CHF Mortality Benefits: Pump Up the Volume

Scenario 1:

Chief Complaint: 32 year old non-pregnant female, non-smoker, occasional alcohol use, no significant past medical history presenting with dyspnea and reports of lower extremity edema. I give you the following tools:

  • Physical exam
  • CXR
  • EKG
  • Pro-BNP
  • Framingham Heart Failure Diagnostic Criteria
  • You don't get to choose an echocardiogram (too easy)

Which ones do you want to use for your evaluation if you're at all concerned about CHF (which you're probably not), specifically systolic CHF (HFrEF)? 

Scenario 2:

Chief Complaint: 57 year old male, 30 pack year smoker, moderate alcohol use, past medical history significant for hypertension, diabetes and hyperlipidemia presenting with dyspnea and reports of lower extremity edema. I give you the following tools:

  • Physical exam
  • CXR
  • EKG
  • Pro-BNP
  • Framingham Heart Failure Diagnostic Criteria
  • You don't get to choose an echocardiogram (too easy)

Which ones do you want to use for your evaluation if you're at all concerned about CHF? (which you probably are)

Best Practice Answer:

Scenario 1: Framingham Heart Failure Diagnostic Criteria, Pro-BNP, and maybe an EKG

Scenario 2: Physical exam and CXR

I'm willing to bet those weren't your selections for each one.

I recently gave a talk on outpatient CHF management and the answer to these scenarios surprised me. I have thought for many years now that the CXR was something that docs ordered blindly without reason in suspected CHF cases. Now that may still be the case but at least they have some good data to back them up. At the same time, Pro-BNPs are becoming more and more of a regular occurrence in diagnostic work ups and they're really best suited for only one side of the CHF equation.

Taking a look at the likelihood ratios compiled below from a 2012 AFP article we can see that the strategy for ruling IN CHF is not the same as ruling OUT CHF.

The healthy 32 year old female has a low pre-test probably of having CHF because of the low prevalence in her specific cohort (young and healthy). Remember that prevalence affects likelihood ratios. We're going to want to use strategies with low negative likelihood ratio (-LR) to rule out the small chance she has CHF. As you can see above, the Framingham criteria for heart failure (especially systolic heart failure) as well as a reduced Pro-BNP level (typically < 300) have a low -LR and would be great tests to use in this case. Hard time remembering how to make a diagnosis with the Framingham criteria? Just remember that if you can make $1 between the two lists you have a diagnosis (see below). One final thing I would add is that EKG's are excellent at ruling out systolic heart failure if normal (98% NPV). [1] Just don't rely on them to rule anything in or rule out diastolic heart failure.

Now what about scenario 2? Use the opposite logic. A patient who has a HIGH pre-test probability of CHF (smoker, hyperlipidemia, older) can more accurately be ruled IN for the diagnosis with a higher positive likelihood ratio (+LR). So in this case I would reach for some rather easy physical exam findings (S3, displaced cardiac apex) and the aforementioned CXR. Remember though that the findings you're looking for are interstitial edema (Kerley B lines) and pulmonary edema. As you can see further down the list, cardiomegaly and pleural effusion don't help you as much.

An interesting side note is how low many of the signs and symptoms we normally report are in the list. Crackles, peripheral edema, PND and orthopnea are all rather low on the list for +LR. 

What Saves Lives in Systolic Heart Failure?

Flavor 1: ACE Inhibitors

Flavor 2: Beta-blockers

ACE inhibitors and beta blockers share lots of similarities. They both have a substantial amount of old but credible data to back them up. They both have been found to improve function (NYHA classification) in patients with systolic heart failure (HFrEF) . They both show improvement in all-cause mortality. [2,3,4] You should use both of these any chance you get in your HFrEF patients. Long term outcomes aren't affected by which one you start first but some would argue to start the ACE inhibitor first due to beta blockers potentially causing further acute decompensation.

Flavor 3: Mineralcorticoid Receptor Antagonists (Spironolactone)

Once your patient is already titrated up on the above medications you should begin thinking about additional therapies if their function isn't optimized (NYHA 2-4). MRA's have been shown to improve all-cause mortality, hospitalization rates, and function status in those already on medical therapy. You want to remember to check the potassium since it's potassium sparing. American Heart Association (AHA) has the unrealistic monitoring schedule recommendation of 2-3 days after starting therapy, 1 week after, then monthly for 3 months then every 3 months after that. Just food for thought.

Flavor 4: ARNI's

Angiotensin Receptor-Neprilysin Inhibitors. Mouthful. These guys are a combo drug with Sacubitril and Valsartan. You already know what Valsartan does. Sacubitril is the Neprilysin inhibitor. What is Neprilysin? It's an endopeptidase that cleaves vasoactive peptides such as BNP. The rationale is to inhibit the inhibitor to ensure a healthy amount of BNP is floating around to vasodilate. Newer RCTs have found a reduction in all-cause mortality and hospitalization rates BEYOND ACE inhibitors. [5] The indications can be somewhat narrow though. Recommendations are to consider ARNI's if the following criteria are met:

  • LVEF < 40%
  • Elevated BNP or heart failure hospitalization in past year
  • SBP > 100
  • GFR > 30
  • Those who have tolerated ACE inhibitor in the past

If that sounds like one of your patients this might be a consideration but just remember that ACE inhibitors have years of good data showing improvement in patient centered outcomes while ARNI's haven't had quite that amount of data to back it up.

Flavor 5: Cardiac Devices

I won't go into too much detail here as placing ICDs and CRTs don't quite fit my scope of practice. But it's important to remember that cardiac devices have been shown to improve outcomes in certain subgroups of patients. Specifically, they have been shown to improve all-cause mortality compared to medical therapy [6]. These results have ranged anywhere between 28-53% improvement based on which type of device was studied. CRT-D was found to have the largest mortality benefit. What is CRT-D vs CRT-P vs ICD? Good question since I didn't know the difference myself beforehand. Here's the breakdown:

  • ICD: Lead in right atrium and right ventricle. Used to shock abnormal rhythms.
  • CRT-D: Lead in right atrium, right ventricle and left ventricle to help synchronize the pumping of blood to the rest of the body. These include an ICD as well.
  • CRT-P: Same as the CRT-D except without the ICD.

What patients receive the most benefit? The 2012 AHA recommendations are divided by classification of recommendations (1-3) and level of evidence (A-C). Here are the Class 1, Level A recommendations:

Quick Shout Out

I didn't discuss treatment modalities that improve morbidity but there are a number of these as well for systolic heart failure. These include digoxin, immunizations (flu, pneumococcal), diuretics, and cardiac rehabilitation. There was subgroup analysis from the DIG trial previously that did show some mortality benefit for digoxin when those patients were within the 0.5-0.8 ng/mL range for serum digoxin concentration. The downside is that when this level increased above 1.2 there was actually an increase in mortality rates. It's also just a tricky medication to use overall that requires close monitoring.

You Forgot Diastolic Heart Failure (HFpEF)!

You're right. I didn't mention diastolic heart failure since there is no substantial evidence to show any mortality benefit in any specific treatment option. Exercise training is the only intervention to show improvement in quality of life and exercise capacity [7]. Unfortunately neither Medicare or Medicaid cover this for diastolic heart failure. So we're stuck with symptom management and contributing factor management for treatment. 

So remember next time you want to rule IN someone with a high likelihood of CHF go for the CXR and cardiac exam. If you want to rule OUT someone with a low likelihood of CHF go for the Pro-BNP, Framingham Heart Failure Criteria, and potentially an EKG. Once you make that diagnosis you can reach for your five lifesaver flavors to help improve mortality rates for your patients.

&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;&nbsp;Click Me!

                                            Click Me!

Quick plug for my new podcast Greyscale. It can be found on iTunes or anywhere you like to get your podcasts. Two episodes currently up with a third on the way. I'll also be launching a Sports Medicine Podcast (The Break) in the near future as well!

BNP: Remember to Stretch Before Dying

murica copy.jpg

Welcome to the The Bad Humors. This is the first in a series of unsuccessful articles that you’ll have the privilege of likely not reading. As well as the first article in the Second Amendment Mini-Series: "The Shotgun Approach." I was feeling patriotic. To know more about what spurred on this entire project, feel free to go to the About page. Now that I got that out of the way you can enjoy the following word salad.

With the reboot of Arrested Development’s (AD) fourth season, my wife and I have been re-watching the entire series to rehash our memories. It’s been great and has reinvigorated my repertoire of AD references that I’ll insert into conversions. Such as the Charlie Brown feeling I used to get my intern year on OB when a triage patient would decline a male provider. However, this also has had unintended consequences. Specifically, Amanda’s hatred of the “British Eyes Only” audio clip that played throughout the third season. Unfounded disdain but somehow understandable. I tend to agree with her as I also get a familiar sense of unexplainable annoyance when I hear it as well. And it is the best way I can explain my relationship with BNP.

BNP - Brain Natriuretic Peptide – first started gaining notoriety in the late 1980’s/early 1990’s alongside it’s now rather defunct cousin Atrial Natriuretic Peptide (ANP). Both are thought to be equally important in counteracting the RAAS and sympathetic activity, effectively limiting sodium retention and vasoconstriction in heart failure. Its use is preferred over ANP as it has a larger range of variability and is thus able to pick up smaller changes. I like to think of BNP like the Dark Knight - the lab that heart failure deserves, but not the one it needs right now. The “right now” part is the thing I have the most beef with, as it is all too convenient get it “right now” as well as trend it. Here are the 3 uses of BNP I’ll be evaluating with extreme prejudice. 

1. BNP trending – Does it make a clinical difference?

2. BNP use in heart failure diagnosis – I mean, how good is it really?

3. BNP in differentiating heart failure and other causes of dyspnea – I actually don’t mind this one.

Question 1: Let’s start with the first one since it’s the use I hate the most, BNP trending. It really makes me cringe in the hospital when I hear anyone say “the BNP is 'blank' trending thus…”. I honestly don’t care what is said after the “thus” and typically will judge someone silently. I’m a big fan of the clinical exam when it comes to a lot of things and progression of heart failure (HF) is definitely one of them. But what does the evidence say? Well, I’m happy to report the evidence shows no difference in every meaningful clinical outcome (30 day mortality, in-hospital mortality, 30 day readmission, and mean length of stay) with trending BNPs vs monitoring clinically. [1] We have the REDHOT II (N=447) trial to thank for that which is really the only RCT that has evaluated BNP trending for acute HF [1]. Until something else comes along to prove otherwise, that’s enough evidence for me to continue to hate on serial BNP trending in the hospital.

But being a Family Physician, I would be remiss if I didn’t address the use of BNP in the outpatient setting. This is something I have personally never done and looking at the evidence opened my eyes a bit on its effectiveness in that setting. There is great, level 1 evidence showing mortality benefit with the use of BNP-guided treatment in HF. This primarily comes from a meta-analysis of 8 RCTs in 2010 which found the risk of all-cause mortality was significantly lower in the BNP-guided group (RR 0.76, 95% CI, 0.63-0.91) but no difference in all-cause hospitalization and mortality benefit in a subgroup analysis of patients older than 75 years. [2] There’s one caveat though in that the BNP-guided groups were found to have more ACE-I and beta-blocker treatment (21/22% vs. 11/12%). Hey, that’s still pretty good. I think I’ll start considering BNP as outpatient  guidance in my patient’s younger than 75 years.


mounds1105 copy.jpg

Question 2: Building on the momentum of my new BFF REDHOT, let’s look at the evidence regarding the use of BNP for the actual diagnosis of HF. In my vast experience of 14 months of MD'dom, I’d say clinicians are typically better about using the BNP properly in this scenario. That is, using it to rule OUT HF but not to rule IN HF. And they’re right. There are Mounds of evidence in supporting the good sensitivity and poor specificity of the test. Renal failure, valvular and coronary heart disease, pulmonary hypertension and sepsis can all elevate BNP. One systematic review from 1989-2005 found pooled sensitivity of 92% and specificity of 65% for diagnosis of heart failure. [3] A different systematic review also in 2006 found a bit more scattered results but with general same trend with sensitivities ranging from 20-100% (Seriously? Overall most studies were >80%) and specificities ranging from 47-89%. There were several others before these studies that corroborated the number crunching as well [4]. So takeaway from this is that you can continue to be confident in BNPs (standard is usually < 100 pg/ml) when ruling out HF. Just don’t diagnose heart failure with it.


Question 3: For the three of you that are still with me (hi Mom), let’s polish off this last question: How useful is BNP in differentiating HF from other causes of acute dyspnea? The answer lies in the masterfully titled “Breathing Not Properly” study in 2002 (N=1586) [5]. This is probably the most cited and famous of the BNP studies. It was conducted in ED and urgent care settings where they looked to distinguish between HF and pulmonary causes of dyspnea in the acute setting. 

What they ended up finding in the study was a pile of obvious. So obvious, one may even… Snicker at it (had to). Plasma BNP was markedly higher in patients with clinically diagnosed HF than those without (675 vs. 110 pg/mml). Sensitivity, specificity, and predictive accuracy were on par with other studies (90, 76, and 83%). While choosing higher values such as >125 and >150 pg/ml decreased sensitivity, increased specificity, and did not change overall predictive accuracy. [5] They also compared it to other modalities (CXR, history of HF, and rales on exam) and were found to be equivalent to or better. The most interesting finding from the study was the death-match between plasma BNP vs. clinical judgment. No surprise here as BNP was more sensitive (90 vs. 49%) but less specific (73 vs. 96%) than clinical judgment. The lessons learned for me here is that again, <100 pg/ml works for me when ruling out HF. Also, if I was stuck in the very common conundrum of “is this a COPD or a CHF exacerbation?” then I would need a rather high BNP to push me into the CHF camp. For my own clinical practice, I will probably use the 675 pg/ml average they found as my bar.

So what did I learn from all this? There is an overwhelming amount of literature on BNP. What I wrote was only the tip of the iceberg with only what I believe to be the very most relevant data to answer my questions. I also learned that I can rest assured that I can continue to ridicule many of the in-hospital BNP use for the above reasons but at the same time incorporate BNP into my outpatient practice. Finally, writing this reminded me that I need to quit writing this and finish the 3rd season of Arrested Development again.

Feel free to blast me in the comments. 

TL; DR - Don't trend BNP inpatient but can consider doing it as an outpatient. Stop using it to diagnose heart failure if you were doing it before.