Bilevel Positive Airway Pressure Versus Nasal Continuous Positive Airway Pressure for Prevention of Extubation Failure in Infants After Cardiac Surgery: A Randomized Controlled Trial

Objective: To evaluate the effect of bilevel positive airway pressure (BiPAP) and nasal continuous positive airway pressure (NCPAP) in respiratory support after extubation in infants undergoing cardiac surgery. Methods: A total of 83 infants who underwent repair of atrial septal defect (ASD) or ventricular septal defect (VSD) after extubation were randomized to the BiPAP group (n= 42) or the NCPAP group (n= 41) between January 2020 and December 2020. The primary outcomes were the extubation failure rate and the level of PCO 2 within 24 h after extubation. Results: The baseline characteristics between the two groups were similar. The introduction of BiPAP for post-extubation respiratory support did not reduce extubation failure rates compared to NCPAP (P>0.05). The PaCO 2 level within 48 h was signicantly lower in the BiPAP group (P<0.05). Additionally, the PaO 2 /FiO 2 in the BiPAP group was signicantly higher than that in the NCPAP group at 6h, 12h and 24h after treatment (P<0.05).There were no statistically signicant differences in duaration on NIV, hospital length of stay, total hospital costs in $ and complications between the two groups (P>0.05). Conclusion: The introduction of BiPAP for post-extubation respiratory support did not reduce extubation failure rates versus NCPAP. However, BiPAP was shown to be superior to NCPAP in improving oxygenation and carbon dioxide clearance.


Introduction
Infants with congenital heart disease (CHD) need respiratory support after surgery because of their young age and poor independent mobility. Noninvasive ventilation (NIV) has been proven to be useful for respiratory support in pediatric patients and can effectively reduce the extubation failure rate in high-risk infants [1][2][3][4][5]. In recent years, the technique of NIV has become increasingly mature, especially the application of bilevel positive airway pressure (BiPAP) and nasal continuous positive airway pressure (NCPAP), which have signi cantly reduced the application of invasive ventilation and shortened the time of mechanical ventilation. NIV has previously been widely used in the eld of premature babies and newborns. Studies have con rmed that BiPAP and NCPAP can be used to treat neonatal respiratory failure, respiratory distress syndrome (RDS) and reduce the incidence of bronchopulmonary dysplasia [2,[5][6][7][8]. However, few studies have focused on the application of NIV in infants with CHD after extubation.
We designed this study to evaluate the e cacy and safety of BiPAP and NCPAP as respiratory support modalities after extubation in infants who underwent cardiac surgery.

Patients and study design
We conducted a randomized, controlled study at the cardiac intensive care unit (CICU) of Inclusion criteria were infants less than 6 months of age with satisfactory anatomical extubation and stable hemodynamics after repair of ASD or VSD. Additional inclusion criteria were infants extubated to NIV support for at least 24 hours to either BiPAP or NCPAP. If patients achieved the following criteria, we would consider extubation: (1) Hemodynamic stability without a large dose of vasoactive drug support; (2) Fraction of inspired oxygen (FiO 2 ) ≤ 40%, peak inspiratory pressure (PIP) ≤ 18 cmH 2 O, and positive end-expiratory pressure (PEEP) 2~4 cmH 2 O; and (3) Arterial blood gas (ABG) showing partial pressure of carbon dioxide (PaCO 2 ) < 50 mmHg, partial pressure of oxygen (PaO 2 ) > 70 mmHg, pH > 7.30, and lactic acid < 2.0 mmol/L. The exclusion criteria were for patients who had congenital thoracic and abdominal malformations, postoperative extracorporeal membrane oxygenation support or preoperative tracheotomy and intubation, parents who decided not to participate, CICU transfer prior to extubation, or death prior to extubation.

Data Collection and De nitions
The primary outcomes were the extubation failure rate and the level of PCO 2 within 24 h after extubation.
For secondary outcomes, the differences in postoperative duaration on NIV, hospital length of stay, total hospital costs in $ and incidence of complications were analyzed. Besides, we analyzed the changes in ABGs (pH, PaO 2 , PaCO 2 , and PaO 2 /FiO 2 ratio) before and after the treatment in the two groups.
Pulmonary hypertension was de ned as a mean pulmonary artery pressure of 25 mmHg or higher [9]. The diagnosis of ventilator-associated pneumonia (VAP) was based on the criteria established by the Centers for Disease Control and Prevention, with diagnosis aided by chest radiographs, positive sputum cultures, transtracheal uid, bronchial washings, and clinical ndings [10]. Extubation failure was de ned as reintubation within 48 h after the rst planned extubation. Endotracheal intubation was performed when patients developed hypoxemia (FiO 2 > 0.60 for target SpO 2 ), respiratory acidosis (pH < 7.20, PaCO 2 > 65 mmHg), polypnea, or elevated serum lactic acid (> 2.0 mmol/L). ABGs were taken before treatment and at 6 h, 12 h and 24 h after treatment and analyzed using an ABL90 FLEX system.

Ventilation Strategies
After extubation, BiPAP or NCPAP was connected to infants through silicone binasal prongs. BiPAP or NCPAP was delivered through a time-cycled, pressure-limited, and continuous-ow ventilator (Infant Flow SiPAP system, CareFusion, California, USA), which detected the inspiratory effort of the infants by means of the Graseby abdominal capsule-triggering device. The initial lower and higher respiratory parameters in the BiPAP group were set at 3~6 cmH 2 O and 8~10 cmH 2 O, respectively; the FiO 2 was 0.21~0.60; and the pressure exchange rate was 20~30 exchanges/min. Respiratory settings were adjusted to maintain blood gas analysis within normal ranges. The respiratory parameter settings in the NCPAP group were set at a pressure of 3~6 cmH 2 O; the oxygen ow was 6~8 L/min; and the FiO 2 was 0.21~0.60. Respiratory settings were adjusted to maintain blood gas analysis within normal ranges. SpO 2 was maintained at 90~95% during ventilation. After extubation, all infants were administered aerosolized budesonide suspension and terbutaline sulfate solution and intravenously injected with methylprednisolone sodium succinate for anti-in ammation. Patients with excessive phlegm were given intravenous infusion of ambroxol hydrochloride to reduce phlegm and lung physical therapy.

Allocation and blinding
Randomization was performed using random numbers generated by a computer. Sequentially numbered and sealed opaque envelopes were used to contain group assignments. When the infants were admitted to the CICU and had satis ed the inclusion criteria, envelopes were opened by the CICU physician, who was not directly involved in the study or the analysis of results. The allocated BiPAP or NCPAP treatment was started immediately. The BiPAP device produces audible noise that cannot be masked. Accordingly, clinicians involved in patient care and researchers assessing study endpoints were not blinded to the nature of the study treatments.

Statistical Analysis
Sample size estimation was calculated with PASS software (version 15; NCSS LLC, Kaysville, UT, USA).
Based on the extubation failure rate in the pre-experiment, assuming the difference between the two independent populations was 10%, α=0.05, and β=0.2, the number of participants needed was 26 in each group. Assuming a 10% attrition rate, the total sample size was 58 (29 per group).
The data of this study were analyzed by SPSS 25.0. To demonstrate normal data distribution, a Kolmogorov-Smirnov test was performed before each analysis. If normally distributed, a t-test analysis was performed; otherwise, the Wilcoxon-Mann-Whitney test was applied. For further characterization, we performed repeated-measure analysis of variance (two-way ANOVA) and displayed the interaction between and within the study groups. A chi-square test was used to compare the qualitative data between the two groups. A P value <0.05 was de ned as statistically signi cant.

Results
A total of 97 infants were screened between January 2020 and December 2020, of which 7 did not meet the inclusion criteria, 4 underwent preoperative tracheotomy and intubation, and the parents of 2 declined to participate. Finally, 83 infants were ultimately enrolled and nished the trial (42 in the BiPAP group; 41 in the NCPAP group) (Fig. 1).
There was no signi cant difference in extubation failure rates within 48 hours between groups (7.1% in the BiPAP group and 9.8% in the NCPAP group, P=0.713, Table 2). Before treatment, there was no signi cant difference in pH, PaO 2 , PaCO 2 , or PaO 2 /FiO 2 between the two groups (P>0.05). Compared with the values before treatment, the PaCO 2 gradually decreased, and the PH, PaO 2 , and PaO 2 /FiO 2 gradually increased after treatment. The PaCO 2 of the BiPAP group was signi cantly lower than that of the NCPAP group at 12 h and 24 h after treatment. Additionally, the PaO 2 /FiO 2 in the BiPAP group was signi cantly higher than that in the NCPAP group at 6h, 12h and 24h after treatment. The pH, PaO 2 , PaCO 2 , and PaO 2 /FiO 2 were similar between the groups during the entire experiment at other time points (Fig 2).
There was no signi cant difference in duaration on NIV, hospital length of stay, total hospital costs in $ between the two groups. Furthermore, there was no statistically signi cant difference in terms of complications. (Table 2).

Discussion
Persistent hypoxemia often occurs after extubation in patients undergoing cardiac surgery, which affects the prognosis of patients and increases the cost of hospitalization. The reason may be blood contact with foreign matter and organ hypoperfusion during the process of CPB, which leads to an increase in systemic in ammatory mediators and results in lung injury [11][12][13]. During pulmonary ischemiareperfusion, in ammatory cells in ltrate, pulmonary interstitial exudation increases, and microvilli on the surface of type II alveolar epithelial cells decrease in abundance. Therefore, patients are more likely to have complications such as alveolar collapse, atelectasis, acute respiratory failure, pulmonary infection, and so on [14,15]. At the same time, during invasive mechanical ventilation, alveolar mechanical ectasia and partial alveolar structure destruction cause lung injury and airway remodeling [16][17][18]. Therefore, for these kinds of patients, the NIV transition is usually required after extubation.
NIV mainly includes NCPAP, BiPAP, high-ow nasal cannula, nasal intermittent positive pressure ventilation, and so on. Compared with invasive ventilation, NIV can signi cantly reduce VAP and other serious mechanical ventilation-related complications, relieve patients' pain and make it easy for patients to accept ventilation [19,20]. Moreover, NIV can reduce respiratory work and provide positive airway pressure by heating and humidifying air ow into the nasopharynx so that infants can obtain different levels and frequencies of respiratory support. In the past three decades, there has been an increasing number of studies on NIV in pediatrics. The main indications are obstructive sleep apnea syndrome, restrictive lung diseases, apnea of prematurity, acute and chronic respiratory failure, etc. [1,4,21,22]. However, there are few studies on the application of NIV in infants who underwent cardiac surgery.
Extubation failure after congenital heart surgery is associated with prolonged hospital stay and mortality [23]. Studies have shown that BiPAP could reduce the need for reintubation, especially in patients with respiratory failure and cardiogenic pulmonary edema [24][25]. In this study, we found that a total of 7 patients were reintubated within 48 h, with an incidence of only 8.4%. However, there was no signi cant difference in the extubation failure rate between the two groups. This may be due to the different research subjects and interventions applied. Multicenter studies with larger sample sizes may be needed in the future to further con rm this conclusion.
Our study showed that in the two groups, the PaO 2 and PaO 2 /FiO 2 gradually increased, the PaCO 2 gradually decreased after treatment. Continuous air ow impact of NCPAP can reduce upper airway resistance, limit thoracic deformation, supply natural work of breathing, maintain alveolar functional residual capacity, prevent alveolar collapse, and reduce the consumption of autologous alveolar surfactant so that NCPAP can improve alveolar ventilation and reduce the use of exogenous pulmonary surfactant and invasive ventilation [26,27]. At present, NCPAP is still widely used as the initial mode of respiratory support in the clinic. BiPAP not only retains the characteristics of NCPAP but also combines the advantages of the pressure support/pressure control ventilation mode. BiPAP provides two different levels of pressure support during the respiratory cycle, and patients can breathe spontaneously completely under high pressure and low pressure, avoiding the problem of man-machine confrontation [28,29]. After a certain period of treatment, PaO 2 /FiO 2 in the BiPAP group was higher than that in the NCPAP group, while PaCO 2 in the BiPAP group was lower than that in the NCPAP group. The reason may be that compared with NCPAP, BiPAP can intermittently give higher pressure support based on PEEP, which is conducive to improving oxygenation and removing carbon dioxide. Therefore, BiPAP has a better respiratory support effect, better oxygenation improvement function and CO 2 emission effect [30][31].
We found that there were no statistically signi cant differences between BiPAP and NCPAP in terms of duaration on NIV, hospital length of stay, total hospital costs in $, and ventilator-related complications. Zoremba et al. noted that short-term use of BiPAP improved lung function within 24 h in their study [32]. Compared with endotracheal intubation, BiPAP is more comfortable, has lower rates of mortality and iatrogenic infection, and helps to avoid ventilator-related complications such as VAP and the need for deep sedation [33][34][35]. Other advantages include oxygenation improvement and respiratory work reduction, resulting in lower myocardial oxygen demand. Diaphragmatic paralysis or dysfunction is common in patients who underwent cardiac surgery, and it can also be used as an indication for the application of BiPAP [36]. Tobias's study showed a decrease in the respiratory rate and PaCO 2 in postoperative patients [37]. These patients were in a state of impending respiratory failure, and BiPAP improved oxygenation, lowered carbon dioxide levels and avoided reintubation.
Our study has several limitations. First, the obvious limitation is the small sample size. This study was conducted at a single center, which limits the generalizability. Second, the medical staff could not be blinded to the randomized mode of support. Although we used objective failure criteria and management protocols, the possibility of a bias might exist. Third, although the short-term effects of NIV on ventilation were considered, the relationship between NIV and long-term clinical outcomes was not shown.

Conclusions
In summary, the introduction of BiPAP for post-extubation respiratory support did not reduce extubation failure rates versus NCPAP. However, BiPAP is shown to be superior to NCPAP in improving oxygenation and carbon dioxide clearance. Before the routine clinical application of this ventilation mode, more research is needed in the future to con rm its effectiveness and safety.