Echos of Failure / Failure of Echos
(Heart Failure)
Congestive heart failure has been on our front burner for a long time, for both inpatient and outpatient care, and for good reason. The complexities of the diagnosis and treatment are vast but still exceed by the frustrating outcomes everyone is experiencing nationwide with this disorder. CHF still has too-high readmission rates, frequent and complex comorbidities, high mortality rates, high cost, medication complexity, poor quality of life, and a very high prevalence. Seemingly, the landscape has been reorganized and become more rational over the last several years with increasing focus on a recognized duality of pathophysiologies – heart failure with reduced ejection fraction (HFrEF) and heart failure with preserved ejection fraction (HFpEF).
The good news is that we now have strong evidence of 4 different medication groups that are all indicated for patients with HFrEF. Ideal care is to have all HFrEF patients taking optimal doses of preferred agents in each of these 4 classes. The bad news is we do not use these therapies often enough or in sufficient doses because the medical community’s uptake of these consensus modalities has been slow. As a system, we’re working actively to fix part of the problem, and increase optimal medication usage.
Additional good news is that there are now some better insights into treatment of diastolic heart failure or HFpEF. The bad news is that there is more heterogeneity within this group and a less straightforward recommended therapeutic approach for HFrEF.
The ongoing enormous national effort to better evaluate and treat heart failure patients hinges on a single quantifiable linchpin – the ejection fraction – to direct appropriate therapy. Consensus is that the EF can be characterized as normal (50% or more), reduced (< 40%), or borderline (40-50%). The good news is that we have multiple readily available measures of ejection fraction used locally including echocardiograms, MUGA scans, cardiac catheterization when performed, and nuclear stress testing. Other areas sometimes use MDCT and MRI, but neither of these is commonly used here. The bad news is that this cornerstone of every treatment decision we made for our heart failure patients too often appears to be just plain wrong.
For starters, there is no recognized gold standard among these various modalities. Whenever research studies are done to determine any test’s accuracy, any such diagnostic is compared to the landmark reference test, but such a gold standard does not exist for EF measurements. Each of the available tests has well-recognized limitations and caveats as to how they should be interpreted. These testing issues prevent any one of them to be considered the “go to” test that will provide a definitive answer. Therefore, whenever we discover a discrepancy between different modalities, we can’t tell which one may actually be the right answer. This is concern number 1.
It is unfortunately also clear that this is not a theoretical consideration, but these modalities do not always agree. Sometimes the discrepancies between these results can be significant. We have recently encountered situations where – by coincidence and not by intentional investigation – echocardiography differed from results of both angiography and nuclear stress test EF by nearly 30%. Not a relative difference of 30%, but an absolute difference of 30%. This is clearly a very big deal. In each of these cases the results were interpreted correctly when the studies were re-reviewed, so there were no issues with respect to the test interpreters. These examples vexingly showed that the nature of each test’s limitations means that there will sometimes be erroneous results, and sometimes with a very large degree of error. This follows the concerns already identified in a landmark paper by Pellikka at all – JAMA Network Open. 2018;1(4):e181456. doi:10.1001/jamanetworkopen.2018.1456 -that looked at over 2,000 patients from 127 sites in 26 countries entered in the STITCH trial. They noted that approximately one quarter of the patients would have been classified differently (HFpEF vs. HFrEF) in this study if a different measure of ejection fraction had been used. That would presumably have significantly changed the assessments and conclusion in the trial. It would also have significantly changed the therapeutic decisions for a large number of patients. They defined a variation between results of 5% or less as being acceptable when 2 different modalities were compared. Approximately half the comparisons fell within that range, meaning about half of the time there was a clinically significant difference between the results when more than 1 modality measuring EF was used. We are seeing the same thing in our patients that this study highlighted so meticulously. This is concern number 2.
Right now we are treating heart failure patients and classifying them pathophysiologically based on their recorded ejection fractions. There is no rule or even clinical suggestion as to which measure of EF should be used. We now have reason to suspect that many of our patients may be categorized incorrectly with respect to HFrEF vs. HFpEF, as has occurred in international studies done under controlled research circumstances. Both inpatient treatment for those hospitalized with heart failure and outpatient care for this group is routinely – and wisely – based on this their HF classification. So is our coding, comorbidity classification, core measures, quality outcomes, and many other aspects of providing care to heart failure patient. Does this mean that we have to be nihilistic about treatment, order redundant and confirmatory testing, or change our management of our patients? Time to pause before we go down any of these unwanted paths.
We are best served by understanding this discrepancy phenomenon, which appears to be underappreciated and infrequently addressed, as per our own documentation in our charts. The Chief of our Division of Cardiology, Dr. Mohan Madhusudanan, has emphasized that the most important thing is clinical judgment, particularly on the part of any treating cardiologist. It is important to interject such medical judgment to classify the patient not based exclusively on the available EF results – most commonly measured by echocardiography – but also to appreciate whether the patient’s course and clinical condition appears to fit their classification. We are familiar with such circumstances and both our cardiologists and the rest of us need to focus more on the clinical correlation, and document our thoughts carefully. For non-cardiologists, it is also incumbent upon us to have the appropriate discussion with our treating Cardiologist so that there is good understanding of how and why the Cardiology Subspecialist classifies the patient into one of these 2 defined categories.
A second step is to look carefully at the report itself. Often we are tempted to have our eyes read the ejection fraction, blurring out much of the rest of the report, particularly with echocardiograms. However sometimes the report will have useful additional information such as right ventricular function, PA pressure, valvular stenoses, and technical factors limiting obtaining appropriate windows, which may raise red flags that the patient’s left ventricular function needs more careful reconsideration to assure appropriate characterization.
No one is suggesting that each study should be reaffirmed by follow-up testing using a different modality. Of note is that a common practice of oncologists using cardiotoxic chemotherapeutic agents is to use sequential MUGA scans because they believe that these give the most consistent and reproducible results, particularly when measured sequentially over time. However, since this is a high cost and high radiation testing strategy, it is certainly not suitable for most of our patients, and MUGA scans also have their limitations with respect to being a true gold standard. This would never work in an inpatient setting to the prohibiting effects on turnaround time it would create.
A similar pattern is seen in the inpatient and outpatient settings when sequential evaluations of left ventricular function show a clinically significant change in function. We use such change (or lack of change) over time to prognosticate with patients, make therapeutic treatment recommendations, and make multiple ongoing clinical decisions. Certainly whenever sequential studies suggest there has been a significant change, then our thinking caps must go on to make sure we are confident of the available data, or see if reevaluation is warranted.
One additional step that behooves us as hospital providers is the critical importance of the lost art of good communication. Our hospitalized heart failure patients are handed off to our outpatient colleagues and they use the discharge summary and the information provided therein as a cornerstone of the continuation of the patient’s therapy after discharge. A direct discussion with the outpatient provider is the best care model, but that currently happens to infrequently. We need to be as clear as possible about the patient’s pathophysiologic classification, the basis of that classification, and any additional pertinent comments about the confidence of that result or any concerns that have been identified about the classification that may warrant additional outpatient follow-up so the patient is classified correctly, and the appropriate treatments are targeted.
EF. Once simple, now maybe not so simple. But keep the faith, we can still proceed in the same fashion we have, but with increased attention to something we all really knew all along. No single test results substitutes for clinical judgment. Use that judgment, and our patients will do better.
