure.

Dr Januzzi: This is an issue for administrators and for clinicians like me in the trenches. There is a tidal wave of admissions to the hospital now, due to the rising incidence and prevalence of heart failure, both in the chronic stage as well as in the acutely decompensated stage. A considerable number of patients on the medicine service at the Mass General Hospital have heart failure either as their primary or secondary diagnosis.

James Fang published similar data in the Journal of the American College of Cardiology, showing the rising incidence of heart failure.

The cost challenges presented by this diagnosis are considerable, whether they are related to the fact that heart failure has terrible short- and long-term prognostic implications, or that heart failure is associated with a dramatic rate of hospitalization.

We are facing a real crisis.

Dr O’Connor: In our hospital, the cost of heart failure is millions of dollars for the over 600 to 700 admissions that we get. We realize that if we can cut even 6 hours off the length of stay, we could save a quarter of a million dollars. There is a lot of interest in figuring out a better way to manage these patients through the system and prevent readmissions that drive up the cost.

Dr Januzzi: It is very clear that something needs to be done. Whether it is evaluating the patient at the front end of their hospitalization or assessing the adequacy of the job we have done at the back end of the hospitalization, there is a lot of focus now on optimizing how we evaluate our patients and take care of them.

For example, we first looked at biomarker-guided evaluation years ago. In 2003, we looked at N-terminal prohormone of brain natriuretic peptide (NT-proBNP) for diagnosis of heart failure in our emergency department at the Mass General. We found that a substantial percentage of patients presenting with shortness of breath were viewed with some degree of indecision as to whether it was pneumonia, heart failure, or obstructive airway disease exacerbation. We found that measuring NT-proBNP reduced that indecision substantially in those patients.

This has ramifications because the patients who experienced long lengths of stay were judged with that indecision. Frequently, their therapy was delayed; they did not get appropriate heart failure-directed care for a day or 2 because they were being treated for a wide range of diagnoses.

By adding a natriuretic peptide test to the mix, we showed fairly substantial cost-effectiveness from correct diagnosis.

That has been subsequently validated in a multicenter randomized trial called the IMPROVE-CHF Trial that we published in Circulation a few years back. Once again, it showed that accuracy of diagnosis has a profound effect on hospital length-of-stay and utilization.

Dr O’Connor: We ran the ASCEND trial and 1 of the things we learned was that time-to-treatment is important for improving outcomes and shortening length of stay. If you are in the emergency room for the initial hospitalization and you do not know what the shortness of breath is due to - - is it a pulmonary embolism, is it an aortic dissection? - - you can go down a very long track.

Dr Januzzi: I agree. In fact, Alan Maisel showed in the ADHERE Registry that door-to-brain natriuretic peptide (BNP) time actually was associated in a linear fashion with a risk for mortality. The faster the diagnosis was made, the faster to diuretic. In large-scale registries like ADHERE, one can identify the subtle nuances of prognosis associated with those differences. I actually talk about the door-to-door diuretic time to our patients with acutely decompensated heart failure and I think we are going to be focusing on that sort of thing in the future.

Dr O’Connor: Should it be a quality metric like door-to-balloon time?

Dr Januzzi: We are certainly looking at a number of different things like door-to-cooling time for our patients with cardiac arrest, and door-to-diuretic time. It is always good to be ahead of the curve. We are now seeing door-to-troponin time for chest pain, so I suspect we will soon be focusing on door-to-diuretic time with heart failure as well.

Dr O’Connor: Let’s say we now have the patient in the hospital and we are treating him and we are diuresing him. If we measure a biomarker near the end of the hospitalization and it still remains quite elevated, what does that mean?

Dr Januzzi: That is a real conundrum, is it not? The patient says that she feels better and is ready to go home. You have the little old lady in bed with her purse next to her, ready to get out of the hospital, and yet her natriuretic peptide level has not fallen, and her weight is down a couple of pounds, and her vital signs seem to be in range. We see this on rounds quite frequently. It is a common scenario.

One thing that has become increasingly clear is that the discharge value for BNP or NT-proBNP is very telling with respect to the short- and intermediate-term prognosis. If a patient is discharged from the hospital with a persistently elevated natriuretic peptide, particularly if it is rising or unchanged from admission, that patient's likelihood for rehospitalization or even death within the next 30 days to 6 months is substantial. We have focused on this point because we think that when you look at a patient, you should use a mixture of subjective and objective tools. Although a biomarker is a bit of a forward-thinking way of considering objective tools - - we typically think objectively about using our stethoscopes, which we still need to use, of course - - using a biomarker to decide that a patient is not recompensated is a very sound thing to do.

For example, at Mass General, we will look at a patient's predischarge natriuretic peptide value. We use NT-proBNT, and if it is rising or still at the level of the hospitalization admission, we keep the patient in the hospital and intensify his therapy.

Dr O’Connor: I like that because I think it is critical.

We have shown in OPTIMIZE-HF that natriuretic peptide is an important indicator for readmission and mortality. We talk to our residents and fellows about looking for a decrease by about a third or a half. If the patient does not achieve this, I advocate that we need to keep them 1 more day or a half a day to optimize their evidence-based medicines. There is a lot of pressure from healthcare administrators to keep length-of-stay short, but in doing so we may lose it on the back end.

Dr Januzzi: That is the thing. What do you think about 30-day rehospitalization rates and heart failure?

Dr O’Connor: It is a challenge. We should look at total hospital stay over a 60-day or 30-day period, not just the 30-day readmission rate. Do you know the number 1 way to reduce your hospital readmission rate?

Dr Januzzi: Yes, keep the patient in the hospital. I think you guys showed that in ASCEND.

Dr O’Connor: In Eastern Europe, where the length-of-stay is 17 or 18 days, they have the lowest readmission rate at 30 days because there are only 12 days left.

Dr Januzzi: Maybe so, but on the other hand, you could argue that prolongation of hospitalization in a high-risk population is justifiable in order to mitigate that 30-day rehospitalization rate. We have all seen those patients. You send them home and then, literally, not even a day later, you get the call from the emergency department that they are back. Some patients objectively seem recompensated - - their wedge pressure went from 30 to 25 mm Hg and they felt better - - but all it takes is a salty meal and they are right back up to where they were and, boom, like that, they are back in the hospital. Having some way to objectively manage them more aggressively, I think, is welcome.

The audience probably knows that more aggressive heart failure care is associated with reductions in natriuretic peptide values. This is true whether therapy is with diuretics, or by correcting the neurohormonal disarray in these patients. Whether you reduce their volume status with loop diuresis or add neurohormonal agents like β-blockers, angiotensin-converting enzyme (ACE) inhibitors or aldosterone blockers, more aggressive care results in a reduction in natriuretic peptide, which is associated with a better prognosis.

Dr O’Connor: The aldosterone antagonist is a very intriguing strategy that we might be underutilizing in our patients who have been hospitalized and are ready to go home. Obviously, we do not have the perfect evidence-based trials, but it looks pretty good.

Dr Januzzi: It is promising.

Dr O’Connor: If you look at ACE inhibitors and β-blockers, we are getting up to 70%, 80%, 90% utilization, but with aldosterone antagonists, we are still at 40%.

Dr Januzzi: There is opportunity there. Better utilization could lead to better care. We have a strategy where we evaluate patients for their risk for rehospitalization. If they are identified as being at lower risk for rehospitalization, there are minimal associated costs. If they are at intermediate risk, we will enroll them in a telemonitoring program of some sort. Whatever the case may be, this is a perfect link into office-based management where we are even more aggressive with biomarker-guided heart failure care.

Dr O’Connor: We measure the natriuretic peptide at the time of discharge, or right before discharge, then the idea is to get the patient back in the clinic within 7 days. We think this is a high-risk group that we can target with higher surveillance and do more follow-up to get them in the clinic early so that everything is adjusted perfectly.

Dr Januzzi: If you get a follow-up NT-proBNP value at a week after discharge, it is very telling about short-term risks. We have shown this in the PROTECT study, which was our outpatient heart failure trial. If you see a patient in the discharge clinic a week later and his NT-proBNP or BNP is rising rapidly, you have identified a patient who has a very high likelihood for a short-term complication. The point here is to improve care while mitigating healthcare expenditures.

Dr O’Connor: Now we have the discharged patient in the outpatient setting. We have prevented that early readmission. Is it cost-effective to be monitoring patients with natriuretic peptides? Do we see a savings? It is not terribly expensive, but it takes a little more effort.

Dr Januzzi: It does take more effort and it is 1 of the things people are concerned about. The aggregate experience has been that utilizing biomarkers to monitor and manage patients results in more visits to the office. In Rudolf Berger’s data in the Journal of the American College of Cardiology^[1] a couple years ago and our experience in PROTECT, patients were seen in the office more frequently when they were treated with a biomarker-guided approach. That makes sense actually, because the goal here is to utilize a tool that identifies something that you would not have already known using your clinical knowhow. Patients are seen more frequently and utilization of necessary medications is increased on the outpatient side.

The question is, does the potential cost savings balance out?

In the PROTECT study, we showed about a 50% reduction in heart failure rehospitalization. This, more than anything else, explains why we saw a cost savings. It is irrespective of the cost of coming to the office more frequently. Thirty percent of heart failure expenditures are related to the hospitalization. We found about a 22% reduction in total costs in the PROTECT study for the patients who were randomized to the NT-proBNP-guided arm.

You can look at it either in terms of the entire time in the study, or by cost-per-day average, which I think is a more reasonable way of thinking about patients who are in the trial for just a short period before they passed away or patients who made it to a full year without any events. Regardless, the per-day costs were lower in the NT-proBNP-guided arm.

That looked pretty good. That is just straight-up total cost of care. We do not have any modeling data just yet.

Dr O’Connor: We saw that data and it looks very promising, particularly the difference by age.

Dr Januzzi: TIME-CHF is an interesting study and their cost-effectiveness data are going to be published in JACC Heart Failure. TIME-CHF was a trial randomizing patients to a NT-proBNP-guided therapy versus a standard management approach. By design, TIME-CHF included older, elderly patients. The majority were older than age 65 and most were above age 70 or 75. TIME-CHF found an interesting interaction between age and outcome. Older patients, whether in the standard-of-care arm or in the NT-proBNP-guided arm, had worse outcomes. No surprise. We see this because older patients are harder to manage.

But even if a patient was over age 75, there was cost-effectiveness when investigators performed incremental cost-effectiveness ratio analyses for NT-proBNP-guided care. This was really very interesting. At various levels of cost-effectiveness, the potential improvement in quality-adjusted life years using NT-proBNP-guided care was more effective than standard-of-care. This goes for straightforward, uncomplicated patients, patients with 1 or 2 comorbidities, patients with renal failure, and even in the elderly patients. It is not that elderly patients do not benefit from NT-proBNP-guided care, rather, it is that they are a higher-risk population, in whom any intervention is going to be more difficult. Even in this context, guided therapy is cost-effective above standard management.

Dr O’Connor: The National Institutes of Health (NIH) have embarked on the GUIDE-IT study. You and I are both involved. Mike Felker is the principle investigator. What will we learn from GUIDE-IT? There has been a meta-analysis that shows a mortality advantage. We already see it in your trials and other trials and other registries about cost-effectiveness. Do we need this trial?

Dr Januzzi: The first principle about medicine is that it is worthwhile to generate hypotheses from small studies. Certainly we have generated a lot of interesting data from our small pilot studies, but in order to really change how we practice, to justify that level 1 indication in the heart failure guidelines, for example, the pivotal multi-centered randomized controlled trials are necessary. It is fair to say that none of the studies that have been done so far -- even PROTECT, which was small but rather definitive -- are sufficient to lift the approach to that level 1 indication that would cause the everyday physician to reach for a biomarker to manage patients. The trials are necessary because heart failure care is complex. We have a narrow range of therapeutics that we can use for these patients, so how we use the things we have at our disposal needs to be reevaluated. That is what biomarker-guided care can do for us.

Dr O’Connor: The other nice thing about GUIDE-IT is getting biomarker-guided therapy out into a broader population of practitioners.

Many studies, including yours and ours, have been centered around people and centers that have experience, knowledge, and expertise with natriuretic peptides. We want to get broader implementation to the community cardiologists.

Dr Januzzi: That is one of the greatest strengths in the way that GUIDE-IT was designed. We will have a wide range of practice types in the trial. This will tell us how patients are managed across the broad range of disciplines that take care of heart failure in the United States of America. If you look at PROTECT, 97% of the patients were on β-blockers at randomization. This is characteristic of a tertiary care referral center. That sets a very high bar for guided therapy to get over, and yet we were able to do so. Now, we are bringing biomarkers to a more general population. I think we are going to see an even bigger benefit to close those treatment gaps, and that, in itself, will improve outcomes.

Dr O’Connor: I think you are right. In wrapping up here, looking down into the future of healthcare, we have accountable care organizations. We cannot just be responsible for episodic care. We are going to be responsible for continuity, starting early with primary care, then hospitalization, back to the clinic, maybe into the nursing home. As a heart center director, we are very interested in how we can manage this type of population. What do you think the challenges are? How can biomarkers help us here?

Dr Januzzi: It is an incredible opportunity that we have to reconsider how we think about the patient with heart failure in terms of their journey from the development of heart failure all the way through to the end, inclusive of hospitalizations and office-based care. The use of biomarkers has promise at each of these different steps. Once a patient develops heart failure and is seen in the office, the use of markers can be a substantial benefit for the clinician for monitoring and managing the chronic phase. They are clearly beneficial for diagnosis and management in the acute phase. It is an intriguing and exciting possibility to think about how they might be used to prevent heart failure as well. At the end of the day, as we think more about the Accountable Care Organization (ACO) concept -- which is a good one, frankly -- we really have to think about the lifespan of the disease. Ultimately, prevention, where biomarkers have been shown to identify patients at risk for developing heart failure, will be how we mitigate the cost of this diagnosis.

Dr O’Connor: This has been a stimulating conversation. I have certainly learned a lot and this field is moving fast.

Dr Januzzi: The way markers have influenced care, in heart failure in particular, has moved forward at a dramatic pace in the last 5 to 10 years, and it is really exciting to be a part of it.

Dr O’Connor: Thank you, Jim, for being a part of this program. I am Chris O’Connor and Jim Januzzi here at Amelia Island.

Dr Januzzi: Thank you for joining us for this program. What would you do? What impact will this program have on biomarkers on your practice? As a reminder, please join me on Medscape Discuss, a supplementary discussion, by clicking the Join the Discussion button now. I am interested in hearing your thoughts and questions around this topic.

You may also now take the CME posttest by clicking on the Earn CME Credit link. Please also take a moment to complete the program evaluation that follows. Thanks.