The results of this study showed that MP, MP normalized to dynamic lung compliance, and MP normalized to PBW were independently associated with weaning outcomes in critically ill ventilated patients, and higher MP and norMP indicated an increased risk of weaning failure. In addition, norMP’s discriminatory performance for weaning failure was better than that of MP. This is the first large-scale study examining the relationship between MP and weaning outcomes in mechanically ventilated patients based on the MIMIC-IV database.
MP is a unifying concept proposed by Gattinoni in the context of acute respiratory distress syndrome. MP integrates several factors of mechanical ventilation, and the total energy delivered by the ventilator to the lung parenchyma can be calculated from a combination of the following parameters: tidal volume, PEEP, plateau pressure, peak inspiratory pressure, and respiratory rate4:12. MP measurement is simple and non-invasive, and the workload required to maintain optimal alveolar ventilation acting on the respiratory muscles per unit time can be achieved without turning off the bedside ventilator. As a result, MP has also become a predictor of clinical weaning in recent years5.6. Studies have indicated that MP is related to functional lung volume size and MP standardized as a surrogate for lung size (e.g., well-inflated lung tissue as determined by CT analysis, lung compliance, or predicted body weight) have higher diagnostic accuracy having4,13,14. Therefore, in examining the correlation between MP and weaning outcomes, this paper also examined whether norMP has a higher discriminating ability for failed weaning than MP.
Of the 3695 ventilated patients in this study, 61.5% were successfully weaned after the first SBT, while 38.5% failed weaning. The 28-day mortality rate for patients with successful weaning was 12.4% and the 28-day mortality rate for patients with failed weaning was 28.9%. A total of 11.1% of patients required reintubation 48 hours after successful SBT weaning, which is consistent with the results of a multicenter observational study by Jaber et al.fifteen. Compared to patients who successfully weaned, the duration of ventilation, length of stay in the ICU, length of hospital stay, 28-day reintubation rate, and 28-day mortality were significantly longer in patients with weaning failure, while the 28-day days without ventilation were significantly reduced, which is consistent with the results of Hayashi et al.16. The study identified several factors associated with weaning failure, including disease severity (SAPS II, SOFA), intake sources (SICU), respiratory load indicators (PEEP, PplatePsummit∆POhRR, MV, FiO2MP, CdynamicCdynamic-MP, PBW-MP), markers of inflammation (WBC), organ function and nutritional status (Albumiun, SCr), fluid management (Uorate), physiological status during weaning (HR, BF, MBP, SPO2, PH, PF) etc. (Table S2). The results showed that the dynamic driving pressure (∆POh) was associated with weaning outcomes, in contrast to the driving pressure (∆P), which may be related to the fact that peak inspiratory pressure is a more conservative measure of elasticity in nature than plateau pressure17. Logistic regression analysis showed that MP, Cdynamic-MP and PBW-MP were independently associated with weaning outcomes, consistent with the results of an observational study by Ghiani et al.7. This indicates that the pathophysiological mechanism of weaning failure is complex. Although the reversible factors leading to weaning failure have been actively addressed, the imbalance between respiratory effort and respiratory muscle performance remains the main reason for weaning failure18. MP includes several factors related to respiratory effort, such as: B. resistance, driving pressure, lung compliance and PEEP. A multitude of studies suggest that MP is more valuable for prognosis than a single parameter index19.20. The present study analyzed the respiratory mechanics indices of mechanically ventilated ICU patients before weaning and assessed the relationship between MP, Cdynamic-MP, PBW-MP or Weaning Results to further optimize the weaning decision. The results of this study showed that the Cdynamic of patients in the weaning failure group was lower, while the PEEP, ∆POhMP, Cdynamic-MP and PBW-MP were higher. In addition, MP, C. with increasing difficulty of weaningdynamic-MP and PBW-MP significantly increased. Comparison of MP and norMP as comprehensive mechanical indicators with individual parameters such as Cdynamic and ∆POhcan reflect not only the weaning outcome of ventilated patients, but also the severity of weaning difficulties.
The study also concluded that dynamic comparison of respiratory mechanics indices (including PEEP, CdynamicMP, Cdynamic-MP and PBW-MP) 24 h before weaning in the successful weaning group showed significant differences, while there was no dynamic change in the weaning unsuccessful group. This shows that the dynamic reduction in respiratory load such as MP 24 h before weaning is a potent predictor of weaning success. In mechanical ventilation, the ventilator’s drive system overcomes the patient’s airway resistance and the elastic resistance of the respiratory system, forcing the gas into the alveoli to participate in the respiratory mechanical process of gas exchange21:22. While PEEP can to some extent indicate the degree of stress inhomogeneity in lung tissue recruitment, it only relates to the static stress associated with PEEP and cannot express the dynamic stress acting on the lung tissue during respiratory ventilation23. In most cases the level of ∆POh is simultaneously affected by the patient’s chest wall compliance and lung compliance. Simply using propulsion pressure cannot accurately describe the state of compliance of the lungs24. MP integrates multiple respiratory mechanics parameters and can assess the impact of a mechanical breath on dynamic and static global lung load and energy load25. Furthermore, based on ROC curve analysis, the discriminatory power of MP on weaning outcome was stronger than that of the traditional pre-weaning RSBI evaluation index and respiratory mechanics index Cdynamic. This shows that the MP, which is a comprehensive respiratory mechanics index, can better reflect the workload on the respiratory muscles prior to weaning.
The size of the ventilated lung volume determines the magnitude of the energy per unit of lung tissue. Due to the different volume of the “baby lung”, the different thoracic compliance and the different load and stress limits that the lung tissue can withstand in different regions, the conversion of MP normalized to ventilatable lung volume may be necessary9:13. Cdynamic-MP, PBW-MP refer to the interaction between various pathophysiological characteristics of the lung (including size, homogeneity and recruitability) that reflect the actual energy expended per unit of ventilated lung volume4,26,27. The results of this study showed that MP, Cdynamic-MP and PBW-MP were all associated with an increased risk of weaning failure, higher MP or norMP scores predicted a greater risk of weaning failure. This indicates that the MP is related to the severity of the lung disease before weaning and patients in the weaning failure group with “sicker” lungs require more intensive ventilation28. Several studies have linked MP > 12 J/min to the occurrence of ventilator-associated lung injury (VILI).19:29 and poor clinical prognosis12,30,31,32. Here, the MP cutoff value for predicting weaning failure was > 11.3 J/min (sensitivity 71%, specificity 65%) and the corresponding area under the curve was 0.745 (95% CI 0.730-0.761). The discrimination power of Cdynamic-MP (AUROC 0.760 [95%CI 0.745–0.776]) and PBW-MP (AUROC 0.761 [95%CI 0.744–0.779]) was higher than that of MP (P< 0.05). This indicates that Cdynamic-MP and PBW-MP can better reflect the functional lung size and the pathophysiological state of the lungs during weaning in patients on mechanical ventilation and are better representatives of the actual energy delivered to the lungs by the ventilator and therefore have a higher predictive value for the weaning results.
There are some limitations to this study. First, we extracted data and calculated MP according to the simplified MP equation proposed by Gattinoni in the volume-controlled ventilation mode in our study. During our screening of the study population in the database, patients missing all required variables for MP calculation were excluded, including all patients with missing Pplatei.e. patients without Pplate Measurement under volume-controlled ventilation before SBT. As this study was a secondary analysis of the data set in MIMIC-IV for the purpose of clinical application, there was no guarantee that the parameters were collected under standard conditions without spontaneous breathing and with an adequate level of sedation. Nonetheless, the mean value of each parameter in each 4-hour time frame in the present study was recorded for 24 hours before weaning, and the dynamic changes in each respiratory mechanics index were monitored. Second, functional lung size was primarily assessed using lung compliance and predicted body weight, but computed tomography of the lungs may be a more accurate method for the amount of functionally ventilated lung remaining, lung inhomogeneity, or recruitability estimation9:33. However, currently there are few datasets containing this type of imaging and further verification in large-scale clinical trials is needed. Finally, studies have shown that T-tube ventilation and pressure support strategies during spontaneous breathing trials can have an impact on weaning outcomes10. Therefore, only patients using T-tube ventilation during weaning were included in this study, and further research on the predictive ability of MP in different SBT modes for weaning outcome is needed.