名媛直播

Evaluation and management of patients admitted with acute decompensated heart failure are on the docket in this episode of "AP Cardiology."

Andrew Perry, MD, interviews Michelle Kittleson, MD, PhD, of Cedars-Sinai Medical Center in Los Angeles. They discuss the etiologies for decompensation, what to do with beta-blockers or renin-angiotensin system inhibitors, and how to set the patient up for successful discharge.

A transcript of the podcast follows.

Perry: Hey, everyone. Andrew here. I hope you enjoyed the last episode on ASDs and VSDs with Dr. Amber Khanna. If you missed it, you should go back and check it out because it's really good.

In this episode, I'm visiting with Dr. Michelle Kittleson, a heart failure specialist at Cedars-Sinai. I encountered , like any good medical trainee should, and looked her up and she's published a lot and written extensively about heart failure and heart transplants. Visiting with her, she is a great educator and teacher and explains things very well with a lot of data behind her suggestions, and also working in the art of medicine as well, which I think she does very well. We discussed the topic of acute decompensated heart failure -- not those patients that are admitted to the ICU, but rather patients who are admitted at the floor. We discussed topics like continuing or discontinuing beta-blockers on admission, RAS inhibitors, diuretic strategies, and a lot more.

A quick warning, there is some background noise in the first 10 minutes that can be a little distracting that I unfortunately was unable to get rid of.

...

Kittleson: I'm Michelle Kittleson. I'm Professor of Medicine at Cedars-Sinai Smidt Heart Institute in Los Angeles, and I'm the Director of Medical Education in Heart Failure and Transplantation.

Perry: Excellent, thank you. I was doing a quick bio review of yourself, and you have written quite extensively on heart failure and transplantation I have seen, so I'm very excited to have you being able to talk with me today.

Kittleson: Awesome! Don't ask me about fever and rash. I won't answer any of those questions, but if it's on heart failure, I sort of know what I'm talking about.

Perry: Okay, great! Let's jump into our case that we have to discuss today. We have Mrs. Clark. She's a 54-year-old woman with ischemic cardiomyopathy. She's had this for the last 3 years. Her ejection fraction a month ago was 30%. She has a three-vessel CABG with a LIMA to the LAD, vein grafts with RCA and obtuse marginal branches. She also has diabetes and CKD 3. She's presenting with a few days of worsening dyspnea on exertion, orthopnea, and she's also gained 8 pounds. Her exam is notable for a heart rate of 104. Her blood pressure is 96/64, and she's satting 92% on nasal cannula. Her neck veins are distended to the angle of the mandible with her lying at 30 degrees, and she has crackles on lung auscultation. Cardiac auscultation is notable for a two out of six systolic murmur at the apex, radiating to axilla, and there's no S3.

We can get into her medications a bit later, but she's diagnosed with acute decompensated heart failure, and she's admitted to the medical ward. First off, having taken care of many of these patients, determining the cause that's precipitated their heart failure decompensation should be one of the main goals of their heart failure admission, but I find that can be really hard to confidently determine. I wanted to ask you about what's your approach to first evaluating patients like these and the cause for their decompensation.

Kittleson: Sure! Absolutely, and I completely agree that the most important thing when you admit a patient to the hospital with decompensated heart failure is to try to figure out why. Because if you can figure out the cause, you are more likely to be able to prevent it in the future. I love the FAILURE mnemonic. I got through medical school mainly through mnemonics, and this one has not let me down. F for forgetting medications or taking drugs that can worsen heart failure. Think calcium channel blockers, NSAIDs, then chemotherapy. A for arrhythmias, especially atrial fibrillation. I for ischemia. L, lifestyle, dietary indiscretion, salt, alcohol, fluid intake. U, upregulation, pregnancy, hyperthyroidism. R for renal failure. E for embolism. Think pulmonary embolism. Then S, FAILURES, like stenosis, aortic stenosis, renal artery stenosis.

That's a checklist I go through in my head every time I admit a patient with decompensated heart failure. I try to organize it by what am I going to find on the history, what am I going to find on the physical, and what am I going to look for in labs. You ask about the dietary medication non-adherence. You ask about medications like NSAIDS or illicit drug use or signs/symptoms that would be consistent with ischemia, PE, infection. On exam, of course, you're looking for Afib with rapid ventricular response, valvular stenosis, regurgitation. On the labs, you pay special attention to the hemoglobin, are they more anemic? TSH, creatinine, troponin, and sometimes you're going to add on a tox screen, VP scan, CT angiogram for PE if the history and physical directs you that way.

I think it's also really important to note that in some cases -- despite an exhaustive history, physical, and lab work -- you can't find a cause for the decompensation. And that's almost more concerning, because it means that their heart failure is so tenuous that nothing can tip them over. That's often a good sign that they may be worsening and you need to think about advanced heart failure therapies.

Perry: Gotcha. I appreciate that. In fact, I found some data to suggest that finding no cause for someone's decompensation can be relatively common, like up to maybe a third of patients.

Kittleson: Yes, and the question is, what is that old saying: Is absence of evidence evidence of absence? Sure, a third of the time maybe we don't find something. Maybe we don't look hard enough, but there are some people who do everything right and they still are admitted time and time again, and those are the patients I worry about the most.

Perry: Now that we're admitting her to the hospital, she's on metoprolol succinate at home. One big question that comes up is whether to stop or continue someone's beta-blocker on admission for acute decompensated heart failure. What's your practice and what data do we have to help guide our decisions?

Kittleson: Great point. As you said, it comes up a lot. There have been a number of observational studies and also a randomized trial on patients continue or discontinue beta-blocker therapy when patients are admitted with decompensated heart failure. In fact, a nice meta-analysis of these trials indicated that the beta-blocker withdrawal leads to significantly increased in-hospital mortality, short-term mortality, short-term rehospitalization. The take-home message is, when in doubt, continue to the beta-blocker.

Of course, the question is why is this the case? Is it a marker for worsening, meaning is the patient worse, therefore, you stopped the beta-blocker and those do worse? Are you identifying a higher risk cohort? The answer is probably not, because in all these studies, the baseline clinical characteristics, the heart rate, and the blood pressure weren't different between groups where the beta-blocker was stopped or continued. What's the more likely signal is if you stop the beta-blocker during hospitalization, these patients are less likely to have it restarted on discharge, less likely to have it restarted at their follow-up appointment, and because beta-blockers save lives in heart failure, this is a missed opportunity to provide the best guideline-directed medical therapy.

There are certainly patients who are sick enough to need their beta-blocker stopped, but they are few and far between. In the classic warm and wet patient, that is the patient that's congested and well perfused -- as judged best by blood pressure and creatinine, continue the beta-blocker. How I think about it is if you don't think you're going to need an inotrope, then they do need the beta-blocker.

Perry: Even for our patient, she's a little tachycardic and her blood pressure is a little on the lower end. We would probably continue the beta-blocker for her.

Kittleson: Absolutely. Her creatinine is not bad. It's not that much above her baseline from what we know. Also, you want to see how she does in the first 24 hours because that really gives you a nice trajectory of how the patient is going to go. You're never wrong to continue it as you gather more information. She has phenomenal diuresis in the first 24 hours? She was clearly well-perfused and the beta-blocker is not a problem. If she has terrible diuresis in the first 24 hours and now her creatinine is worsening, wait a minute. Maybe poor perfusion is coming to bear and you need to hold off on the beta-blocker.

Perry: Now, a slight tangent. Some patients are getting aggressive or maybe even moderate escalation of their beta-blocker therapy as an outpatient. How commonly do those patients then present with decompensated heart failure and perhaps we might think, "Perhaps their beta-blocker has been increased too quickly in the outpatient setting"?

Kittleson: I think it is important to . While your angiotensin receptor neprilysin inhibitor, ACE inhibitor, angiotensin receptor blocker, those medications you can up titrate very aggressively because they have acute hemodynamic benefit of afterload reduction. The beta-blockers don't. They're only for long-term mortality benefit, so I go very slowly, every week or 2 watching how the patient tolerates these slow and gradual increases. That being said, the goal is always to preserve some level of beta-blockade, even if it's as a placeholder. If a recent doubling of their beta-blocker dose seemed to have precipitated the acute decompensated heart failure hospitalization, start by lowering, it if the patient tolerates it, rather than stopping it altogether simply as a marker so you don't forget to continue it long term, and the next go around you try more slowly to up titrate.

Perry: I follow. One of our next questions is we'll throw in the bit that she has a bit of an AKI. I mentioned she has CKD. Her baseline creatinine is 1.5 and she's coming with a creatinine of 1.9. She's also on lisinopril as an outpatient. We frequently see lisinopril being stopped because it is seen as a nephrotoxin and a nephrotoxic, if you have an AKI, you want to avoid the nephrotoxins. For our patient, decompensated heart failure, stop or continue the lisinopril on admission?

Kittleson: Such a good question. Ultimately, the right answer is what works. But the question is do we have data to tell us what the right answer will be? Can we predict the future in this patient, meaning which direction is her AKI going in? Is she going to get better? Is she going to get worse? How's the ACE inhibitor playing a role? We have some insight into this from a . There was a subgroup analysis of the SOLVD trial and this was one of the early ACE inhibitor in symptomatic heart failure trials. This is an outpatient study, but there are lessons here that apply.

In this study, they looked at an interaction between early-worsening renal function and randomization to enalapril and mortality. In the overall population, worsening renal function was associated with increased mortality. In the placebo group, this association was very clear. In the enalapril group, early-worsening renal function had no adverse prognostic significance. There was an interaction between enalapril and worsening renal function on mortality. What that tells us is in the enalapril group, even if the renal function worsened after beginning it, mortality was not compromised or survival was not compromised. I think that's a really important lesson that yes, you can expect that an ACE inhibitor or RAS inhibition may cause a small worsening in renal function, but that in and of itself is not a reason to withhold your ARNI [antiotensin receptor-neprilysin inhibitor], ACE, or ARB.

In addition, it's important to remember that the azotemia you see with ACE and ARBs is usually completely reversible unless tubular necrosis has set in, so if you discontinued the agent, then the creatine should return to normal. What do you do in this woman? That's a bit of an outpatient lesson, the analysis for the SOLVD trial, but how do we apply it to the inpatient setting?

I think the way we apply it is unless the patient has significant hyperkalemia or oliguric renal failure, continue the ACE inhibitor for the acute hemodynamic benefit of afterload reduction. Also, reassuring ourselves, because we know, at least in the outpatient setting, that that small degree of worsening renal function that can be attributed to the ACE inhibitor is not associated with increased mortality and knowing that if the ACE inhibitor truly is/was the problem, stopping it can lead to reversal, and you can take a day or 2 to sort that out before you stop it knee-jerk on admission for this relatively small worsening of renal function. It's also important to remember that one of the mechanisms of cardiorenal syndrome is the venous congestion of volume overload, so simply diuresing this woman effectively, for which ACE inhibitors can help their afterload reduction, may actually improve the renal failure.

Perry: Interesting. I'll have to look through that study. Patients who had a worsening renal failure, even started on enalapril, which you might think is somewhat counterintuitive, no change in their survival afterwards?

Kittleson: Exactly right.

Perry: That's a great pearl. Thank you. Our patient, she's being admitted. A recent echo within a month, 30%, with an ejection fraction of 30%. Do we need to repeat her echocardiogram on admission because some could argue that her clinical status has changed and now she's decompensated?

Kittleson: I love this question. This is one of my favorite things in the world, one of my favorite rules in life: Don't do a test that won't change your management. Put another way, if you want to get a test because "you just want to know what it shows," then that's not a reason to get the test. I sometimes joke with my residents, that if you order an echo every time a patient is admitted with acute decompensated heart failure, you're just putting the echo attending's kids through private school. Now not to say that isn't a worthy goal, because we all have tuitions to pay for, but not necessarily the best reason to order the echo. Of course, this is in jest. I love echo attendings. They're amazing. They're hardworking, but trying to get the point across to my residents.

Let's do a thought experiment, then. How is this echo going to change your management? If her ejection fraction is now 20% versus 30%, does that tell you anything? Not really. If there's severe MR, you still wouldn't go for a MitraClip or percutaneous mitral valve intervention because the MR has to be severe after optimization to qualify. I never check an echocardiogram on a patient with decompensation because I always want to know what happens when they're compensated.

Now, say I want to know what the RV function looks like because I think they might need an LVAD. I would again say, "Don't you want to look at that RV once they're as compensated as you can possibly get them to get a sense of whether LVAD support would be okay?"

Whether you're looking for a MitraClip indication or you're looking for LVAD consideration, risk stratification, whether you're looking for other valvular lesions, it behooves you to optimize the patient first before you get the echo.

Perry: Love it. Now, speaking of optimization -- and we've mentioned the topic of decongestion a couple of times or decongesting -- how do you choose your dose of diuretic for your patient? Let's just say our patient takes 40 mg of furosemide by mouth daily that we can adjust on admission.

Kittleson: Yes. The , the classic DOSE trial, which came out in the New England Journal in 2011, is basically our guide and bible, for diuretic titration. It's the only randomized trial of diuretic therapy in decompensated heart failure. So, 308 patients were divided into a 2x2 factor design, low dose versus high dose loop diuretic and then bolus versus continuous infusion. Low dose was their IV dose equal their total daily oral dose. High dose was 2.5 times that. Then was it administered as bolus every 12 hours or by continuous IV infusion. I think the design of the study was great.

I was a little surprised by the endpoint. The primary efficacy endpoint was the patients' global assessment of symptoms by visual analog scale at baseline to 72 hours. Then the safety endpoint was change in creatinine. It sort of seems to me that these endpoints were designed from a power standpoint, meaning we have 308 patients, how can we have an endpoint that we can have a power to detect a difference? But I don't find those endpoints particularly clinically relevant.

However, some of the secondary endpoints I find very useful and some of the secondary endpoints were likely to require a dose increase at 48 hours, net fluid loss at 48 hours, increase in creatinine greater than 0.3. Which group was more likely to require a dose increase at 48 hours? The bolus versus the continuous, the low dose versus the high dose. Bolus low dose is more likely to require a dose increase at 48 hours. Which group resulted in more net fluid loss? The high-dose group. Which group had a higher increase in creatinine? The high-dose group, but by 0.3 mg/dL, so not necessarily a clinically relevant increase in creatinine.

Again, I'll say it again. The bottom line is, the right answer is what works. I find what seems to work based on the secondary endpoints of the DOSE trial is a high-dose, continuous infusion. This is what I'll do: If a patient is on 40 p.o. daily as an outpatient, when they walk through the door, I'm going to put them on something high dose, like 80 IV b.i.d. I'll start with boluses because I want to give them the benefit of the doubt. But in the first 24 hours, if I don't have a good diuresis -- what I do I define as a good diuresis? I want at least 2 liters negative. I would love 3 liters negative because I kind of judge every plus of edema is about 5 to 10 pounds of fluid. If I have 2+ lateral edema, they might have 10 to 15 pounds of fluid to lose. If they lose 2 pounds a day, they're there for 5 to 7 days. If they lose 4 pounds a day, they're there for half as long. Every day a patient spends in the hospital is a chance to get C. diff or fall out of bed. Do the shorter the hospitalization, the better.

So as I want to make that diuresis as fast as possible and their blood pressure and their creatinine tell me they will tolerate rapid diuresis, then I want to go gangbusters. If 80 IV b.i.d. doesn't get me minimum 2 liters negative in the first 24 hours, more like three, I'd be happy. Then I will switch to a drip. Lasix 10 mg an hour, 20 mg an hour. You get so much great bang for your buck. I don't like t.i.d dosing. I don't like it for two reasons. First, I find that if you don't respond to 80 IV b.i.d., it's not a frequency problem. It's a threshold problem, so just increasing the frequency is not going to help. Second problem, what's t.i.d. dosing? It's 8 a.m., 2 a.m., 8 p.m. I mean who wants to give a patient a dose of Lasix at 8 p.m.? Cruel and unusual punishment.

I like the bolus and drip. Drip you can turn off between 10 p.m. and 4 a.m. The patient gets to sleep. They get to pee a lot. You can speed up their hospitalization. There's nothing more satisfying than a patient that goes down 5 liters and 5 kilos 2 days after admission, feels like a million bucks, thinks you're the best doctor in the world. That's the strategy I employ.

Perry: I have never tried the drip and turning it off between 10 p.m. and 4 a.m. That is a great idea. I really like that.

Kittleson: Your patients will love you for it. There's nothing like sleep in the hospital.

Perry: So we diurese her. Let's think ahead a little bit on discharge. I know there's some subcurrents and maybe a subtle fan club for . I kind of want to ask where you stand in these conversations.

Kittleson: Yes. As with most things in medicine, I consider myself a curmudgeon. I like the old. I like the tried. I like the true. I'm very skeptical of new-fangled things. Torsemide, it has increased bioavailability. It has a longer half-life. Yet, there's been some observational studies that support no clear benefit of torsemide over furosemide. As I've said before and I'll say again. The right answer is what works. For me, the right answer is often what's the cheapest. Furosemide is cheap and it often works. Here's the problem. If the heart failure is so bad that I'm trying to juggle my choice of oral diuretic to maintain euvolemia, this patient has bigger problems at hand. Torsemide is not going to be the magic bullet that takes someone who's refractory to furosemide with frequent hospitalizations and turn them around. If you're all of a sudden so desperate that oral furosemide isn't working and you're reaching for torsemide, you're reaching for bumetanide. It's telling you this patient has advanced heart failure and needs something like transplant or LVAD consideration.

Perry: I appreciate that. Now as we are decongesting, we frequently make our patients hypokalemic and hypomagnesemic. I remember as an intern one of the first things that I was told was make sure that K is greater than 4 and the mag is greater than 2. For that first year, I wondered where does this come from and why are we so religious about this threshold?

Kittleson: Yes. Changes in potassium, of course, and magnesium are very common when you're diuresing someone. That being said, there's been data from a number of trials that indicate that the change in potassium during hospitalization isn't associated with mortality. That being said, I do respect hypokalemia and hypomagnesemia for two reasons. First, you'll often notice that if you aggressively diurese a patient they will feel horrible. They will feel great because their shortness of breath is better, but they will feel horrible because they're crampy, kind of like I imagine I would feel if I ever ran a marathon and all my insensible losses and electrolytes had gone out of whack.

What I try to do is be incredibly aggressive for that reason. If you're actively diuresing them and trying to replete the potassium, you are going to be behind the eight ball. It's kind of like I remember back to diabetic ketoacidosis as an intern. Whatever you think the potassium is, it's really lower than you think it is because the glucose makes the potassium inaccurate, and their total body potassium is so low and it's going to bottom out. I think of the ongoing aggressive diuresis in the same way. I like to give lots of spironolactone and you say, "But wait a minute. This woman's creatinine is a little higher than her baseline." I remember back to those observational studies, and if your creatinine is high, you might have dangerous hyperkalemia from spironolactone, to which I would answer, "Yes, as an outpatient, sure, but you're checking the potassium every day in the hospital. If you give spironolactone, it will offset the need for those horrible potassium horse pills, right?"

Potassium isn't a gel cap. Potassium is an awful big pill that patients hate to swallow. If you give spironolactone, you will offset some of that potassium wasting and you will also make the patients happier for having to take less potassium pills, less burning potassium through their IV. It's a win-win. I oral mag them out. I mean 400 mag ox, 800 mag ox twice daily, 800 t.i.d. You can really never overdose on magnesium. Think of all those pregnant ladies on their mag drips to relax their uterus. Plus, if you give it orally, they'll just get diarrhea, so I love to max out the mag. It's a great way to control their cramping and then be aggressive with the potassium.

The second reason I really like to be aggressive with potassium and magnesium is not just for their quality of life, but because of the horrible iatrogenesis imperfecta of hypokalemia-induced defibrillator shocks. You only have to see that happen once to never want to see it happen again. It's worst feeling in the world to know that as you're trying to make the patient feel better, you've actually made them feel worse. Staying on top of potassium with the spironolactone, with aggressive K repletion is important. Even if an observational study tells me it's not associated with mortality, my experience tells me that if you want their cramping to be less, you want their quality of life to be better as they're in the hospital, the best thing to do is aggressive repletion.

Perry: Perfect, I love it. Now, let's think about teeing up our patient for a successful discharge and a successful next few months outside of the hospital. We haven't gone over her medications in full at the beginning, so let's just review those now. She came in prescribed on aspirin, atorvastatin, lisinopril, metoprolol succinate, and furosemide. When we're thinking about her discharge medication reconciliation or thinking about other medications, how might we adjust these medications in terms of dose increases or in terms of additional medications?

Kittleson: You gave me a really neat slam dunk there allowing me to talk about the miracle of the . So, so exciting. I think a lot of people remember where they were when JFK was shot, depending on your generation, or when the Challenger space shuttle exploded or when Princess Di was killed in Paris. Anyway, heart failure transplant cardiologists remember where they were when PARADIGM-HF came out in August of 2014. I remember exactly where I was. I was about to board a plane for our Kittleson family vacation at the lake house in Wisconsin. It came up on my email from the New England Journal table of contents, and I was transfixed by the first positive trial in heart failure in 10 years. ; PARADIGM-HF, 2014. It was a miracle and a game-changer when it came out. Sacubitril-valsartan [Entresto] beat out ACE inhibitors in active control, not a placebo-control, an active-control trial. ACE inhibitors, which have been the reigning king and queen of heart failure since 1987 and the . Really incredible.

This woman, no brainer, she's got to be on sacubitril-valsartan. The miracle wonder drug of wonder drugs for heart failure patients. Of course, we must remember that the reason it's sacubitril-valsartan and not sacubitril-lisinopril combination is because of the risk of angioedema. Sacubitril being a neprilysin inhibitor. It doesn't just inhibit neprilysin, which degrades natriuretic peptides and thus increases natriuretic peptide levels. Sacubitril also inhibits the breakdown of bradykinin, so sacubitril leads to more bradykinin, just like ACE inhibitors lead to more bradykinin. The combination of lisinopril and sacubitril will lead to angioedema. Big, big no-no, contraindication, class 3 warning by the guidelines and easy fodder for any board question. Thus, we have a 36-hour washout whenever you have someone on an ACE inhibitor and you want to prescribe sacubitril-valsartan.

I love to do it in someone in the hospital, best time ever, plus you have your awesome, amazing, in-hospital pharmacist to help you get the insurance company authorizations on board. Stop her lisinopril when you find she's nearing the point of euvolemia because then you see her creatinine is stable. There was no problem with that ACE inhibitor after all. Stop the ACE inhibitor, 36 hours later give her sacubitril-valsartan. When I do it as an outpatient, I tell the patients 48 hours. It's just easier to give the instructions when they're an outpatient as just a 2-day window.

The only one contraindication we have to remember is angioedema. Now the guidelines aren't 100% clear on this. For the PARADIGM-HF study, angioedema was a contraindication and it wasn't just angioedema to ACE inhibitors, but angioedema in general. What are you supposed to do if they get hives to peanuts or shellfish or sulfa? The data is not clear. I tend to just stick with lisinopril in those patients because it's a data-free zone until we get more experience, but assuming this woman doesn't anaphylax to other stuff, sacubitril-valsartan is 100% the right way to go. Once she tolerates sacubitril-valsartan, you've already kept her on her beta-blocker, wisely so. You will then add spironolactone to her regimen and what a wonderful thing just to get it on during the hospitalization.

My goal before someone is discharged is to have them on a low dose of everything and then have the outpatient doc titrate things up over time. If she was on 24/26 sacubitril-valsartan, her usual dose of metoprolol XL, spironolactone 12.5, beautiful. Now the super-new excitement, so where PARADIGM-HF was in 2014 is where we are now with in August of 2019, super exciting. How could it be an SGLT2 inhibitor, a diabetes medication reduces heart failure hospitalization and death in non-diabetic heart failure patients? Craziness, right?

Would you give her dapagliflozin [Farxiga]? As long as her eGFR is greater than 60, you would. Now this one I might wait for an outpatient just because I think the eGFR greater than 60 is very important for the dapagliflozin, and because of that, I'd like to see a good stability of her basic metabolic profile as an outpatient before I add it on. That one, I think, can wait. But the sacubitril-valsartan, metoprolol, spironolactone, big three she should be on before discharge.

Perry: I feel like sometimes there's hesitancy to start or change all those medications within the hospital without close follow-up. Do you have any comments about those sorts of hesitations?

Kittleson: Yes, make sure there's close follow-up. It's a very easily reconciled problem. No, I'm a big believer in make the most out of the hospitalization. "Oops, I'm not sure about follow-up" is never the excuse because either the patient belongs to the attending who you happen to be rounding with, in which case that attending is therefore responsible that the baton has been passed, or it's one of the attending's colleagues within that heart failure group, in which case the communication is clear. Here in Los Angeles, we have a lot of patients who come from lots of faraway places. I mean, in Los Angeles 10 miles can be far away, because it takes an hour to drive there.

In that situation, you pick up the phone, you make a call, and you fax the discharge summary. There's never a reason not to use that hospitalization for everything it's worth. Any cardiologist knows, even if you gave them no information, when they see that patient for their first post-discharge visit after decompensated heart failure, what are they going to do? Check a basic metabolic panel. That's sort of standard heart failure cardiology 101. So even in the worst-case scenario, the cardiologist knew nothing, they will still do the right thing.

Perry: Perfect. We have talked a lot about floor management of acute decompensated heart failure. That's a lot of bread and butter of internal medicine just in general. One of the last closing questions I want to bring up is about patients who are sicker, who require ICU-level care, who are in cardiogenic shock. This question, I think, is starting to raise up or some people would advocate for early initiation of mechanical support in patients with acute cardiogenic shock rather than titrate afterload and inotropes. I wanted to ask and get your thoughts and impressions on those types of patients who are different from this patient we've been discussing.

Kittleson: We'll start with inotropes. There is what I like to call the holy trinity of heart failure: volume overload, hypotension, and renal dysfunction. I like to think that if you have volume overload plus one of these other things -- hypotension or renal dysfunction, then what you're doing isn't working and you need more information. The best place to get that information is often with a PA [pulmonary artery] catheter, meaning someone is volume overloaded, and as you diurese them, either the blood pressure is going down or their creatinine is going up. Either they're really not wet like you think they're wet or they're so poorly profused they need inotropic support to help on with that diuresis. The PA catheter will answer those questions for you.

Your two options in that setting when you're evaluating the holy trinity of heart failure is either a PA catheter or inotropic support empirically. Which of those is the right answer? You could say I have a patient, say this woman, for example, she came in and her creatinine was 1.9 and she was clearly volume overloaded. You gave her 80 IV b.i.d. of Lasix and in 24 hours she peed 200 cc's, still clearly volume overloaded on exam, and now her creatinine is 2.4. You could say to yourself, "I don't really want to stick her with a PA catheter. It's an invasive procedure." You could start some inotrope, dobutamine at 3, milrinone at 0.2, whichever one you want. The right answer is what works.

Then 24 hours after that, her creatinine is 1.8. She's peed 4 liters. Then what you did worked, but you're still kind of at a loss. Is she now better because she's on a "better part of the Starling curve," and now I'm going to wean off this inotrope and she's going to be fine? Or is this a marker that she's sick enough that we need advanced heart failure therapies? Because you could argue every time they need that inotropic support, even if it's temporary, even if it can be weaned off, you've lost a little bit. A few of those nephrons have died through the stress of that decompensation. A little bit of her heart reserve has gone away. I would argue any time you need an inotrope, even if you think it's going to be short term and you're going to wean it off, it's good to know exactly how bad the patient is doing.

Now if the patient is 85 years old and not a candidate for any advanced therapies, no, I don't think that information would change your management.The need for inotropes will trigger a discussion about goals of care and shared decision-making in that sense, but the PA catheter is not useful. In a 50-something-year-old woman with lots of options ahead of her, you need to know exactly how bad she's doing.

Then, when it comes to the mechanical support in these patients with cardiogenic shock, again, that's why I think it's nice to have the PA catheter in place. Because if a patient looks "okay" and then you give them a little bit of inotrope and they do "okay," you can often underestimate how bad things really are. If you have the numbers, it can help you direct yourself appropriately to triage to recovery back to optimal guideline-directed medical therapy versus triage to temporary mechanical circulatory support, durable mechanical support, or transplantation. I do believe that while inotropes are good empirically, the information you will gather from the PA catheter can be invaluable.

Perry: I totally agree. I think PA catheters for a while had been getting a bad reputation and I think underutilized in many situations.

Kittleson: Here's what I think. The classic trial was the , JAMA 2005, randomized patients with cardiogenic shock to PA catheter or no PA catheter. There was no difference in survival, rehospitalization, ICU stay. But here's the thing: PA catheters were no better, but they were also no worse. This is what I think the problem can be. I remember being a medical student, rounding in the surgical ICU, and God bless those surgeons. They can do things I could never do. I would faint dead away in the operating room, but they would say things like, "The patient is not doing well. Let's throw in a Swan." But the very act of throwing in the Swan, it's not a therapeutic maneuver. It's a diagnostic test. Putting it in isn't going to make the patient better. It's what you do with the information. That's really the issue.

I believe if you know what you're going to use it for, if you have a plan for the information you're going to find, then it is going to help you. That's why I take the "bad rap" with a grain of salt because I know, I have confidence that, if you are not clear on the volume status or you're not clear on their cardiac index, then you know what you're going to do with the information and it will benefit the patient.

Perry: I appreciate a lot of your pearls and guidance in answering a lot of these different questions. I think one thing that you've come to back again is this phrase "what works, works."

Kittleson: That's right.

Perry: Settle back on that.

Kittleson: I love working with my residents and fellows. One thing I love the most is parsing out what is the art of medicine and what is the science of medicine. Certain things are non-negotiable. ACLS is non-negotiable. Giving a diuretic to someone in heart failure is non-negotiable. But then the choice is so much of an art and it's so much based on experience. I love to sometimes tell the residents or fellows, "I'm not going to micromanage you. I wouldn't do it exactly that way. Let's do it your way and let's reconvene in 24 hours." Nothing bad will happen to the patient, but let's get a sense of the art and what works. The best way to learn something is by trying it and modeling yourself after practices that have worked through your experience and training and watching your mentors, role models and attendings do it that way. When it comes to something like dose of diuretics, when it comes to the choice of inotropic support, so much of that is art.

I posted a poll on Twitter the other day, because I just love my Twitter polls. They're fascinating. It was a patient very similar to this one, who came in with decompensated heart failure, warm and wet, and no response to 80 IV b.i.d. of Lasix. The choices I gave were 80 IV t.i.d., 160 IV b.i.d., drip at 20 mg an hour, or add some metolazone. Something like 2,000 people responded and it was basically 25% for every choice. You know what that tells us? The right answer is what works. If you can diurese the patient effectively, safely, then the patient will do well. Now, of course, I have my clear preferences for a Lasix drip in that setting, a furosemide drip, I should say, because I know in my experience it is the quickest and most effective way to get the job done, but that doesn't necessarily mean the other answers won't get us to the same place.

Perry: No, I do recognize you have a very good Twitter game going on. Remind me. Your Twitter handle?

Kittleson: @MKittlesonMD.

Perry: @MKittlesonMD. Perfect. I'll have that in the show notes.

Kittleson: Cool, thanks so much.

Perry: Any last pearls or words of wisdom before we sign off?

Kittleson: If I had my top take-home messages for acute decompensated heart failure, number one would be try to get a sense of how sick the patient is when you meet them, and that is are they warm and wet or are they cold and wet? The best way to decide that is how is their blood pressure and how is their creatinine and how do those parameters respond to your initial diuresis? The second thing I would make clear is always try to get a sense of their long-term trajectory, even when you're meeting them for the first time. Is this going to be a quick, easy diuresis for a few days, pop them back home, everything is fine? Or is this the beginning of the downward trajectory to advanced heart failure as indicated by a more cold and wet presentation or an inability to find triggers for their decompensation? Have those warning signs in the back of your head so that you don't miss that slippery slope and triage them appropriately as needed.

Perry: Excellent, perfect. Great closing words. Again, thank you so much for visiting with me, and I'm sure the audience will really appreciate it as well.

Kittleson: It was an absolute pleasure. You ask the best questions. Thanks so much.

Perry: Thank you. That was a lot of information, and there was a lot of data and papers that were referenced. ... I think she summarized the conversation nicely there at the end, so I'll skip that part, but I will say I was very happy to hear her use the phrase "iatrogenesis imperfecta." That was a phrase taught to me by a senior resident when I was an intern. I like using that phrase along with iatrogenesis fulminans, but you never want to be the one who causes iatrogenesis fulminans. Anyhow, thanks for listening and we'll see you next time.

is a cardiology fellow at the University of Washington Medical Center in Seattle.

Comment