Comparison of Success Rate of Extubation Between Spontaneous Breathing Trial with Pressure Support Ventilation and T-piece in Patients with Expiratory Flow Limitation

Background: Expiratory flow limitation (EFL) is an inability to exceed a certain flow regardless of the pressure exerted, which appears in 60% of patients with extubation failure. However, an appropriate weaning method for the patients with EFL remains unknown. Objectives: We aimed to evaluate an effect of different techniques of spontaneous breathing trials (SBT) on a success rate of extubation in the patients with EFL. Methods: We conducted a non-inferiority randomized controlled trial comparing between 30-minute pressure support ventilation (PSV) and 30-minute T-piece in ventilated patients with EFL who got ready to wean. The primary outcome was successful extubation during a 72-hour post-extubation period. Secondary outcomes were a reintubation rate within 7 days, time to reintubation, and a SBT success rate. Results: A total of 99 ventilated patients with EFL consisting of male (50.5%) with the median age of 70 [23] years were recruited. The most common cause of acute respiratory failure was intrapulmonary cause (68.7%). The eligible patients were randomized into 2 groups with a ratio of 1:1. The success rate of extubation during the 72-hour period in the PSV group was non-inferior to the T-piece group: 82% and 81.63%, respectively (95%CI -0.148 to 0.156, p=0.0475). There was also non-inferiority in the reintubation rate within 7 days (22.9% in the PSV group VS 15.2% in the T-piece group; 95%CI -0.081 to 0.235, p=0.005) and the SBT success rate (96% in the PSV group VS 93.9% in the T-piece group; 95%CI -0.065 to 0.108, p<0.001). There was no significant difference in the median time to reintubation between these 2 groups (55 [95.5] hours in PSV VS 25.33 [48] hours in the T-piece group, p=0.683). Conclusions: Among patients with EFL, the 30-minute SBT with PSV was non-inferior to the T-piece SBT in terms of the successful extubation during the 72-hour period, successful SBT, and reintubation rate within 7 days. This was the first study demonstrating that the different techniques of SBT did not affect the weaning outcomes in the patients with EFL.

Theoretically, EFL might play an important role in liberation from mechanical ventilation. However, the evidence of suitable techniques for weaning in the patients with EFL was limited. Practically, SBT is considered as a standard and effective weaning method 16,17 which consists of low-level pressure support ventilation (PSV) and T-piece. The majority of previous randomized controlled trials of different weaning methods demonstrated the similar extubation outcomes between the PSV and T-piece 18,19,[21][22][23][24][25] . In contrast, the recent multicenter study from Spain 27 showed a significantly higher rate of successful extubation in the PSV, compared with the T-piece (82.3% in the PSV group and 74% in the T-piece group, p=0.001) but showed the similar rates of reintubation (11.1% in the PSV group and 11.9% in the T-piece group, p=0.630). It might be explained that the heterogeneity of research protocols, including duration of SBT and the use of non-invasive ventilation (NIV) prophylaxis after extubation, led to the inconsistency of the results.
Noticeably, 60% of the patients with unsuccessful extubation had EFL 9,28 . Previous studies showed that positive end expiratory pressure (PEEP) 29,30 , upright position 31 , bronchodilator 32 , and diuretics 33 possibly improved extubation outcomes in the patients with EFL.
Up to our knowledge, there were no studies of weaning methods focusing on the patients with EFL.
We hypothesized that a SBT technique with PEEP might be non-inferior or superior to other SBT techniques in terms of successful extubation among the patients with EFL. Thus, we conducted the first study to evaluate effects of 2 SBT techniques on extubation outcomes, namely a rate of successful extubation, a SBT success rate, a reintubation rate within 7 days, and time to reintubation in the patients with EFL.

Study design:
We conducted a single-center non-inferiority randomized controlled trial, which enrolled ventilated patients with EFL between June 2020 and January 2021.

Patients:
Eligible patients were age of ≥ 18 years with EFL, who were assisted with mechanical ventilation for at least 24 hours and considered as readiness for liberation from mechanical ventilation. The criteria for readiness to wean were as followed: (1)

Protocol:
All mechanically ventilated patients were screened for EFL by a PEEP reduction maneuver 9 . The patients were considered to have EFL if expiratory flow did not increase when PEEP was decreased by 3

Outcomes:
The primary outcome was the success rate of extubation, defined as free of mechanical ventilation during the 72-hour post-extubation period. Secondary outcomes were the success rate of SBT, reintubation rate within 7 days, time to reintubation, and reasons for reintubation. Exploratory outcomes were changes of the ratio of arterial oxygen partial pressure to fractional inspired oxygen (PF ratio), partial pressure of arterial carbon dioxide (PaCO 2 ), and SpO 2 at 2 hours after extubation, duration of NIV prophylaxis, and reasons for NIV intolerance.

Statistical analysis:
We calculated the sample size for the non-inferiority trial with the margin of -0.15, an α=0.05 and a power of 80%. According to Subira C. et al study 27 , the successful extubation rates in the PSV group and the T-piece group were 82.3% and 74%, respectively, therefore 39 patients were required in each group. We estimated a drop-out rate of 10%, so the total number of patients in each group were 43.
Categorical variables were shown as numbers and percentages, and compared by Fisher's exact test. Continuous variables were shown as mean or median, and compared by independent t-test or Wilcoxon signed rank test, depending on the distribution of data. Intention-to-treat and per-protocol principles were used for the outcome analysis. A 1-sided α=.05 was considered statistically significant. Statistical analyses were performed by STATA version 16. Graphs for subgroup analysis were generated by R program.

Study participants:
Approximately, 167 patients receiving mechanical ventilation for at least 24 hours were screened.
Sixty-eight patients were excluded due to the absence of EFL. A total of 99 patients met the inclusion criteria and all were recruited into the study (Figure 3).

Figure 3 Flows of participants
Of 167 screened patients, 99 (59.28%) were recruited and randomized into 2 groups with a ratio of 1:1.
No patients dropped out of the study. Data of all patients were used for analysis.

Primary outcome:
The success rate of extubation during the 72-hour period in the PSV group (41 out of 50, 82%) was non-inferior to the T-piece group (40 out of 49, 81.63%); 95%CI -0.148 to 0.156, p=0.0475 (Table 2 and Figure   3). However, for the patients who passed SBT, the rate of successful extubation was inconclusive: 41 out of 48 (85.42%) in the PSV group and 40 out of 46 (86.96%) in the T-piece group; 95%CI -0.155 to 0.124, p=0.0588 (Table 3 and Figure 5).

Figure 4
Forest plot showing respiratory outcomes comparing between SBT with PSV and SBT with T-piece (Intention-to-treat analysis)

Secondary outcomes:
The SBT success rate in the PSV group was non-inferior to the rate in T-piece group: 96% in the PSV group VS 93.9% in the T-piece group; 95%CI -0.065 to 0.108, p<0.001 (Table 2 and Figure 4). Similarly, the reintubation rate within 7 days in the PSV group was non-inferior to the rate in T-piece group: 11 out of 48 compared with the T-piece group ( Figure 6).
The main reason for reintubation in the PSV group was secretion obstruction (27.27%), while the main reason in the T-piece group was pneumonia (42.86%). Other reasons for reintubation in both groups were shown in Table 5.

Exploratory outcomes:
The number of patients receiving NIV prophylaxis tended to be higher in the PSV group (66.67% from the PSV group and 45.65% in the T-piece group, p=0.061). Nevertheless, the duration of NIV use tended to be shorter in the PSV group (18.5 [18] hours in the PSV group and 23 [6] hours in the T-piece group, p=0.476). HFNC was replaced in all patients unable to tolerate NIV. The durations of HFNC use in both groups were similar (24 [27] hours and 24 [30] hours in the PSV and T-piece group respectively, p=0.551).
The major reason for NIV intolerance was patients' discomfort (62.5% and 68% in the PSV and the T-piece group, respectively). For the gas exchange at 2 hours after extubation, the changes of PF ratio, PaCO 2 , and SpO 2 were not different. (Table 4)

Subgroup analyses:
The subgroup analyses for successful extubation showed SBT with PSV was non-inferior to SBT with T-piece in the patients receiving mechanical ventilation for less than 4 days, patients without ARDS, and patients using NIV prophylaxis. However, the effect of these 2 SBT techniques was inconclusive in other subgroups (Figure 7).  On the contrary, the reintubation rate within 7 days tended to be lower in the T-piece group. It was possible that the workload from T-piece was harder than PSV, so the patients who could tolerate this most demanding test might have more physiologic reserve. In other word, pressure support augmented spontaneous breathing and reduced the work and oxygen cost of breathing while PEEP maintained airway opening and prevented atelectasis, which masked the actual respiratory muscle function of the patients in the PSV group. Thus, after withdrawal of this support, the PSV group tended to have greater reintubation events.
Additionally, the T-piece group received longer duration of NIV prophylaxis (18.50 [18] hours in the PSV group VS 23 [6] hours in the T-piece group), which minimized the risk of reintubation. Besides, sub analysis in the reintubated patients found the higher percentage of patients using NIV prophylaxis in the T-piece group (7 (63.64%) patients in the PSV group VS 5 (71.43%) patients in the T-piece group, p=0.061).
In addition, we compared our results with those from previous published studies (Table 6). Although EFL was one of the factors contributing to extubation failure, our findings showed the failure rate of extubation in the patients with EFL was comparable to other population. Furthermore, our study agreed with previous studies that PSV was not worse than T-piece in terms of the success rates of extubation and SBT.
Our study has several limitations. Firstly, the investigators and attending physicians could not be blinded due to the different devices used during the SBT, but the statisticians were blinded. Secondly, physicians' compliance to the NIV prophylaxis protocol was less in the PSV group (NIV duration < 24 hours), which confounded the reintubation outcome. Thirdly, some ventilators used in our study could not demonstrate the two different flow-volume loops on the same display, so the graphic comparison to determine the EFL had been made by manually taking 2 separated snapshots of the 2 consecutive flow-volume loops at PEEP 3 and at 0 cmH 2 O (ZEEP) with a mobile camera and then overlapping each other with an auto-align function in Adobe Photoshop (Adobe Inc., California, USA). However, this method had not been validated as the standard test to determine EFL, thus there might be a chance for misinterpretation of EFL. In addition, for intrinsic PEEP (iPEEP) measurement, we used a hyperventilation technique to abolish respiratory efforts before performing expiratory occlusion maneuver. As a result, the patient efforts might not be totally blunted, possibly leading to some errors of the measurement. Besides, the expiratory occlusion maneuver could not be performed in 8.1% of all patients because of no program for this maneuver available in some ventilators.
Furthermore, our findings showed the non-inferiority effects of these 2 SBT techniques in the patients with EFL with nearly zero intrinsic PEEP, thus in the patients with EFL with increased intrinsic PEEP, the SBT technique with PEEP might result in different outcomes. Finally, this was a single-center study with limited numbers of the participants. Notwithstanding, our initial findings pointed out to a need for further study in a larger population.

Conclusions
Our study was the first study demonstrating the non-inferior effect of the 2 SBT techniques, namely PSV and T-piece, on the weaning outcomes in the patients with EFL. Further research is needed to confirm our results.