May 2016

Identifying Low-Risk CHF Patients for ED Discharge


You’re working an average shift in your local community ED when you encounter Mr. P, a fifty-five year old man with a history of an ischemic cardiomyopathy, coronary artery disease with a stent to his LAD, hypertension, and hyperlipidemia. Mr. P has noted 3 days of increasing dyspnea on exertion and orthopnea, with a moderate increase in lower extremity edema. He reports compliance with his medications, but admits he had a hamburger and fries the night before his symptoms began, and suspects all the salt he ingested has something to do with his decline. He called his PMD with his symptoms who suggested he come to the ED.

He is on room air at rest with an O2 sat of 97%. His BP is 155/95, HR is 73, and RR is 14. He is in no distress, breathing comfortably, with fine rales at bilateral lung bases. He has no cardiac murmurs or gallops, mild jugular venous distension, and 1+ pitting edema confined to the feet and ankles.

You order an ECG, CXR, and labs. His ECG is unchanged from prior, with evidence of LVH but no ischemic changes. His CXR reveals unchanged cardiomegaly and mild pulmonary edema. His labs reveal a creatinine of 1.3 (baseline 0.9), a BUN of 24, and otherwise normal electrolytes. His troponin is < 0.03 and his BNPeptide is 350.

Convinced that you are dealing with a CHF exacerbation, you order the patient a dose of IV furosemide and put in an order to admit to the hospitalist. The hospitalist calls back and suggests that perhaps the patient is low-risk enough that he could be discharged home on a few extra doses of furosemide. He ends of up accepting the patient, but after your shift you wonder if perhaps the hospitalist was correct. You decide to search the literature for articles addressing the use of observation units or ED discharge in low-risk CHF patients, and for articles to help select which patients can be considered low risk.

PICO Question

Population: Adult patients presenting to the ED with acute, decompensated heart failure.

Intervention: Clinical assessment and/or a clinical prediction rule use to select low-risk patients for observation unit admission or discharge.

Comparison: Hospital admission.

Outcome: Mortality, MI, hospital readmission, hospital length of stay, healthcare costs.

Search Strategy
Articles were selected by an emergency physician with a special interest in heart failure management. No formal search strategy was employed.


Article 1: Rame JE, Sheffield MA, Dries DL, Gardner EB, Toto KH, Yancy CW, Drazner MH. Outcomes after emergency department discharge with a primary diagnosis of heart failure. Am Heart J. 2001 Oct;142(4):714-9.

Article 2: Schrager J, Wheatley M, Georgiopoulou V, Osborne A, Kalogeropoulos A, Hung O, Butler J, Ross M. Favorable bed utilization and readmission rates for emergency department observation unit heart failure patients. Acad Emerg Med. 2013 Jun;20(6):554-61.

Article 3: Collins SP, Jenkins CA, Harrell FE Jr, et al. Identification of Emergency Department Patients With Acute Heart Failure at Low Risk for 30-Day Adverse Events: The STRATIFY Decision Tool. JACC Heart Fail. 2015 Oct;3(10):737-47.

Article 4: Auble TE, Hsieh M, Gardner W, Cooper GF, Stone RA, McCausland JB, Yealy DM. A prediction rule to identify low-risk patients with heart failure. Acad Emerg Med. 2005 Jun;12(6):514-21.

Bottom Line

Despite concerns regarding high rates of hospital admission for acute heart failure in the US, and despite the use of such rates as a prevention quality indicator by the Agency for Healthcare Research and Quality (AHRQ), there has been little research into effective strategies to reduce admission rates in the ED. Much of the evidence that exists is related to long-term management strategies to reduce hospital admission, rather than ED-based strategies at identifying patients who do not require admission. The current evidence is limited, and four such articles were reviewed.

Article #1 was an observational study of patients admitted to the observation units of two Emory-affiliated hospitals. They found that among patients in the OU, those who eventually required admission to the hospital had higher overall BNP, BUN, and creatinine levels compared to those discharged home, and the median LVEF was lower in admitted patients. Readmission rates were similar between those who were admitted from the OU and those discharged home.

Article #2 was a retrospective chart review of patients discharged from the ED of Parkland Memorial Hospital in Dallas, TX. They found that “failure of outpatient therapy” (defined as return visit to the ED, hospital admission, or death) occurred within 3 months in 61% of patients. Only elevated respiratory rate was found to be a predictor of treatment failure.

Article #3 and article #4 were both aimed at deriving clinical prediction rules to identify patients at low-risk of adverse outcomes. Both studies were limited by lack of validation, the use of complex rules that would not be clinically useful, and in the case of article #3 a failure to identify a large enough proportion of the population as “low-risk” to have any significant impact.

This body of evidence provides much to stimulate discussion, but unfortunately does not provide any guidance to the emergency physician. To date there is no validated clinical decision to identify heart failure patients at low risk of decompensation who are appropriate for discharge from the ED, and it would appear that clinical factors are not, in fact, very good predictors of failure of outpatient therapy. The first article does at least provide evidence that observation units can help reduce hospital admission and do so without increasing rates of outpatient failure. It seems reasonable to admit low-risk patients to the observation unit for diuresis and blood pressure management, but as to which patients can be discharged, the evidence is unfortunately lacking.