Further prognostic stratification in patients with oscillatory ventilation with exercise: is there more to it than just ups and downs?

Abstract

Abstract Introduction Exercise Oscillatory Ventilation (EOV) has numerous definitions, but the common foundation is the presence of an oscillatory phenomenon of the ventilation/minute graph with a given amplitude and frequency. Recently, it was proposed that the presence of a delay in O2 consumption (VO2) peak to minute ventilation (VE) peak during ventilatory oscillation was a predictor of worse prognosis in patients with heart failure (HF) and left ventricular ejection fraction (LVEF) <50%. The aim of this work was to assess whether these characteristics add further prognostic value to the subset of patients with HF and EOV. Methods This was a single-centre retrospective cohort of consecutive patients with HF and LVEF <50% that underwent cardiopulmonary exercise testing (CPET) from 2016–2020. EOV was defined as per Vainshelboim 2017 (≥3 consecutive cyclic fluctuations of ventilation during exercise, average amplitude over 3 ventilatory oscillations ≥5L and an average length of three oscillatory cycles 40s to 140s). The presence of EOV was evaluated by 3 independent observers after observing a VE over time plot. For the creation of this plot, data was collected as a rolling average of 5 consecutive breaths. The presence of EOV was established if at least 2 observers agreed on the classification. Afterwards, a second graph was plotted, with both VO2 and VE over time and the mean delay between VO2 peak to VE peak during EOV was manually calculated (Figure 1). The primary endpoint was a composite of time to all-cause death, heart transplantation or left ventricular assistance device (LVAD) implantation. Results Of the 285 patients, 78 (27%) were classified as having EOV. These HF patients were mostly male (75%) with a mean age of 58±12 years, mean LVEF 31±10% with 63% having ischemic etiology. During a median follow up of 27 (17 to 43) months, 4 LVAD implantations, 12 heart transplantations and 18 deaths. The rate of primary outcome was 19% and 36% at 1- and 2-years. The amplitude, frequency, and maximum number of EOV cycles were not associated with the development of the primary endpoint. Only the mean delay between VO2 peak to VE peak during EOV was predictive of time to primary endpoint, even when adjusted for peak VO2 and VE VCO2 slope (adjusted HR 1.06 95% CI 1.009–1.114) (Figure 2). The cut-off of 5 seconds for mean VO2 peak to VE peak delay seems to be the most useful to predict the primary outcome at 2 years, with a sensitivity of 48% and specificity of 84%. Conclusion This novel parameter seems to be the only EOV-related parameter to enable further stratification of prognosis in a cohort of patients with severe HF, with the best cut-off of mean delay between VO2 peak to VE peak during EOV being 5 seconds. Funding Acknowledgement Type of funding sources: None.