Non-Invasive Ventilation: a Safe and Effective Respiratory Support Method in Hypoxemic Acute Respiratory Failure Due to Pneumonia with or without Acute Respiratory Distress Syndrome

Introduction: The benefit of non-invasive ventilation (NIV) in cases of hypercapnic acute respiratory failure (ARF) has already been proven. Still, its safety and efficacy as a respiratory support method for patients with hypoxemic ARF hasn’t been studied so well. Aim: The aim of our study was to examine the safety and efficacy of NIV in hypoxemic ARF of primary lung origin. Materials and methods: This was a prospective observational cohort study of patients with hypoxemic ARF due to community acquired pneumonia with or without acute respiratory distress syndrome (ARDS) treated using NIV. They were divided into four groups: pneumonia without ARDS, mild, moderate, or severe ARDS. Their clinical and ABG parameters were recorded before initiation of NIV, at 1 hour and 24 hours after ventilation onset and at transition to non-intensive NIV or before endotracheal intubation in NIV failure cases. Results: A total of 63 patients were included. NIV trial was successful in 85.71% of them, while 14.29% experienced NIV failure. In the general population, we observed a significant difference in PaO2/FiO2 only before transition to non-intensive NIV in comparison to the value at admission. This trend was seen in the patients with pneumonia without ARDS and moderate ARDS, but not in those with mild and severe ARDS. The clinical parameters showed improvement early in the course of treatment both in the entire study population and all subgroups. Conclusions: NIV is an effective and safe option for respiratory support in patients with severe CAP only when an adequate etiological treatment has been applied.


INTRODUCTION
The benefit of NIV for the survival of patients with certain conditions (COPD, cardiogenic pulmonary edema and difficult weaning from invasive mechanical ventilation) has already been proven. 1 However, the safety and efficacy of NIV as a respiratory support method for patients with de novo hypoxemic ARF hasn't been studied very well. To date, only ten randomized controlled trials have been published on this topic. [2][3][4][5][6][7][8][9][10][11] Only two of them include only non-hypercapnic immunocompetent patients and can serve as strong evidence for giving recommendations for the use of NIV in this type of patients. That is why the European Respiratory Society and the American Thoracic Society don't give any recommendations for the use of NIV for de novo hypoxemic ARF in their new 2017 guidelines 12 , making the topic highly controversial.

AIM
To determine the degree of clinical and arterial blood gas (ABG) values improvement during the course of NIV treatment in patients with hypoxemic ARF due to severe community acquired pneumonia CAP with or without ARDS.

MATERIALS AND METHODS
This was a prospective observational cohort study of patients with hypoxemic ARF due to CAP with or without ARDS. It was conducted between 2015 and 2018 in the ICU of a specialized hospital for pulmonary diseases. The trial was approved by the local Ethics Committee of the same hospital and conducted in accordance with the Good Clinical Practice guidelines and the principles of the Declaration of Helsinki.
Consecutive patients over 18 years of age with severe CAP and PaO 2 < 60 mmHg that could not be corrected with oxygen therapy were included in the study. The patients were split into four groups: patients with pneumonia without ARDS, and with pneumonia with mild, moderate, or severe ARDS. ARDS severity was assessed using the criteria of the Berlin definition 13 : acute onset of the disease; bilateral opacities on the chest X-ray not fully explained by effusions, lobar/lung collapse or nodules; respiratory failure of non-cardiac origin; and PaO 2 /FiO 2 on at least 5 cm H 2 O of CPAP between 300 and 200 mmHg for mild ARDS, between 200 and 100 mmHg for moderate ARDS, and under 100 mmHg for severe ARDS. Patients with any of the following conditions were excluded: pregnancy, cardiogenic pulmonary edema, lung carcinoma, active tuberculosis, severe encephalopathy, cardiac or respiratory arrest, hemodynamic instability, unstable arrhythmia, acute myocardial infarction, excessive sputum production, hematemesis or hemoptysis, facial trauma, uncontrolled vomiting, pneumothorax without a chest tube in place, and pleural effusion.
We provided NIV only via dedicated critical care respirators. The initial expiratory positive airway pressure (EPAP) was set at 5 cm H 2 O and increased until alveolar recruitment with oxygenation improvement was achieved. Inspiratory positive airway pressure (IPAP) was started 2-4 cm H 2 O above the EPAP and increased gradually until a tidal volume of 6 ml/kg was achieved. Respiratory rate was 20-25/min and the fraction of inspired oxygen -0.6-1. In the first few days after treatment initiation, the patients were on NIV for more than 16 hours a day. When their oxygenation started to improve, the ventilator-free time was extended until we were able to correct their hypoxemia with oxygen therapy alone.
Deterioration of oxygenation, unresponsive to changes in the ventilator settings, impaired consciousness, hemodynamic instability, and inability to protect the airway were defined as criteria for NIV failure with the need for endotracheal intubation.
We monitored the patients' respiratory rate (RR), heart rate (HR) and ABG parameters before initiation of NIV, at 1 hour and 24 hours after ventilation onset and at transition to non-intensive NIV (less than 16 hours a day) or before endotracheal intubation.
Because of the non-parametric fashion of the data, results were presented as median and interquartile range (IQR). The differences between medians of paired variables were analyzed with the Wilcoxon signed-rank test. A p-value < 0.05 was considered statistically significant. The statistical analysis was done using the IBM SPSS v.25 package.
The NIV trial was successful in 85.71% (n=54) of the patients, while 14.29% (n=9) experienced NIV failure and were intubated. Five (7.94% from the whole study population and 55.56% from the NIV failure group) died and the other three were extubated successfully. The indications for endotracheal intubation were: deterioration of oxygenation (n=7), high leak (n=1) and impaired consciousness (n=2). Amongst the deceased, three were with mild, three with moderate, and two with severe ARDS. None of the NIV success patients died. Folia Medica I 2021 I Vol. 63 I No. 3 First we conducted a statistical analysis to determine the dynamics of the clinical and ABG parameters in the whole study population (Table 1). In the general study population, we observed a statistically significant difference in oxygenation (defined by the PaO 2 /FiO 2 ) only between the initial and the last value, while PaO 2 improves on the first hour after ventilation onset. The clinical parameters showed rapid improvement early in the course of treatment.
To assess the dynamics of the monitored parameters according to disease severity, we allocated the patients into four groups. For every group, a separate statistical analysis was conducted (Tables 2-5). The results show that in the "pneumonia without ARDS" and "moderate ARDS" groups, the dynamics of ABG parameters follows the trend observed in the entire population. In the severe ARDS group, PaO 2 /FiO 2 was significantly higher at 1 hour, then deteriorated and improved again later. In the mild ARDS group, we could not observe any significant improvement of PaO 2 /FiO 2 in comparison to the initial values.
The clinical parameters improved significantly at 1 hour after ventilation onset in all groups, except for the mild ARDS, where the RR and HR decrease at 24 hours.

DISCUSSION
The results of this study suggest that the use of NIV helped avoiding invasive ventilation in 85.71% of the patients with severe CAP in the present cohort. The NIV failure rate in our study was lower than that reported in other studies (between 20% and 70.3%). [14][15][16][17] The great difference between the NIV success rates in the different studies is mainly due to the variable patient selection criteria. In time, the criteria become stricter and this leads to reduction of failed NIV attempts. This thesis is confirmed by Demoule et al., who report an increase of NIV success in de novo hypoxemic ARF patients in the last 15 years. 18 The use of NIV in all groups was associated with a significant improvement of RR and HR shortly after ventilation onset with a tendency of normalization at the end of intensive NIV treatment. This means that there is a reduction in the work of breathing, the O 2 consumption, heart muscle strain, and improvement in the psychological state of patients. Similar effects of NIV have been reported by other authors. [14][15][16] Improvement in PaO 2 /FiO 2 ratio in the whole study population was observed only after a few days of complex antibiotic and supportive treatment. A possible explanation for this phenomenon is the fact that PaO 2 /FiO 2 is the main indicator for ventilation/perfusion mismatching and mirrors the degree of lung tissue damage. This means that it can improve significantly only after reduction of the inflammatory process itself. The late improvement of PaO 2 / FiO 2 we observed correlates with the results reported in other studies. 19,21,22 In the patients with pneumonia without ARDS and moderate ARDS, the central tendency mirrors that of the : comparison between the parameters before NIV initiation and at 1 hour after ventilation onset; р 2 : comparison between the parameters before NIV initiation and at 24 hours after ventilation onset; р 3 : comparison between the parameters before NIV initiation and at transition to non-intensive NIV/ before endotracheal intubation Table 4. Dynamics of the monitored parameters in the group of moderate ARDS (n=26) Before initiation of NIV : comparison between the parameters before NIV initiation and at 1 hour after ventilation onset; р 2 : comparison between the parameters before NIV initiation and at 24 hours after ventilation onset; р 3 : comparison between the parameters before NIV initiation and at transition to non-intensive NIV/ before endotracheal intubation entire population. Therefore, we can conclude that these groups shape the results observed in the cohort as a whole. Interestingly, we observed no improvement of PaO 2 /FiO 2 in the mild ARDS group and a significant improvement of the same parameter early in the course of treatment in the severe ARDS group, which disagrees with the general perception that NIV is suited mainly for cases of mild to moderate ARDS. 23 Although interesting, the results might be biased by the unequal number of patients in the four study groups. If there were milder and more severe ARDS cases, the results could have been different and maybe more coherent with the observed central tendency. Therefore, further research is needed in order to draw any definite conclusions.
Limitations of the study 1. Small number of patients, particularly in the individual groups. The whole study population was not that small (a total of 63 patients), but none of the groups had more than 30 patients.
2. The patient groups do not have equal number of patients, which could have biased the results and shaped the central tendency in a wrong way.
3. The study was conducted at a specialized hospital for pulmonary diseases with great experience in NIV. Careful patient selection and the extensive training of the ICU personnel might have influenced the positive results, particularly in the severe ARDS group.