High-fl ow nasal cannula oxygen therapy and noninvasive ventilation for preventing extubation failure during weaning from mechanical ventilation assessed by lung ultrasound score: A single-center randomized study

Shan-xiang Xu, Chun-shuang Wu, Shao-yun Liu, Xiao Lu

Emergency Department, the Second Affiliated Hospital of Zhejiang University School of Medicine, Hangzhou 310009, China

KEYWORDS: High-fl ow nasal cannula oxygen; Noninvasive ventilation; Lung ultrasound; Extubation

INTRODUCTION

Mechanical ventilation is associated with the occurrence of complications whose incidence rates increase with the duration of respiratory support. Optimizing weaning to reduce the duration of ventilation is crucial.[1,2]The extubation of the patient marks the end of the weaning period; however, extubation is permitted only after a successful weaning trial involving spontaneous ventilation without positive end-expiratory pressure ( PEEP) or with a T-tube. Furthermore, between 20% and 30% of extubated patients who meet the criteria for being extubated develop respiratory distress within 48 hours after extubation, which is characterized by an inability to maintain an eff ective airway. In this context, the reintroduction of mechanical ventilation with reintubation or noninvasive ventilation (NIV) using a face mask is generally required.

Respiratory distress after extubation is associated with increased morbidity and mortality rates. Its multifactorial pathophysiology leads to a loss of pulmonary aeration during the weaning process.[2,3]Lung ultrasound can accurately quantify the loss of pulmonary aeration before, after, and during the weaning trial by calculating the lung ultrasound score (LUS). A previous prospective dual-center study of 100 patients demonstrated that the intensity of lung aeration loss observed during the weaning trial was predictive of the development of postextubation respiratory distress within 48 hours after extubation.[4]An LUS score ≥14 points could identify patients at high risk of developing postextubation respiratory distress.[5-7]Another study reported that among 80 patients weaned from mechanical ventilation, a 30% reduction in respiratory distress in the postextubation period was achieved.[8]

Meanwhile, high-fl ow nasal cannula oxygen (HFNCO2) appears to be an effective new therapeutic option when compared with other oxygen delivery devices (e.g., nonrebreathing oxygen masks, high-humidity face masks, high-humidity tracheostomy collars, and Venturi mask). The major benefi ts of HFNCO2include continuous alveolar recruitment and a reduction in airway collapse (due to the eff ect of continuous positive airway pressure [CPAP]). Both of these physiological effects are key factors in achieving adequate minute ventilation and sufficient oxygenation.[9,10]Some studies have indicated the superiority of HFNCO2among the extubation patients,[11,12]and several randomized clinical trials using different NIV respiratory devices have reported reduced incidence rates of postextubation pneumonia and re-intubation, improved weaning outcomes, and shorter hospital stay lengths. Such benefi ts of NIV and HFNCO2may be correlated with the similar physiological effects observed when applying CPAP during the early postextubation period.[13-16]It is unknown whether HFNCO2and NIV in combination would impart clear physiological advantages among high-risk patients as assessed by the LUS score. The testing of a targeted therapeutic strategy in a group of high-risk patients defined as having an LUS score ≥14 points at the end of the weaning trial could lead to a reduction in the incidence of extubation failure and associated rates of morbidity and mortality.

METHODS

Study population and recruitment

A prospective, cross-sectional study was conducted in an emergency department-based intensive care unit (ICU) of the Second Affiliated Hospital of Zhejiang University School of Medicine from November 2017 through November 2019. The inclusion criteria were an age of older than 18 years and a recent history of being mechanically ventilated on tracheal intubation for more than 48 hours. The exclusion criteria were a diagnosis of chronic obstructive pulmonary disease with a moderate to severe status (stages 3 and 4), defi ned by a forced expiratory volume in one second of less than 50% of the theoretical value; a history of chronic respiratory disease; paraplegia status with a level higher than T8, severe ICUacquired neuromyopathy that have a low rate of extubating successfully; a history of tracheostomy for any reason; and the receipt of a clinical decision not to intubate.

Performance of the study

At the institution in this study, LUS is routinely performed prior to extubation at the end of the weaning trial involving the pressure support ventilation plus zero end-expiratory pressure (PSV-ZEEP) protocol or once spontaneous breathing on a T-piece has been achieved successfully. Before conducting the spontaneous breathing trial, the following items were confirmed: (1) absence of large pleural effusion defined using transthoracic LUS at an interpleural distance of at least 5 cm; (2) lack of positive fluid balance; (3) absence of significant atelectasis that would require fibroaspiration; (4) a hemoglobin level of at least 8 g/L; and (5) ongoing antibiotics usage if the Clinical Pulmonary Infection Score was less than six points. Just after extubation, patients were randomized to the control group or intervention group.

Randomization was performed using the REDCap software program (REDCap Consortium). The attending physician remained unaware of the LUS scores of patients in the control group at the end of the spontaneous breathing test, and these patients received standard care. Conversely, for patients in the treatment group, the physician knew the LUS scores recorded during the spontaneous breathing trial and the therapeutic management selection after extubation was as follows: patients with an LUS score ＜14 points (at low risk of extubation failure) were extubated and received standard preventive care without NIV or HFNCO2, while patients with an LUS score ≥14 points (at high risk of extubation failure) were extubated with a second review of therapeutic optimization to identify and address any persisting risk factors for postextubation respiratory distress. Patients with an LUS score ≥14 points received HFNCO2therapy (Optiflow?; Fisher & Paykel Healthcare, New Zealand) combined with sessions of preventive NIV (4-8 hours per day for 4-8 sessions total) during the fi rst 48 hours after extubation.

The primary endpoint of this study was the incidence of extubation failure within 48 hours of planned extubation, defined as the need for reintubation or the implementation of therapeutic NIV; deaths occurring in this period were also included in this analysis. By extension, high-risk patients in the treatment group (LUS score ≥14 points) who develop signs of acute respiratory distress despite the implementation of prophylactic HFNCO2therapy are classified as cases of extubation failure. Meanwhile, secondary endpoints included the number of ventilation-free days spent in the ICU following the planned extubation after randomization, the length of the total stay in the ICU, and the ICU mortality rate at 28 days.

Lung ultrasound

A Mindray M9 Echography (Mindray Co., China) and a 2- to 4-MHz round-tipped or convex probe was used for examination. In each patient, 12 lung areas were assessed in the right and left lungs. A value was assigned to each area according to the most severe LUS results detected in the corresponding intercostal space as follows: 0 = normal aeration, defined by the presence of lung sliding with horizontal A lines or fewer than two isolated vertical B lines; 1 = moderate loss of lung aeration, defi ned by the presence of either multiple well-defined and spaced B1 lines; 2 = severe loss of lung aeration, defi ned by multiple coalescent vertical B2 lines issued either from the pleural line or from juxtapleural consolidations; and 3 = lung consolidation, defined by the presence of a tissue pattern containing either hyperechoic punctiform images. The LUS score was calculated as the sum of the values of all areas and ranged from 0 to 36 points.

Description of measures adopted to avoid or reduce the bias

LUS was performed before randomization to avoid infl uencing the ranking in the at-risk group. Patients met the same study inclusion criteria of inclusion, but only those with an LUS score ≥14 points entered the analysis concerning the primary endpoint. LUS scores were evaluated before randomization in both groups, but those for the patients in the control group were not made known to the investigator.

The randomization of patients (by blocks) was conducted by center into two groups in a 1:1 manner at the time of study inclusion on an e-CRF server type (REDCap). To minimize bias associated with knowing the patient’s group, the parameters of ventilator support (e.g., number of hours of HFNCO2) were predefi ned in the treatment group, and no provision of preventive NIV or HFNCO2therapy was authorized in the control group.

Data collection

The following data were collected in an e-CRF (REDCap): age, sex, body mass index, Sepsis-related Organ Failure Assessment score, LUS score at admission, and the reason for ventilation/intubation. The respiratory parameters of the two groups at three time points (H0: end of weaning trial; H24: 24 hours after extubation; H48: 48 hours after extubation) were also recorded.

Statistical analysis

The primary analysis was fi rst intention-to-treat and then per-protocol using the Stata version 12 software program (Stata Corp., College Station, USA). All statistical tests were performed at the alpha risk level of 5% (excluding interim analysis). Continuous variables are presented as mean and standard deviation values, subject to the normality of distribution (Shapiro-Wilk). In the case of non-normality, they are presented as median, quartile, and extreme values. Meanwhile, qualitative variables are expressed as numbers and associated percentages.

RESULTS

During the study period, a total of 142 patients who successfully completed a spontaneous breathing trial (60 minutes) and underwent extubation were considered for this study. Forty-four patients were excluded for the following reasons: eight had chronic obstructive pulmonary disease with moderate to severe status (stages 3 and 4), six had paraplegia with a level of greater than T8, fi ve had severe ICU-acquired neuromyopathy, 18 had undergone tracheostomy, and seven patients had received a clinical decision of “do not intubate”. Thus, there were 98 patients who were deemed fi nally eligible for inclusion in this study and who underwent randomization (Figure 1). The demographics, clinical features, and outcomes of these patients are presented in Table 1. The mean participant age was 45±14 years, and there were more male than female patients in both groups. Patients’ characteristics (body mass index, Sequential Organ Failure Assessment score) and the duration of invasive mechanical ventilation before study inclusion were similar between the two groups (P＞0.05). Of the included patients, 49 were assigned to the control group, and 49 were assigned to the treatment group. In the control group, 13 patients had the LUS scores ≥14 points, while 36 patients had scores ＜14 points. In the treatment group, 16 patients had the LUS scores ≥14 points, while 33 patients had scores ＜14 points. The length of ICU stay (9.4±3.1 days vs. 7.2±2.4 days) was significantly different between the two groups, and the reintubation rates at 48 hours (18.4% vs. 10.2%) and seven days (22.4% vs. 12.2%) were signifi cantly diff erent between the two groups (P＜0.05). The 28-day mortality rate (6.1% vs. 8.2%) was not signifi cantly diff erent between the two groups.

Figure 1. Flowchart of the study. ICU: intensive care unit; LUS: lung ultrasound score; NIV: noninvasive ventilation; HFNCO2: high-fl ow nasal cannula oxygen.

For the patients with LUS scores ≥14 points, the rate of extubation failure within 48 hours was 30.8% (4/13) in the control group and 12.5% (2/16) in the treatment group, demonstrating a significant difference (P＜0.05). Among patients with LUS scores ＜14 points, 13.9% (5/36) of patients required re-intubation in the control group, and 9.1% (3/33) of patients required the same in the treatment group (P=0.61). Considering extubation failure at seven days after weaning, 38.5% (5/13) of patients with LUS scores ≥14 points in the control group and 18.8% (3/16) of patients with the same in the treatment group showed extubation failure at this time point, with a signifi cant diff erence between the two groups (P＜0.05). Meanwhile, in patients with LUS scores ＜14 points, 16.7% (6/36) of patients in the control group and 9.1% (3/33) of patients in the treatment group experienced extubation failure at seven days (P=0.35).

The evolution in the respiratory parameters of the two groups is shown in Tables 2 and 3. For patients with LUS scores ≥14 points, at the end of the weaning trial, there were no significant differences in the respiratory rate, pulse rate, mean arterial pressure (MAP), hydrogen ion concentration (pH), partial pressure of carbon dioxide (PaCO2), alveolar oxygen partial pressure (PaO2)/fraction of inspired oxygen (FiO2), or oxygen saturation (SPO2) results between the two groups. At 24 and 48 hours after extubation, there were also no significant differences in the MAP, pH, or PaCO2results between the two groups; however, those with lower respiratory and pulse rates and higher PaO2/FiO2and SPO2values were more often found in the treatment group. Among patients with LUS scores ＜14 points, no signifi cant diff erences were apparent in the respiratory rate, pulse rate, MAP, pH, PaCO2, PaO2/FiO2, or SPO2results between the two groups after the weaning trial and at 24 and 48 hours after extubation. Overall, the PaO2/FiO2and SPO2values of patients with LUS scores ＜14 points were higher than those of patients with LUS scores ≥14 points.

Table 1. Characteristics of the study participants at baseline

Table 3. Evolution of respiratory parameters among patients with an LUS score ＜14 points

DISCUSSION

The main results of the present study can be summarized as follows: first, NIV+HFNCO2therapy could reduce the incidence of extubation failure at both 48 hours and seven days, even among patients at high risk of developing postextubation respiratory distress (LUS score ≥14 points). Second, among patients at low risk of developing postextubation respiratory distress (LUS score ＜14 points), the regular treatment is as efficient as NIV+HFNCO2therapy. Third, NIV+HFNCO2therapy may decrease the respiratory rate, increase the PaO2/FiO2and SPO2values, and may improve the patient’s discomfort as a result of lacking symptoms of airways dryness among patients at high risk of developing postextubation respiratory distress (LUS score ≥14 points) after receiving the treatment for 24 hours.

To reduce the risks associated with extubation failure, the adoption of tools able to predict the onset of respiratory distress after extubation becomes essential to maximize the weaning period and improve the prognosis of ventilated, critically ill patients. LUS, which is easy to use, enables physicians to quantify both the overall and regional lung aeration using a validated scoring system. The lung ultrasound is measured at the bedside reproducibly, without the need to involve special equipment and without any significant risk of adverse effects on patients. Pulmonary transthoracic ultrasound is increasingly being used in the ICU because it is noninvasive, does not require transport outside of the ICU, and does not necessitate the exposure of the patient to radiation. Evaluating the recovery of ventilatorassociated pneumonia by assessing lung re-aeration during antibiotic treatment was compared with using computed tomography by Bouhemad et al.[4]Ultrasound was validated by the same group as a means to assess alveolar recruitment resulting from the application of PEEP in adult patients with acute respiratory distress syndrome.[5]Atelectasis may persist for up to 24-48 hours after extubation following anesthesia and paralysis, even in patients with healthy lungs.[17]Oxygen therapy is almost invariably employed after extubation to correct residual oxygenation impairment. Because of its positive eff ects on the respiratory system, HFNCO2is an appealing approach to reversing postextubation atelectasis and improving oxygenation.[18]Few studies have been published to date on the use of NIV+HFNCO2therapy after extubation in patients with high risk of developing postextubation respiratory distress as assessed by LUS.

The rate of failed extubation is variable but may reach 20% or more,[19,20]and the ideal approach to avoiding the need to reintubate has yet to be determined. Whether HFNCO2is benefi cial postextubation has been examined in patients after cardiothoracic surgery and abdominal surgery and in general ICU patients at both high and low risk for reintubation.[21]In the context of postextubation respiratory failure, HFNCO2is consistently better tolerated than NIV. However, although HFNCO2seems noninferior to NIV with regard to intubation and mortality rates after cardiothoracic surgery and in high-risk ICU patients, the fi ndings are more controversial when it is used after abdominal surgery.[22]It remains to be elucidated whether these dissimilarities stem from a variable effect stemming from thoraco-abdominal coordination or other causes. In low-risk hypoxemic patients, support with HFNCO2seems to prevent intubation to a certain degree as compared with conventional oxygen therapy. NIV as a treatment for acute respiratory distress after extubation is often implemented even though the level of evidence for doing so remains inadequate. NIV is recommended (should probably be implemented for grade 2+) to treat acute respiratory distress after extubation occurring in the immediate postoperative period of supramesocolic abdominal surgery, thoracic surgery, and surgery for aortic aneurysm.[23-29]However, no benefit has been demonstrated for mitigating other causes of extubation failure (i.e, probably should not be implemented for grade 2-). This last recommendation is based on the results of randomized studies that reported no reduction in the risk of re-intubation, and a higher mortality rate associated with late reintubation in the group treated with therapeutic NIV.[30,31]NIV is the first-line method for promoting lung recruitment and improving CO2elimination in patients with respiratory distress. In most cases, however, it cannot be administered continuously because of limited clinical tolerance. An NIV session for one hour is generally followed by an interruption for two hours, a period during which conventional oxygenotherapy is administered. The period of NIV interruption likely results in lung derecruitment and decreased alveolar ventilation.[32]

To our knowledge, this is the first study to describe the use of NIV+HFNCO2therapy in a population at high risk of developing postextubation respiratory distress. NIV+HFNCO2therapy may have a good eff ect on mitigating high-risk postextubation respiratory distress.

CONCLUSIONS

Among high-risk (LUS score ≥14 points) adults weaned from mechanical ventilation and assessed by LUS, NIV+HFNCO2does not reduce the mortality rate within 28 days but reduce the length of ICU stay and lessen the rate of extubation failure at both 48 hours and seven days. NIV+HFNCO2may play an important role in preventing reintubation in high-risk adult patients, although further studies are needed to better define which individuals might benefit the most and to discern the optimal timing of application. Findings of ongoing randomized trials will hopefully help to answer these questions.

Funding:None.

Ethical approval:The experimental protocol was conducted according to ethical principles in human and animal experimentation and was approved by the Commission on Ethics of the Zhejiang University School of Medicine (protocol number 2017-050).

Confl icts of interests:No benefi ts in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

Contributors:SXX proposed the study, and wrote the fi rst draft. All authors contributed to the design and interpretation of the study and to further drafts.