In clinical practice, it was difficult to find true patients who failed weaning from ventilation after extubation. SBT provided an effective way to screen suitable intubated patients. However, its role was overestimated in difficult-to-wean patients. In this study, SBP and MAP and RR remarkably increased during a SBT. It implied that cardiac afterload and the work of respiratory muscles increased during a SBT. When cardiopulmonary workload increased to a certain degree extent, the body failed to compensate for the changes and SBT failed as a result. In this study all patients successfully passed SBT. However, 29 cases (48%) developed postextubation respiratory failure, and 14 cases (23%) were re-intubated or died after extubation. These results revealed that SBT was not a sensitive way to screen out patients with postextubation respiratory failure or re-intubation, especially in difficult-to-wean patients. It was necessary to find a way to predict extubation outcomes precisely.
Comparisons of E/Ea in different groups before and after SBT
TTE and diaphragm ultrasound provided an effective and sensitive way. In this study, E/Ea and DE were closely related to respiratory failure within 48 hours and re-intubation within 1 week after extubation, respectively. These conclusions were published in the previous study (16). In this study, E/Ea at different positions had the similar predicting value of respiratory failure within 48 hours after extubation.
In 2020, Sanfilippo F etal (6) performed a meta-analysis to manifest the association of weaning failure from MV with TTE parameters. They found that weaning failure was significantly associated to a higher E/Ea ratio, not LVEF and the ratio of mitral Doppler early peak diastolic velocity to late peak diastolic velocity (E/A). This meta-analysis also reported that significant differences between E/Ea in three groups (P = 0.04) including septal, lateral and average, with moderate heterogeneity (I2 = 68.6%). In 2011, Papanikolaou J etal (23) reported that E/Ea before SBT at different positions was all closely related to weaning outcomes. However, up to date, few studies directly compared the differences between E/Ea at different positions. In this study, no differences were found at different positions between each other. It could be reasonably explained that all E/Ea at different positions could reflect cardiac diastolic dysfunction according to the guideline (24). Only the cut-off values of E/Ea at different positions might be different, but the predicting value of extubation outcomes could be similar between each other.
On the other hand, △E/Ea might reflect the of myocardial reserve to cope with the stress of SBT. Were acute alterations in E/Ea during a SBT associated with extubation outcomes? The conclusions in different studies were not completely in accordance. In 2010, Caille V etal (25) found that there were no differences between E/Ea (lateral) before and after SBT in three groups according to the level of LVEF (>50%, 35-50%, <35%). However, Bedet A etal (26) in 2019 revealed that during a SBT, echocardiographic surrogates of left atrial pressure (E, E/A, and E/Ea) increased remarkably in the failure group, but not in the success group. Similar results were reported in the studies by Moschietto et al (22) and E. Gerbaud etal (27) that a significant increase in the E/Ea ratio was observed in weaning failure group during a SBT, while no variation was observed in weaning success group.
Furthermore, Papanikolaou J etal (23) found that SBT-induced acute alterations in E/Ea tended to reach statistical significance between weaning success group and weaning failure group, though they were not predictive of weaning outcomes. In 2019, another interesting study by Kaltsi I etal (28) reported that 11 failed-to-wean patients in the first SBT received a 24-hour infusion of levosimendan, and 8 patients successfully weaned. E/Ea remained constant at 10 throughout the second SBT, whereas it increased from 10.5 to 12.9 during the first SBT. At the same time, LVEF increased from 30±10 to 36±3%.
In this study, △E/Ea during a SBT was only slightly higher in RF group than that in ES group. There was no statistical difference in two groups. However, the combined factor showed the highest diagnostic accuracy in predicting respiratory failure, compared to E/Ea (average) or △E/Ea (average) individually. △E/Ea could reflect the changes of cardiac diastolic function precisely during a SBT, just same as SBP and MAP and RR remarkably increased during a SBT. Therefore, the combined factor might be identified as the most suitable predictor. These results implied that both cardiac diastolic function in itself and cardiac tolerance during a SBT could be closely associated with respiratory failure. However, these conclusions were drawn carefully. Comparisons of the ROC curves in three variables showed no statistical differences between each other. Futhermore, in three studies mentioned above (22, 26, 27), SBT was performed during 1 to 2 hours, remarkably longer than 30 minutes in our study. Whether SBT with a longer time could induce remarkably the changes of cardiac diastolic function and thus the combined factor could screen out more patients with weaning failure needed further investigation.
Comparisons of DE in different groups before and after SBT
In this study, DE at different positions had the similar predicting value of re-intubation within 1 week after extubation. Some studies reported that both DE (right) and DE (left) were closely associated with extubation outcomes (11,12,14). Due to easily measured, only DE (right) was analyzed to predict weaning failure in other studies (19,29-31). However, it was common for unilateral diaphragm dysfunction in patients required MV beyond 48 hours (11). Thus, in this study all patients enrolled must be observed distinct bilateral diaphragm images except only five patients of poor echocardiography images or ultrasonic windows of DE. Up to date, no studies directly compared the predicting value of DE at different positions on extubation outcomes. In this study, no differences were shown between DE after SBT at different positions. These results implied that the assessment of unilateral diaphragmatic function could reflect overall diaphragm function, often affected by the common factors such as systemic inflammatory response, malnutrition, chronic hypoxia and so on.
Furthermore, most studies paid more attention to DE after SBT, not that before SBT (8,9,32,33). Similarly, in this study only DE after SBT was associated with re-intubation within 1 week. DE reflected the interaction of diaphragmatic function and MV. Accordingly, DE before SBT could not give expression to diaphragm contractility truthfully. 30mins T-tube SBT induced potential diaphragm contraction. △DE (average) was remarkably higher in NI group than that in RI group. These results implied that patients in NI group had stronger diaphragm reserve function.
In 2019, Xia Zhang etal (4) reported that in mechanically ventilated patients with chronic obstructive pulmonary disease, a cut-off value of ΔDE30−5 (the variation between 30 and 5 min during a SBT) > 0.16 cm was associated with a successful extubation with a sensitivity of 84% and a specificity of 83.3%, respectively. ΔDE30−5 might more sufficiently reflect the endurance of diaphragmatic contraction than DE. Another study by Palkar A etal (15) in 2018 showed that DE in extubation failure group was lower than that in extubation success group during a SBT. A decrease in DE of < 16.4% measured serially between assist control mode and SBT performed better than DE alone to predict extubation success with a sensitivity of 84.9% and a specificity of 65%.
In this study, combination of DE (average) and △DE (average) showed the highest diagnostic accuracy in predicting re-intubation, compared to DE (average) or △DE (average) individually. The combined factor reflected not only basic diaphragm function without MV, but also diaphragm potential reserve function. Although comparisons of three ROC curves also showed no statistical differences, there was a tendency that the combined factor could predicted re-intubation better. In clinical practice, it might be a more resonable choice to combine DE and △DE for assessing re-intubation.
Our study had several limitations. Firstly, we did not compare DE and E/Ea in different modes of SBT. In theory, free of ventilation in a mode of T-tube trial could stimulate the cardiopulmonary potential function at the greatest extent. Only a meta-analysis of Sanfilippo F etal (6) reported that there were no group differences in E/Ea according to the type of SBT. DE was primarily related to the inspired volume (34), regardless of whether it depended on the muscle effort or the ventilatory support. Some studies suggested that DE should be examined in the mode of absence of the ventilatory support (9,35). However, up to date few studies directly compared DE in different modes of SBT. Therefore, future study should be designed to compare directly cardiac function and diaphragmatic function in different modes of SBT. Secondly, diaphragm was the main respiratory muscle, but not the only one. Diaphragm dysfunction could be partly compensated by other respiratory muscles such as extra-diaphragmatic inspiratory muscles and abdominal wall expiratory muscles during a SBT (5). In fact, a proportion of patients could be successfully weaned from the ventilator despite having diaphragm dysfunction. Therefore, a combined ultrasound approach, including a systematic ultrasound evaluation of the heart, lungs and respiratory muscle pump might play an important role in difficult to wean patients (5,36-38). Thirdly, DTF different to DE, had been proved a good indicator reflecting the active contractility of diaphragm (9,33). It was valuable to manifest the effects of DTE at different positions during a SBT on extuabtion outcomes in future study.