By Van Selby, MD

Assistant Professor of Medicine, University of California, San Francisco Cardiology Division, Advanced Heart Failure Section

Dr. Selby reports no financial relationships relevant to this field of study.

SYNOPSIS: In patients with unexplained dyspnea, a score based on six noninvasive criteria can predict the likelihood of heart failure with preserved ejection fraction.

SOURCE: Reddy YNV, Carter RE, Obokata M, et al. A simple, evidence-based approach to help guide diagnosis of heart failure with preserved ejection fraction. Circulation 2018 May 23. pii: CIRCULATIONAHA.118.034646. doi: 10.1161/CIRCULATIONAHA.118.034646. [Epub ahead of print].

Exertional dyspnea is a frequent complaint among patients referred to cardiology clinics. Heart failure with preserved ejection fraction (HFpEF) is common among such patients, but the diagnosis can be challenging without invasive testing. Reddy et al sought to develop a simple risk prediction score based on readily available clinical data. They retrospectively analyzed a cohort of 414 consecutive patients with unexplained dyspnea who were referred for right heart catheterization with exercise testing at the Mayo Clinic. A diagnosis of HFpEF was confirmed if the patient demonstrated a pulmonary arterial wedge pressure ≥ 15 mmHg at rest or ≥ 25 mmHg during exercise. Multivariable logistic regression was used to identify clinical variables associated with the presence of HFpEF.

Of the 414 patients studied, 267 were diagnosed with HFpEF, and the rest were diagnosed with noncardiac dyspnea. The clinical predictors in the final model were obesity, atrial fibrillation, age > 60 years, treatment with two or more antihypertensive medications, an E/e’ ratio > 9 on echocardiography, and a pulmonary artery systolic pressure > 35 mmHg on echocardiography. The authors developed a weighted composite (the H2FPEF score) using these six predictors, with scores ranging from 0-9. The score was strongly associated with the likelihood of HFpEF, with the odds of HFpEF doubling for every one-unit increase in score (odds ratio, 1.98; P < 0.0001). The area under the curve (AUC) for predicting HFpEF was 0.841 (P < 0.0001), with a score of 1.0 representing a perfect test. The H2FPEF score was validated in a separate cohort of 100 patients with similar performance (AUC, 0.886; P < 0.0001).

The authors concluded the H2FPEF score enables discrimination of HFpEF from noncardiac causes of dyspnea and is useful in the evaluation of patients with unexplained exertional dyspnea.


HFpEF can be challenging to diagnose when obvious signs and symptoms are absent. Right heart catheterization with exercise is considered the gold standard for diagnosing HFpEF. However, given the invasive nature and required resources, it is not feasible to refer all patients with suspected HFpEF for such testing. Using data from patients referred to the Mayo Clinic catheterization laboratory for evaluation of unexplained dyspnea, Reddy et al developed an easy-to-use score that predicts the likelihood of HFpEF in such patients using data obtained during routine clinical evaluation.

The H2FPEF score is simple to calculate: 3 points for atrial fibrillation, 2 points for obesity, and 1 point for each of the other criteria. A score of 0-1 was considered sufficient to eliminate HFpEF. High scores (6-9) usually can establish the diagnosis of HFpEF with reasonable confidence. In patients with scores of 2-5, a diagnosis of HFpEF can neither be confirmed nor ruled out, and further testing (such as invasive exercise study) should be considered. The authors also provided a more sophisticated calculator (available in the supplementary material published with the article) that can be used if even more precise risk estimation is desired. The H2FPEF score performed markedly better than other available algorithms for determining the likelihood of HFpEF. For example, when compared to criteria proposed in expert guidelines from the European Society of Cardiology (primarily based on natriuretic peptide levels and echocardiographic parameters), the H2FPEF score performed substantially better in this Mayo cohort (increase in AUC of +0.169; P < 0.0001).

Although the test has advantages over other risk prediction tools, there are important considerations to keep in mind when using the test. Even in patients with a score of 1 (low risk), the prevalence of HFpEF was approximately 20%. When suspicion for HFpEF persists and an alternative explanation for a patient’s exertional dyspnea cannot be identified, it is reasonable to consider invasive exercise hemodynamic testing. The test performs much better when the score is high; more than 90% of patients with scores above 5 were confirmed to have HFpEF by invasive testing.

The H2FPEF score was derived only using data from patients treated at the Mayo Clinic, and it is unclear how the same test would perform outside of this highly specialized, tertiary referral setting. The final model contained multiple variables that are well-known predictors of HFpEF (i.e., echocardiographic evidence of diastolic dysfunction), but left out other predictors that are previously shown to predict HFpEF (such as natriuretic peptide levels). The authors did not find a strong enough association to include NTproBNP level in the final score, but NTproBNP data were missing for 24% of patients. Does this mean we should ignore NTproBNP levels when evaluating a patient for HFpEF? Probably not.

Tools like H2FPEF cannot replace clinical judgment. In patients with clear signs and symptoms of vascular congestion, a diagnosis of HFpEF can be made regardless of the score. Similarly, in patients with a clear alternative etiology for dyspnea, the score should not distract from the obvious causes.

Despite its limitations, the H2FPEF score provides a useful framework for estimating the likelihood of HFpEF in select patients with exertional dyspnea that remains unexplained despite appropriate initial evaluation.