Quantifying oscillatory ventilation during exercise in patients with heart failure

Background: This study examined the validity of a novel software application to quantify measures of periodic breathing rest (PB) and oscillatory ventilation during exercise (EOV) in heart failure patients (HF). Methods: Eleven male HF patients (age=53±8yrs, ejection fraction=17±4, New York Heart Association Class=III(7)/IV(4)) were recruited. Ventilation and gas exchange were collected breath-by-breath. Amplitude and period of oscillations in ventilation (V̇E), tidal volume (V_T), end-tidal carbon dioxide (PETCO2), and oxygen consumption (V̇O2) were measured manually (MAN) and using novel software which included a peak detection algorithm (PK), sine wave fitting algorithm (SINE), and Fourier analysis (FOUR). Results: During PB, there were no differences between MAN and PK for amplitude of V̇E, V_T, PETCO2, or V̇O2. Similarly, there were no differences between MAN and SINE for amplitude of V̇E or V_T although PETCO2 and V̇O2 were lower with SINE (p<0.05). In contrast, the PK demonstrated significantly shorter periods for V̇E, V_T, PETCO2, and V̇O2 compared to MAN (p<0.05) whereas there were no differences in periods of oscillations between MAN and SINE or FOUR for all variables. During EOV, there were no differences between MAN and PK for amplitude of V̇E, V_T, PETCO2, and V̇O2. SINE demonstrated significantly lower amplitudes for V_T, PETCO2, and V̇O2 (p<0.05) although V̇E was not different. PK demonstrated shorter periods for all variables (p<0.05) whereas there were no differences between MAN and SINE or FOUR for all variables. Conclusion: These data suggest PK consistently captures amplitudes while underestimating period. In contrast, SINE and FOUR consistently capture period although SINE underestimates amplitude. Thus, an optimal algorithm for the quantification of PB and/or EOV in patients with HF might combine multiple analysis methods.