The weaning process consists in liberate patients from mechanical ventilation (MV) removing the endotracheal tube after well tolerate spontaneous breathing.1,2 To determine the optimal moment for extubation remains a challenge, so current guidelines recommend the implementation of a spontaneous breathing trial (SBT) as a tool to predict weaning outcome.2–4 However, 10%–25% of patients who are extubated following a successful SBT need to be reintubated, which is associated with higher mortality.5,6 The T-piece trial (T-T) and low level of pressure support ventilation (PSV) are the most common modes of SBT used in clinical practice but there is still no consensus on which of them are associated with better outcomes.1,4,7,8 While some studies suggested that PSV is related to a higher proportion of successful extubation, some meta-analysis advocate for the T-T method.3,7,9,10 Recently, Li et al., have also published a meta-analysis suggesting comparable predictive power of successful extubation between them in critically ill patients.11

Since loss of aeration and diaphragm dysfunction have been related to weaning failure, they should be taken into account when evaluating an extubation attempt.12–21 Some authors evaluated the utility of electrical impedance tomography (EIT) to predict SBT failure during the weaning process.12–14 Specifically, EIT showed a significant decrease in tidal variation of impedance and end-expiratory lung volume after ventilator disconnection during a SBT.12,13 In this regard, SBT failure compared to success was characterized by a greater lung de-recruitment and more inhomogeneity of ventilation distribution measured by EIT.14 Transthoracic lung ultrasound (LUS) is also a non-invasive, radiation-free, bedside tool, that allows to detect the amount of lung aeration loss or derecruitment of the lung during the SBT and could be an accurate predictor of weaning failure.15,16,22 Different studies agree that there is a greater lung derecruitment evaluated by LUS in patients who develop weaning failure after successfully pass the SBT (both T-T and PSV), in comparison with other who are successfully weaned.16–18,23 Moreover, current literature have suggested that diaphragmatic ultrasonography (DU) could be an accurate tool to identify diaphragmatic dysfunction, to monitor respiratory work-load and, also, to predict extubation failure during weaning from MV and SBT.20,21,24

However, heterogeneous results have been published and there are no recommendations on which SBT to use based on patient characteristics to achieve better outcomes. Nevertheless, differences in aeration between different SBT have not been compared, so further investigation is needed to determine the best approach. In that sense, we hypothesized that both SBT, T-T and PSV, could generate different changes in aeration and respiratory drive which could be related to differences in successful extubation rates. Therefore, this study aimed to evaluate the changes in lung aeration evaluated with EIT and LUS, and respiratory effort estimation using DU, generated by two different SBT (T-T vs PSV).

A total of 58 patients were enrolled in the study. Among them, 43 were finally included, 20 (46%) being randomized to the T-T group and 23 (53%) to the PSV group (Fig. 1). Table 1 shows the principal characteristics of included patients. Briefly, there were no differences in demographic variables, comorbidities and chronic treatment as well as, in the reason for ICU admission, severity scores, and treatment received during ICU stay. Furthermore, no differences were found in extubation failure, in the NIRS used after extubation (used in almost 50% of patients in each group), ICU LOS, or in-ICU mortality between groups.

Figure 1.

Flow chart of participants.

SBT: spontaneous breathing trial; T-T: T-piece trial; PSV: pressure support ventilation.

(0.14MB).

No differences were found in physiological parameters (RR, SBP, HR and SpO2) evaluated at any study timepoints between groups (see supplementary material OS Table 1).

Aeration changes

Aeration changes evidenced by EIT and LUS in the two SBT groups at the different protocol monitoring times are shown in Fig. 2. At baseline, no significant differences were found in EELI nor LUS between PSV and T-T (Fig. 2A). During the study period in both SBT groups, EELI and LUS score showed a trend to loss of aeration (trend to decrease EELI score and to increase LUS score) from BSL to EoSBT and OTE in PSV and T-T although without differences between them (Fig. 2A). Fig. 2B shows the ΔEELI and ΔLUS between the different study timepoints in both study groups, showing in all the cases a loss of aeration. In the T-T group, the greatest aeration loss occurred during the SBT period (EELI and LUS). Only differences between study groups in ΔEELI between EoSBT and OTE were statistically significant, showing higher loss of aeration in PSV compared to T-T [−1992.40 (−3458.76, −5.07) AU vs 48.09 (−1319.71, 1084.07) AU, p = 0.027] (Fig. 2B). Percentage of variation calculated between BSL and OTE, was greater when LUS was used (Δ%LUS 60.6%) in comparison with EIT (Δ%EELI 4.7%), p ≤ 0.001 (Fig. 2C).

Figure 2.

Aeration according to EIT and LUS by SBT (A), aeration changes by SBT (B) and percentage of variation between BSL and OTE of EELI and LUS (C).

PSV: pressure support ventilation; T-T: T-piece trial; EELI: end-expiratory lung impedance; LUS: lung ultrasound; AU: arbitrary units; BSL: basal; EoSBT: end of spontaneous breathing trial; OTE: one hour after orotracheal extubation.

(0.35MB).

The subanalysis of aeration by groups of SBT regarding the use of NIRS (NIV or HFNC) at one-hour post-extubation is shown in supplementary material OS Fig. 2. There were no significant differences between groups according to receive NIRS at OTE.

Diaphragmatic ultrasound measures

Fig. 3 shows the data of DU in both study groups at different timepoints (DE, TF and TI). There were no differences between any of them at baseline, neither in the other timepoints (Fig. 3A). However, DE tended to decrease during extubation in both SBT groups, coinciding with a loss of aeration. During the weaning process, the decrease trend in DE was greater in the T-T group. Furthermore, there were no differences regarding ΔTF between SBT groups (Fig. 3B).

Figure 3.

Diaphragm ultrasound measurements by SBT groups (A) and delta diaphragm thickening fraction by SBT (B).

PSV: pressure support ventilation; T-T: T-piece trial; BSL: basal; EoSBT: end of spontaneous breathing trial; OTE: one hour after orotracheal extubation; TF: diaphragm thickening fraction.

(0.28MB).

The main findings of this study are that: 1) there are no differences between PSV and T-T in generating aeration loss measured by EIT and LUS, and 2) there are no differences between PSV and T-T regarding respiratory effort or tidal volume estimated using TF and DE respectively.

For this study and in order to assess changes in aeration and respiratory drive between two SBT strategies, we used the two previously described with the greatest difference in respiratory effort and degree of demand between them: T-piece and PSV with 5-cmH2O PEEP.7,8,11,31 In addition, we performed the SBT during 30 min since no differences have been found between 30 min and longer duration (120 min) SBT on checking breathing capacity related to successful extubation rates.32,33 Supporting that, in our study we observed that using the T-T test for 30 min there was no greater loss of aeration than PSV (measured by EIT and LUS), despite in this group the greatest aeration changes occurred during the SBT itself. Moreover, there were no differences in the estimated tidal volume with DE or estimated respiratory effort with TF, compared to the PSV group.

Different studies have found significant lung derecruitment after an SBT based on both, EIT or LUS as a tool to measure pulmonary aeration. Those studies showed that lung derecruitment is greater among patients who develop weaning failure after successfully pass the SBT, in comparison with patients who are successfully weaned.13,14,16,17,23 In the present study, we found a loss of aeration during the weaning process regardless the tool used to evaluate lung derecruitment, or the SBT strategy used. Until now, there is no agreement in the literature regarding superiority of a specific SBT to others in determining extubation failure. Although it has been suggested that SBT with a ventilator support could overestimate patients capability to maintain the cardiorespiratory load after extubation,34 we found no aeration differences between the two SBT during the weaning process. However, while we included a general population, other studies as Cabello et al.8 included specific subpopulation of difficult-to-wean subjects with a high prevalence of left ventricular heart failure where ventilator support could have modified cardiovascular and respiratory response in comparation with T-piece. Moreover, our absence of differences in aeration between both SBT groups could also be explained by using a T-T of only 30 min duration, as well as the small sample size.

Analyses at different study timepoints, showed that the greatest aeration loss occurred in the T-T group was during the SBT period. These results are in accordance with previous studies,7,25,32 where only 30 min SBT allows discrimination of patients who are probably going to fail without producing a greater loss of aeration and could justify the benefit of a rest period with reconnection to the MV after a successful T-T trial before extubation.35 Moreover, we found a higher loss of aeration measured by EELI in PSV compared to T-piece between EoSBT and OTE, probably related with the higher aeration loss during the SBT in the T-T group, as commented. That could also support findings of Subirà et al. who recommended a less demanding ventilation strategy during weaning trial to avoid delay on extubation in patients able to spontaneously breathe without the ventilator, although without assessing pulmonary aeration.7

In addition, literature supports both, EIT and LUS, to measure pulmonary aeration. However, they have not been compared between them. Our results showed that between BSL and OTE timepoints, the Δ%LUS was greater than Δ%EELI which could reflect a higher sensitivity of LUS over EELI in order to detect loss of aeration. This phenomenon could be explained by a greater number of thoracic regions evaluated using LUS.

Current literature suggests that DU could be an accurate tool to identify diaphragmatic contractility in patients under MV to predict extubation failure, to measure loss of aeration through DE changes as an estimator of tidal volume, as well as, to monitor diaphragm effort.24,30,36,37 On the one hand, DE performed during an SBT in intubated patients has proven to be useful to predict weaning failure. In this regard, the cut-off ranged from 10 to 14 mm during normal spontaneous breathing have been selected to diagnose diaphragmatic disfunction.24,38 On the other hand, TF has shown significant correlation with invasive techniques such as oesophageal and transdiaphragmatic pressure monitoring (the gold-standard to assess inspiratory diaphragmatic effort in mechanical ventilated patients)36,39 as well as to predict extubation failure during an SBT (cut-off ranged between 30 and 36%).24,37 In this line, Xia et al.40 showed that patients with a greater loss of pulmonary aeration, had increased diaphragmatic contractility, indicating an additional respiratory effort to compensate lung volume loss. In our study, there were no differences in respiratory effort between groups, in accordance with no differences in aeration loss. Furthermore, in the present study we observed that when MV were disconnected during the SBT in the T-T group, DE decreased. In fact, excursion is directly proportional to inspired volume in both the supine and sitting position, regardless of whether it depends on the force of the diaphragm contraction by itself or passive displacement by ventilator support pressure.24 Consequently, in our study we observed that DE as an estimator of lung tidal volume, tended to decrease with extubation, suggesting a loss of aeration in both groups.

Our study has several limitations to consider. First, prophylactic use of NIV or HFNC after extubation was not protocolized, but at the discretion of the treating physician. Then, it could have been used associated to patients with more comorbidities and higher risk factors for extubation failure. Therefore, since these are not homogeneous criteria, the results of lung aeration could be influenced by the use of NIRS. However, in our cohort of patients, there were no differences between groups regarding the use of NIRS after extubation and this would not be a confusion factor when comparing both strategies. Second, investigators and attending physicians were not blinded to weaning trial randomization group, but clinicians were blinded to their results and clinical management was not affected by the result of the techniques. Third, although it was not the aim of this study, the sample size and the study design did not allow conclusions to be drawn on clinically relevant outcomes such as weaning failure, hospital mortality, nor stratify aeration loss according to weaning failure. Besides, due to this study included general ICU patients, the results could not be extrapolated to specific group of patients as difficult weaning population, in whom SBT modality may have more clinical importance. Finally, this is an unicentric study with a limited number of included patients and further studies should be designed to generalize the results.

T-piece and PSV as different strategies to perform SBT do not show differences in aeration loss, nor estimated respiratory effort or tidal volume measured by EIT, LUS and DU. Further studies will be needed to evaluate the impact of those SBT strategies in a higher sample size to generalize these results.

RBC, JMC and PPT conceived and designed the study.

RBC, RM, ID, CC, JMC and PPT, contributed to patient and data recruitment.

RBC, FJP, JRM, JMC and PPT made important intellectual contributions and actively participated in the interpretation of the data.

RBC wrote the paper.

JMC and PPT done contributions to the writing and conceptual exposition of the results.

All authors contributed to critical examination of the paper for important intellectual content and approval of the final manuscript.

The authors declare that no funds, grants, or other support were received for this study.

The authors have no relevant financial or non-financial interests relevant to this article to disclose.

The study was conducted in accordance with the ethical principles of the Declaration of Helsinki, and approved by the Institutional Review Board of the PSMAR (Hospital del Mar, Barcelona, Spain, number 2015/6444/I).