Methods

This study was conducted in conjunction with the Cone Health Advanced Heart Failure Program, the cardiac catheterization laboratory at Moses H. Cone Memorial Hospital and the Department of Kinesiology at the University of North Carolina at Greensboro, Greensboro, North Carolina. All study procedures were concurrently approved by the institutional review boards of Cone Health and the University of North Carolina at Greensboro. All patients provided written informed consent before receiving sedation and any study-related procedures. There was no exclusion based on gender, race, or ethnicity. All patients scheduled for a right-sided heart catheterization with a left ventricular ejection fraction (LVEF) ≤40% within the previous 6 months, with continued signs and symptoms consistent with HFrEF, and having not received intravenous inotropic therapy ≤7 days of the catheterization were considered for inclusion. Patients were excluded if they (1) were determined to have severe lung disease (diagnosed as such in the patient’s medical history or, if spirometry data were available, a forced expiratory volume in 1second ≤1 L and/or a forced expiratory volume in 1 second to forced vital capacity ratio of ≤0.50, or required use of home O ₂ ), (2) were expected to receive >2 mg of midazolam and/or >50 μg of fentanyl, or (3) were expected to need supplemental O ₂ after receiving sedation.

Appropriate medical history was obtained to determine the patient’s candidacy for participation in the study, which included results of the patient’s last LVEF measure and results of the last pulmonary function test (if available). In addition, age, gender, height, weight, and current list of medications were obtained. The results of the catheterization were also obtained after the procedure.

VO ₂ at rest was measured using a commercial, open-circuit, breath-by-breath gas analysis system (Ultima-CPX, MGC Diagnostics Corp., St. Paul, Minnesota). System calibration was performed before each study according to the manufacturer’s specifications. After the completion of the catheterization procedure, while the patient remained in the catheterization laboratory procedure room (supine on the laboratory table), the patient breathed through a mouthpiece with a noseclip occluding nasal ventilation. After a 5-minute run-in (acclimation) phase, sampling was performed for an additional 5 minutes. The reported value of VO ₂ at rest was the average during the 5-minute sampling period.

Estimated VO ₂ at rest was calculated according to each of the 3 widely used empirical formulas listed in Table 1 . In accordance with these equations, BSA was calculated according to the formula by D. Du Bois and E. Du Bois : BSA (m ² ) = 0.007184 × weight (kg) ^0.425 × height (cm) ^0.725 .

All statistical analyses were done with JMP version 10 (SAS Institute Inc., Cary, North Carolina). With the exception of the mean differences, all continuous variables are reported as mean ± SD and categorical variables are reported as the total number (% of total). The mean differences are reported as mean ± SE.

Single-sample t tests were performed to compare the mean of differences (estimated − measured) against the hypothetical mean difference of 0 and was considered statistically significantly different when p ≤0.05.

The Bland-Altman method for comparing methods of measuring the same parameter was used to assess the agreement between measured and estimated VO ₂ . In accordance with this method, % error of the estimation of VO ₂ at rest ([estimated − measured]/measured) for each patient was plotted against the corresponding average of the measured + estimated values. The limits of agreement were the mean difference ± 1.96 SD, and poor agreement was considered when the limits of agreement exceeded ± 25% error.

Two different analyses were used to determine if differences between the measured and estimated VO ₂ were clinically significant. First, the absolute % error in estimated VO ₂ was calculated by dividing the absolute value of the difference (estimated − measured) by the measured VO ₂ . An absolute error ≥25% was considered the cut point for clinically significant error to occur. Second, the actual clinically significant error rate of the sample was calculated based on the cardiac index (cardiac output/BSA). For each equation, data were plotted with the cardiac index derived from the measured VO ₂ along the x axis and the cardiac index derived from the estimate on the y axis. Break lines were laid over the plots at values representing the clinically important cutoffs for cardiogenic shock (< or ≥1.9 ml/min/m ² ) and hypoperfusion (< or ≥2.2 ml/min/m ² ), resulting in plots with 4 quadrants: 2 representing no clinically significant error (classification by the estimated VO ₂ = classification by measured VO ₂ ) and 2 representing clinically significant error (classification by the estimated VO ₂ ≠ classification by measured VO ₂ ). For each level, the number of patients falling in the clinically significant error quadrants for cardiac index was summed and divided by the total sample.