Can non-invasive ventilation failure be predicted with pulse oximetry in COPD patients?

Introduction: In individuals presenting with an acute exacerbation of Chronic Obstructive Pulmonary Disease (COPD), the excessive workload imposed on the respiratory muscles due to heightened inspiratory demand results in muscular fatigue, leading over time to a reduction in respiratory effort and the development of hypercapnia. Noninvasive Mechanical Ventilation (NIMV) stands as the cornerstone treatment modality in the management of these patients. Nevertheless, NIMV proves ineffective in approximately one-fifth of cases. Delayed initiation of endotracheal intubation exacerbates the risk of prolonged intensive care unit stays and hospitalizations, along with an increased likelihood of in-hospital mortality. Therefore, the early prediction of NIMV failure assumes critical importance. Objective: Fluctuations in pleural pressure occurring throughout the respiratory cycle give rise to similar oscillations in arterial pulse. Pleth Variability Index (PVI) is defined as the variability in the pulse oximeter waveform during respiratory oscillations and is correlated with changes in pleural pressure. In fact, it serves as a noninvasive indicator of respiratory effort. The purpose of our study is to investigate whether the changes in PVI during the first 1-2 hours of Noninvasive Mechanical Ventilation (NIMV) application have a predictive value for early NIMV failure. Material and Methods: Between June 2022 and December 2022, a total of 56 patients diagnosed with Chronic Obstructive Pulmonary Disease (COPD) and suffering from hypercapnic respiratory failure were enrolled in our study at the Pulmonology Department. These patients underwent Noninvasive Mechanical Ventilation (NIMV) therapy. Arterial blood gas analysis, vital parameters, Glasgow Coma Scale scores, and Pleth Variability Index (PVI) values were recorded both at the initiation of NIMV treatment and 1-2 hours after its commencement. NIMV failure was defined as the requirement for endotracheal intubation within 24-48 hours. Results: A positive correlation was observed between PVI change (1-2 hours after NIMV application) and pH change (r=0,448), while a negative correlation was found between PVI change and pCO2 change (r=-0,499) (p<0,001, p<0,001). Logistic regression analysis performed to predict NIMV failure revealed that the initial Glasgow Coma Scale score, a pulse rate exceeding 100 beats per minute measured after NIMV (1-2 hours), and the absence of an increase in PVI compared to its baseline value after NIMV were associated with a 92% Area Under the Curve (AUC) value in predicting NIMV failure. Conclusion: This study demonstrates that PVI is closely associated with changes in pH and pCO2, suggesting its potential relevance in predicting NIMV failure.