An analysis of the treatment effect of two modes of oxygenation on patients with radiation pneumonia complicated by respiratory failure

1. Background

With the global increase in the number of patients with thoracic cancer and the characteristics of stereotactic radiotherapy (SBRT), such as high efficiency, low toxicity, cost effectiveness, and ease of use, SBRT has become widely used in the treatment of thoracic cancer [1, 2]. Human lung tissue is sensitive to radiation and studies have demonstrated that the lethal dose is 10 Gy. Although the incidence of radiation pneumonia has decreased with the continuous improvements in radiotherapy technology, it is estimated that radiation pneumonia is common in patients with cancer who receive radiation treatment, with an average incidence of 9–28%. In patients with breast cancer who receive radiotherapy, although the incidence is relatively low, it also reaches 5–15% [3, 4, 5, 6, 7]. The acute phase of radiation lung injury refers to the acute pulmonary toxicity that occurs during radiation therapy or within 1–6 months after the completion of a course of chest radiation treatment, and symptoms may appear prior to imaging changes. The most common manifestations are breathing difficulties, coughing, and hemoptysis, and sometimes the patients are complicated by varying degrees of fever. The computed tomography (CT) severity scale for radiation pneumonia is 0–5, according to the National Cancer Institute’s Common Terminology Criteria for Adverse Events version 4.0 (CTCAE 4.0). Severe radiation pneumonia (grade 2 and above) usually requires hospitalization for oxygen therapy, and in severe cases (more common in grade 3 and above), mechanical ventilation, infusion, empirical antibiotics, and intravenous steroids are required [8, 9]. Radiation-induced lung injury is a complex pathophysiological process involving multiple mechanisms interacting with each other. In the course of the disease, damage to the alveolar type II epithelial cells and capillary endothelial cells leads to the reduced secretion of lung alveolar surfactant and increased capillary penetration, resulting in alveolar collapse, atelectasis, and alveolar interstitial edema, which affects the diffusion process of gas in the alveoli. Patients with severe disease may suffer from persistent and severe hypoxemia, especially in the acute stage of the disease. If hypoxemia cannot be corrected timeously, it will often lead to a poor prognosis [10, 11]. In a retrospective study involving 191 patients who underwent chest radiotherapy from 1988 to 1998, 49% of patients exhibited the pulmonary exudative changes of radiation pneumonia, of which 36% ( n = 69) had mild changes and 13% ( n = 25) had severe changes. The three-year survival rates of patients without, with mild, and with severe radiation pneumonia were 33.4%, 38.2%, and 0%, respectively ( P = 0.003). The study demonstrated the necessity of the early diagnosis and treatment of radiation pneumonia of grade 2 or above [12]. With the continuous progress in oxygen therapy technology, effective oxygen therapy measures have become the key to treating radiation lung injury complicated by hypoxemia in the acute stage. In clinical practice, the non-rebreather mask (NRBM) and high-flow oxygen nasal cannula (HFNC) are the most common apparatus for providing high-concentration non-invasive oxygen treatment. The purpose of this study is to compare the improvement in respiratory function in patients with acute radiation pneumonia complicated by respiratory failure when using these two non-invasive oxygen inhalation techniques that can provide high-concentration oxygen.

2. Methods

3. Results

A total of 163 patients were included in this study, 38 of whom were excluded based on the above exclusion criteria, 125 patients were included in the per-protocol analysis, NRMB group ( n = 63), and the HFNC group ( n = 62) (Supplementary Figure 1). The number of patients with COPD was higher in the HFNC (15, 24.2%) than the NRBM group (6, 9.5%; P = 0.035), but there was no significant difference between the groups in terms of the other characteristics (Table 1).

The PaO 2 ÷ FiO 2 , respiratory rate, and tidal volume at 0, 4, 8, 12, 24, 48, and 72 h after admission of two groups of patients were collected for repeated ANOVA. (1) There were differences in the PaO 2 ÷ FiO 2 , respiratory rate, and tidal volume between the two groups at each time point ( F = 258.177, 294.121, and 134.372, respectively; 0.01). (2) The PaO 2 ÷ FiO 2 , respiratory rate, and tidal volume were different under the different modes of oxygenation ( F = 40.671, 168.742, and 55.353, respectively; 0.01). (3) The PaO 2 ÷ FiO 2 , respiratory rate, and tidal volume changed over time in different oxygen modes ( F = 7.480, 9.115, and 12.165, respectively; 0.01; Fig. 1).

OPEN IN VIEWER

Figure 2.

a: Kaplan-Meier analysis of two groups under different modes of oxygenation. b: Kaplan-Meier analysis of tracheal intubation with different grades of CTCAE4.0.

There were 7 cases (30.4%) in the HFNC group and 20 cases (58.9%) in the NRBM group (Table 2). Regarding the comfort of patients, the incidence of mucus dryness and sensation of asphyxia in the HFNC group and the NRBM group was 14 (22.6%) vs. 32 (50.8%) ( P = 0.001) and 17 (27.4%) vs. 33 (52.4%) ( P = 0.004), respectively. Based on the visual analog scale, the comfort level of patients in the HFNC group was significantly better than that in the HFFM group ( P = 0.0004) (Table 2).

At different time periods (4–8, 8–12, 12–24, 24–48, and 48–72 h), patients in the HFNC group and NRBM group who underwent tracheal intubation and assisted breathing were: 0 (0.0%) vs. 7 (11.1%), 8 (12.9%) vs. 9 (14.3%), 7 (11.3%) vs. 7 (11.1%), 5 (8.1%) vs. 7 (11.1%), and 3 (4.8%) vs. 4 ( 6.3%), respectively. The total number of patients receiving invasive mechanical ventilation at 4–72 h was 23 (37.1%) in the HFNC group and 34 (54.0%) in the NRBM group, and the Fisher’s precise test showed that these differences were statistically significant ( P = 0.043; Table 2).

Kaplan-Meier analysis of the HFNC and NRBM groups with tracheal intubation as the outcome showed that the length of time from hospitalization to tracheal intubation and mechanical ventilation in the two groups was 55.3 ± 3.2 vs. 45.9 ± 3.6 h, respectively ( P = 0.035; Fig. 2a). A Kaplan-Meier analysis of the patients at a CTCAE4.0 grade of 2 and 3, with intubation as the outcome, showed that the time from hospitalization to tracheal intubation and mechanical ventilation in these two groups was 53.1 ± 3.2 vs. 47.1 ± 3.7 h, respectively ( P = 0.271; Fig. 2b). A Cox regression analysis was conducted based on the APACHE-II score (mean score of 23.0 for all enrolled patients), CT imaging grade, and oxygen therapy methods, and the risk of tracheal intubation in patients with an APACHE-II score $>$ 23 was 2.557 times that of patients with a score ⩽ 23. The risk of tracheal intubation in the NRBM group was 1.948 times that of the HFNC group (Table 3). In patients with an APACHE II score $>$ 23, the NRBM oxygen therapy was identified as an independent risk factor for tracheal intubation and mechanical ventilation.

All patients enrolled in this study had varying degrees of mild to moderate hypoxemia. The HFNC group exhibited a faster and more effective improvement in PaO 2 ÷ FiO 2 than the NRBM group. As shown in the treatment time profiles of the different oxygen treatment methods, the oxygenation index of patients in the HFNC group increased rapidly and significantly within 0–12 h when compared with the NRBM group. The rate of improvement in the oxygenation index resulting from the two oxygen treatment methods gradually decreased during the 12–72 h window, but the PaO 2 ÷ FiO 2 in the HFNC group was still higher than in the NRBM group. Similarly, it can be seen from the profile plot that the improvement in the respiratory rate and tidal volume of the patients in the HFNC group was also significantly better than in the NRBM group. Most notably, at 4 h after admission, the respiratory rate of patients decreased rapidly to 28 ± 3 times/min, and the tidal volume increased rapidly from the initial 296 ± 30 to 370 ± 53 ml. As a result of the improvement in the tidal volume and respiratory rate of the patients, the hypoxemic state of the patients in the HFNC group was alleviated over a short time period. From a pathophysiological viewpoint, the cause of hypoxemia in acute radiation pneumonia, especially 1–3 months after thoracic radiotherapy, is thoracic radiotherapy-related lung tissue injury, i.e., the damage to the alveolar type II epithelial cells and capillary endothelial cells resulting in the reduced secretion of lung alveolar surfactants and increased capillary penetration, alveolar collapse, atelectasis, and alveolar interstitial edema, which affects the diffusion process of gas in the pulmonary alveoli. The reason for the significantly better improvement in the respiratory function indexes of patients in the HFNC group when compared with those in the NRBM group is that in addition to providing a high concentration of oxygen (up to 100%), HFNC can output a constant oxygen concentration at constant temperature of 37 ∘ C and a relative humidity of 100%, which can guarantee the normal functioning of airway mucosal cilia [14]. The promotion of sputum dilution is conducive to the excretion of airway secretions, which has obvious advantages over the ordinary nasal catheter and mask oxygen inhalation [15]. The maximum oxygen flow output of the HFNC can reach 60 L/min, which can effectively increase the ventilatory volume of the alveoli and reduce the ventilatory volume of the physiological dead space, resulting in a decrease in the inspiratory phase work of breathing, and at the same time, quickly and effectively reducing the symptoms of dyspnea. The oxygen inhalation device can also produce a certain positive airway pressure in the expiratory phase, which is similar to positive end-expiratory pressure, to prevent alveolar collapse and effectively improve the oxygenation state by increasing the functional residual capacity [16].

Previous studies have compared the application of Venturi mask oxygen inhalation, non-respiration oxygen storage mask, and non-invasive assisted breathing and HFNC in patients with respiratory failure after extubation. Compared with traditional oxygen inhalation, HFNC not only improved patient’s oxygenation status rapidly and reduced the reintubation rate but also showed significantly higher comfort level [17, 18]. Consistently, in this study, we compared the comfort of HFNC and NRBM in patients with radiation pneumonia and respiratory failure by evaluating mucus dryness and sensation of asphyxia, as well as visual analogue scores. The comfort level of patients in the HFNC group was significantly better than that of the HFFM group. The possible reasons were that NRBM provided high-concentration oxygen by wrapping the patient’s mouth and nose (similar to NIV) without re-inhalation, but the communication was restricted. At the same time, the fear of suffocation occurred, which reduced the comfort of this treatment. Also, the airway was not fully humidified. High-flow oxygen inhalation may cause sputum discharge disorder. The dryness of the airway also aggravated the discomfort, therefore reducing the comfort level of this therapy.

In this study, we found that there was a significant difference in the incidence of tracheal intubation between the HFNC group (23, 37.1%) and the NRBM group (34, 54.0%; 0.05). In addition, seven patients (11.1%) in the NRBM group received tracheal intubation 4–8 h after admission, while no tracheal intubation occurred in the HFNC group during this time period. The Kaplan-Meier analysis revealed that in the HFNC group, the time until patients received invasive assisted ventilation through tracheal intubation was longer than in the NRBM group (55.3 ± 3.2 and. 45.9 ± 3.6 h, respectively; 0.05). The need for this ventilation is based on the improvement in the patient’s respiratory function, indicating that the therapeutic effect of the HFNC is greater than the NRBM. However, the Kaplan-Meier analysis revealed that the time to ventilation was 53.1 ± 3.2 h for patients with a CTCAE 4.0 grade of 2 and 47.1 ± 3.7 h for patients with a grade of 3, but this difference was not significant ( $P>$ 0.05). In an animal study [19], it was found by observing the CT scans at different stages after radiotherapy that the most obvious CT changes after radiotherapy were peribronchial opacification, thickened interlobular septa, and reduced lung volume. The changes in CT images caused by radiation pneumonia can be divided into the exudation stage (ground glass and patchy changes), the mixing stage (patchy and fibrous strip shadows), and the consolidation stage (fibrous strip changes). It is well known that chest CT findings reflect the lung tissue density, which may appear approximately one week after clinical symptoms in early-stage or mild patients. The time from the end of the last radiotherapy session to admission in the HFNC and the NRBM group was 0–1 month for 32 (51.6%) and 34 patients (54.0%), 1–2 months for 15 (24.2%) and 19 (30.2%), 2–3 months for 9 (14.5%) and 6 (9.5%), and 3–6 months for 6 (9.7%) and 4 (6.3%), respectively. Therefore, 75.8% of patients in the HFNC group and 84.2% in the NRBM group were in the exuviate stage of the disease within two months of the end of radiotherapy, but the chest imaging findings of patients might not be consistent with the pathological state of the lung tissue at this time, resulting in the lung function of patients being overestimated. This study may explain why in the present study, although the CTCAE version 4.0 can evaluate the severity of pulmonary inflammation, it does not truly reflect the state of pulmonary function, and therefore, the grade cannot be an independent risk factor for invasive mechanical assisted ventilation. In addition, conducting a multivariate Cox regression analysis on the APACHE II score, CTCAE version 4.0 grade, and treatment mode, we found that the APACHE II score and treatment mode were independent risk factors for invasive assisted ventilation through tracheal intubation in the present study. Patients with an APACHE II score $>$ 23 were 2.557 times more likely to need tracheal intubation than those with a score ⩽ 23, which is also consistent with a recent study on the value of the Apache II score in predicting mortality in critically ill patients. The study found that an Apache II score of 15 provided the best diagnostic accuracy to predict the mortality rate in critically ill patients [20]. In a retrospective study of the use of the HFNC in patients with aspiration pneumonia post-stroke, treatment with a HFNC significantly reduced the incidence of endotracheal intubation when compared with treatment with a Venturi mask [21]. Similarly, in the present study the risk of tracheal intubation in the NRBM group was 1.948 times higher than the HFNC group, indicating that treatment using a HFNC can reduce the incidence of tracheal intubation in patients, which may be related to the fact that this treatment improved the ventilation function of the lungs and quickly corrected hypoxemia in the patients. In this study, traditional non-invasive mask-assisted ventilation was adopted for some patients before tracheal intubation to avoid invasive assisted breathing. There were 7 cases (30.4%) in the HFNC group and 20 cases (58.9%) in the NRBM group (Table 2). Although the NRBM group had a higher proportion of NIV, the incidence of tracheal intubation in the NRBM group was still significantly higher than that in the HFNC group (54.0% vs. 37.1%), which may be related to the early application of HFNC through airway humidification and flow rate adjustment to produce PEEP, which improved the diffusion function and ventilation function of the respiratory system (Fig. 1a, c).

There were some limitations of this study. First, there were 15 (24.2%) COPD patients in the HFNC group and 6 (9.5%) in the NRBM group ( 0.05). However, these patients were at the stable stage of the disease. Their lung function was assessed before radiotherapy and able to tolerate. All patients completed radiotherapy. The heterogeneity of patients between the two groups may affect the results of this study. The pathophysiological changes in the lung tissues in the acute stage of radiation pneumonia are similar to those of acute respiratory distress syndrome (ARDS). In a study on the risk of non-invasive positive pressure ventilation and treatment using a HFNC for tracheal intubation in patients with ARDS, it was found that although the HFNC could improve oxygenation status and reduce the work of breathing in patients, it may delay invasive assisted ventilation and lead to a poor prognosis [22, 23]. However, there is a lack of uniform criteria for tracheal intubation in this study, which may lead to significant differences in treatment methods among clinicians, thereby affecting patient outcomes. In the present study, it was found that the incidence of tracheal intubation was lower and the time to transition to invasive assisted ventilation in patients in the HFNC group was higher than in the NRBM group. Unfortunately, our study did not provide a statistical analysis of mortality. In the early stage of radiation pneumonia, the lung tissue is characterized by hypoxic ischemic necrosis of the type I and II alveolar epithelial cells and the exudation and infiltration of lymphocytes from the capillaries to form lymphocytic alveolitis. This process takes approximately a few weeks, and at this stage, the disease is still under control. As the disease continues to progress, numerous cells, cytokines, and molecular organisms participate in the regulation of the immune response and initiate the repair of damage, eventually leading to greater lung damage than that caused by the radiation itself [24, 25, 26, 27]. In this complex and multi-factorial pathophysiological process, non-functional fibrosis and repair are dominant, including collagen fibers, collagen deposition, alveolar septal thickening, and capillary occlusion, resulting in radiation-induced pulmonary fibrosis [10, 11]. At this time, pulmonary compliance is significantly reduced, and the ventilation and gas exchange functions are seriously impaired. Therefore, the enrolment of patients at the early, reversible phase may be part of the reason for the rapid improvement in respiratory function and oxygenation in this study. Further studies are needed to determine whether treatment with a HFNC has the same effect of improving the respiratory function of patients in the mixed phase, with patchy and fibrous strip shadows, and in the consolidation phase, with fibrous strip shadows. In addition, this study did not adopt the same flow-rate setting for the HFNC group but adjusted the flow-rate setting of the inhaled gas according to the patient’s tolerance. The heterogeneity of this setting method may lead to a bias in the results.

5. Conclusion

In patients with radiation pneumonia complicated by respiratory failure, when compared with the NRBM, the HFNC can improve the oxygenation state over a short period, reduce the respiratory rate, increase the tidal volume, extend the time until mechanical ventilation, and reduce the incidence of tracheal intubation within 72 h after admission.

Author contributions

XD and DSM conceived the idea and conceptualised the study. CL collected the data. JJ and LH analysed the data. XD and LD drafted the manuscript. DSM reviewed the manuscript. All authors read and approved the final draft.

Availability of data and materials

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Funding

This study was funded by the Medical Science Research in Hebei Province (20170729) and Medical Science Research in Hebei Province (20201083). The funding body had no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.