Heart Failure: Practice-Changing Developments for Hospitalists

Bruce Jancin

September 09, 2020

A recently validated, easy-to-use calculator of predicted 7-day mortality risk in patients presenting with acute decompensated heart failure is well worth incorporating into hospitalist clinical practice, Dustin T. Smith, MD, said at HM20 Virtual, hosted by the Society of Hospital Medicine.

The risk prediction tool, called the Emergency Heart Failure Mortality Risk Grade (EHMRG), can help guide clinical decision making as to whether a patient presenting with acute heart failure is appropriate for early discharge or should instead be admitted for inpatient monitoring and more aggressive therapy, explained Dr. Smith, a hospitalist at Emory University in Atlanta.

In addition to the EHMRG, other highlights of his wide-ranging update on recent practice-changing developments in heart failure directly relevant to hospitalists included the introduction of a simple, evidence-based tool for differentiating heart failure with preserved ejection fraction from other potential causes of unexplained dyspnea on exertion in euvolemic patients, and a study debunking what has been called the potassium repletion reflex in patients with acute heart failure undergoing diuresis.

The ACUTE Study

Heart failure is an area of special interest for Dr. Smith. He has been surprised to find that virtually no hospitalists, emergency medicine physicians, or cardiologists he has spoken with have heard of the EHMRG or its validation in the ACUTE (Acute Congestive Heart Failure Urgent Care Evaluation) study. Yet this is a very handy tool for hospitalists, he observed.

The EHMRG algorithm utilizes nine variables for which data is readily available for every patient who arrives at the emergency department with acute heart failure. The variables are age, arrival by ambulance, heart rate, systolic blood pressure, potassium level, oxygen saturation, troponin, serum creatine, and presence or absence of active cancer. The information is entered into a cell phone app, which spits out the patient's estimated 7-day mortality risk. The algorithm divides patients into one of five risk groups ranging from very low to very high. With the addition of data input as to the presence or absence of ST-segment depression on the 12-lead ECG, the weighted algorithm will simultaneously generate an estimated 30-day mortality risk.

ACUTE was a prospective, observational, real-world validation study of EHMRG involving 1,983 patients seeking emergency department care for acute heart failure at nine Canadian hospitals. The actual 7-day mortality rate was 0% in the very-low-risk group, 0% in the low-risk group, 0.6% with an intermediate-risk EHMRG, 1.9% with high risk, and 3.9% in the very-high-risk group. The corresponding 30-day mortality rates were 0%, 1.9%, 3.9%, 5.9%, and 14.3%.

The University of Toronto investigators also asked participating physicians for their clinical estimates of 7-day mortality risk while blinded to the EHMRG predictions. The algorithm proved more accurate than physician predictions across the board. Indeed, physicians consistently overestimated the mortality risk for all categories except the very-high-risk one, where they underestimated the true risk (Circulation. 2019 Feb 26;139[9]:1146-56).

Given that heart failure remains year after year at the top of the list of most frequent causes for hospital admission, and that there is compelling evidence that many low-risk patients get hospitalized while potentially unsafe early discharges also occur, the EHMRG score fills an important unmet need.

"I think this can help inform us as to who with acute heart failure potentially needs to come into the hospital and who doesn't," Dr. Smith said. "I think the sweet spot here is that if you're in the low- or very-low-risk category, your 7-day mortality is less than 1%; in fact, in this study it's zero. But once you get to category 3 – the intermediate category – you're talking about a 7-day mortality of 1%-2%, which I think is high enough to warrant hospital admission for treatment and to watch them, not just send them home."

The H2FPEF Score

Diagnosis of heart failure with preserved ejection fraction (HFpEF) is a challenge in euvolemic patients with clear lungs and dyspnea on exertion. Investigators at the Mayo Clinic have developed and subsequently validated a weighted score known as the H2FPEF score that's of great assistance in this task. The score is based upon a set of six simple variables universally available in patients undergoing diagnostic workup for the numerous potential causes for dyspnea on exertion. Together these six variables comprise the acronym H2FPEF:

  • Heavy: One point for a BMI greater than 30 kg/m2.

  • Hypertension: One point for being on two or more antihypertensive drugs.

  • Atrial fibrillation: Three points for paroxysmal or persistent AF.

  • Pulmonary hypertension: One point for having a Doppler echocardiographic estimated pulmonary artery systolic pressure greater than 35 mm Hg.

  • Elder: One point for age greater than 60 years.

  • Filling pressure: One point for a Doppler echocardiographic E/e' ratio above 9.

The total score can range from 0 to 9. (Circulation. 2018 Aug 28;138[9]:861-70).

Each 1-point increase in the score essentially doubled a patient's risk of having HFpEF as opposed to pulmonary embolism or some other cause for the dyspnea.

"I really like this H2FPEF score. The score works very, very well. Once you get to a score of 6 or above, the probability of HFpEF is more than 90%, which is pretty powerful. I think this is worthwhile," Dr. Smith said.

In their derivation and validation cohorts, the Mayo Clinic investigators used as their gold standard for diagnosis of HFpEF invasive hemodynamic exercise testing with a pulmonary artery catheter in place to measure pressures. A score that enables hospitalists to lessen the need for that kind of costly invasive testing is most welcome.

"Here's how I'd use this score: With an H2FPEF score of 0-1, HFpEF is unlikely. With an intermediate score of 2-5, additional testing is warranted. If the score is high, 6-9, I think HFpEF is likely," the hospitalist said.

Dr. Smith isn't the only big fan of the H2FPEF score. In an editorial accompanying publication of the score's validation study, Walter J. Paulus, MD, PhD, hailed the H2FPEF score as "a unique tour de force" which constitutes a major advance beyond the confusing diagnostic recommendations for HFpEF issued by the European Society of Cardiology and the American Society of Echocardiography, which he said have been "met by skepticism qualifying them as overcomplicated and even triggered disbelief in the existence of HFpEF."

Particularly interesting were the variables rejected for inclusion in the H2FPEF score because they failed to achieve statistical significance as predictors, even though they're often considered important in defining HFpEF, he noted. These included left atrial volume index, sex, and levels of circulating N-terminal probrain natriuretic peptide, wrote Dr. Paulus, professor of cardiac pathophysiology at VU University, Amsterdam.

Debunking the Potassium Repletion Reflex

Longstanding conventional wisdom holds that patients hospitalized for heart failure need to maintain a serum potassium above 4.0 mEq/L.

"I'm sure you've all written orders to keep the potassium greater than 4.0 mEq/L and the magnesium above 2mEq/L about a million times, like I have," Dr. Smith said.

But it turns out this traditional practice, which involves a huge cost in terms of time, money, and health care resources, is supported by weak evidence – and an important recent study has now debunked what the investigators termed the potassium "repletion reflex."

The investigators at the University of Massachusetts identified 4,995 patients admitted with exacerbation of acute heart failure and a normal admission serum potassium level of 3.5-5.0 mEq/L. More than 70% received potassium repletion at least once within a 72-hour observation window, during which 2,080 patients maintained a low-normal serum potassium below 4.0 mEq/L, 2,326 had a mid-normal level of 4.0-4.5 mEq/L, and 589 had a high-normal level of more than 4.5 mEq/L but not more than 5.0 mEq/L.

The study had three endpoints: in-hospital mortality, transfer to the intensive care unit, and hospital length of stay. After statistical adjustment for comorbidities, demographics, and severity at admission, there was no difference between the low- and mid-normal serum potassium groups in any of the three endpoints. In contrast, the high-normal potassium group had a significantly longer length of stay, by a median of 0.6 extra days. The high-normal group also had a 78% increased likelihood of ICU transfer and a 51% increased risk of in-hospital mortality, although neither of these differences reached statistical significance (J Hosp Med. 2019 Dec 1;14[12]:729-36).

"A potassium greater than 4.5 mEq/L may be associated with increased risk of worse outcomes," Dr. Smith observed. "I think the sweet spot may be 3.5-4.5 mEq/L based on this study."

He reported having no financial conflicts regarding his presentation.

This article originally appeared in The Hospitalist, an official publication of the Society of Hospital Medicine.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.