FACTORS OF NON-INVASIVE VENTILATION (NIV) FAILURE IN PATIENTS WITH ACUTE HYPOXEMIC RESPIRATORY FAILURE DUE TO SEVERE COMMUNITY ACQUIRED PNEUMONIA (CAP)

Abstract

Background: NIV is recommended as a first line of treatment for acute hypercapnic respiratory failure even in patients with acidosis. On the other hand, experts have a controversial opinion when it comes to a NIV trial for acute hypoxemic respiratory failure. Most of them don’t recommend NIV in severely hypoxemic patients because many studies report failure rates from 20 to 70,3% in this particular setting. Over the years, the use of NIV for acute hypoxemic respiratory failure has increased and the failure rates have dropped, mainly because clinicians make better patient selection and they are more aware of the factors, indicating pending NIV failure.Aim: The aim of our study is do determine the NIV failure rate in a cohort of patients with severe CAP, treated in an intensive care unit (ICU) of a specialized center for pulmonary diseases and to study the factors that are associated with NIV failure.Materials and methods: We studied a prospective cohort of 56 patients with severe CAP that developed acute hypoxemic respiratory failure and were put on NIV. 15 of them had pneumonia without ARDS; 9 – mild, 24 – moderate and 8 – severe ARDS. All of them were ventilated with pressure-supported modes (S, S/T, AVAPS) or CPAP only, taking into account the protective ventilation strategy. We recorded the patients’ age, CURB 65 and SAPS II score on admission and their heart rate (HR), respiratory rate (RR) and parameters of oxygenation, obtained from an arterial blood-gas analysis (ABG) on admission, 1 h and 24 h after initiation of NIV. Then we compared those parameters between patients that succeeded and those that failed an initial NIV trial.Results: Of all 56 patients, undergoing a NIV trial, only 8 (14%) failed and were intubated. 5 of them died in the ICU and the other 3 were extubated successfully. The reasons for NIV failure were: insufficient correction of hypoxemia in 6 patients, large leak in 1 and delirium in 2. After conducting a Mann-Whitney U test, we found statistically significant differences in age (median: 56,5; IQR: 18,5 vs. median: 67,5; IQR: 26,5; p=0,027), PaO2/FiO2 on the 1st (median: 161; IQR: 81,47 vs. median: 120,88; IQR: 50,13; p=0,039) and 24th hour (median: 183,56; IQR: 71,45 vs. median: 118,18; IQR: 56,47; p=0,011) after ventilation onset and HCO3 on admission (median: 23,59; IQR: 5,23 vs. median: 18,6; IQR: 7,15;p=0,006), on the 1st (median: 24,5; IQR: 5,33 vs. median: 20,35; IQR: 6,78, p=0,013) and 24th hour (median: 25,45; IQR: 7,13 vs. median: 21,6; IQR: 4,4; p=0,01) after ventilation onset between the groups of NIV success and failure. To investigate the strength of association between these parameters and NIV failure, we conducted a Kruskal-Wallis H statistical analysis and computed the correlation coefficient of Cohen W. It showed that all of the above listed factors have a strong association with NIV failure.Conclusion: In severe CAP with or without ARDS, causing acute hypoxemic respiratory failure, NIV can be a safe option for respiratory support with close monitoring of PaO2/FiO2 and HCO3, which may indicate upcoming failure.