Exercise Systolic Reserve and Exercise Pulmonary Hypertension Improve Diagnosis of Heart Failure With Preserved Ejection Fraction

Figure 1. Study flow. We screened 1,126 unique patients who underwent simultaneous exercise echocardiography and cardiopulmonary exercise test. Patients who underwent subsequent exRHC (n = 22) were included in the exRHC cohort. Patients with unexplained dyspnea were included in the DST cohort (n = 326). We excluded healthy subjects (n = 172), patients with incomplete data (n = 62) and patients with a baseline reason for dyspnea (n = 544). Note that some patients had more than 1 reason for dyspnea. Valvular heart disease was defined as more than mild valvular stenosis, more than moderate left-sided valvular insufficiency, or previous valvular surgery. DST, diastolic stress test; exRHC, exercise right heart catheterization; LVEF, left ventricular ejection fraction; PAPm, mean pulmonary artery pressure.

Figure 2. DST parameters associated with elevated exPAWP. Results of non-invasive septal E/e' (A), S' (B), systolic PAP (C), cardiac index (D), and mean PAP/CO slope (E) in the exRHC cohort at rest and peak exercise. Red: patients with elevated exPAWP (n = 14), green: patients with normal exPAWP (n = 8). P-values from linear mixed models. *p < 0.05 in multiplicity-adjusted comparison of peak values. exPAWP, pulmonary artery wedge pressure during peak exercise; PAP; pulmonary artery pressure.

Figure 3. Proposed decision tree for diagnosis of HFpEF on DST. (A) Derivation of the decision tree in the exRHC cohort. Step 1: the existing approach using exE/e' is maintained. Step 2: exS' and mPAP/CO slope are determined, HFpEF is considered high probability if exS' <9.5 cm/s and mPAP/CO slope ≥3.2 mmHg/L, and low probability if both parameters below these thresholds. Thus, all patients with exPAWP ≥25 mmHg are identified. A single patient is false positive using this approach. (B) Application of the decision tree to the non-invasive DST cohort. Of 291 patients with normal exE/e', 116 (40%) had exS' >9.5 cm/s and mPAP/CO slope <3.2 mmHg/L, we propose that probability of HFpEF is low in these patients. A total of 50 patients (17%) had low exS' and elevated mPAP/CO slope, we propose that probability of HFpEF is high in these patients. In the remaining 125 patients, we propose to perform additional investigations before establishing a diagnosis of HFpEF. CO, cardiac output; exE/e', highest septal E/e' recorded during exercise; exS', S' at peak exercise; mPAP, mean pulmonary artery pressure.

Figure 4. Performance of decision tree in the DST cohort. To evaluate the performance of the decision tree (Figure 3), the probability of HFpEF according to the decision tree (high/intermediate/low) was compared to surrogate HFpEF indicators peak VO2 (A) and logistic H2FPEF score (B), which calculates the probability of elevated exPAWP in percentage through clinical and echocardiographic parameters (17). Multiplicity-corrected P-values from Kruskal-Wallis test. HFpEF, heart failure with preserved ejection fraction; Peak VO2, peak oxygen uptake, *p < 0.001 compared to low; p < 0.001 compared to intermediate.