This is the first clinical trial evaluating the efficacy and safety of mechanical in-exsufflation in combination with manual chest physiotherapy and endotracheal suctioning in a cohort of sedated critically ill patients on MV. The combination of CPT and MI-E was associated with a significant increase in the amount of retrieved pulmonary secretions when compared with CPT alone. Additionally, MI-E was safe and resulted in a short-term improvement in respiratory system compliance.
In a mixed population of 180 intubated and critically ill patients, Ferreira De Camillis et al. found a significant improvement in the weight of suctioned mucus using MI-E rather than CPT (17). In this previous study, CPT included manual chest compressions, applied upon left and right lateral decubitus, followed by manual hyperinflation. Unfortunately, manual chest compression was poorly described, challenging extrapolation on the benefits of MI-E or inefficacy of the applied CPT techniques. Conversely, in a cross-over study in 43 invasively ventilated ICU patients, Coutinho et al. found negligible effects of MI-E performed prior to TS (16). Coutinho’s study presented some incongruencies between planned primary outcome (volume of secretions) and assessed outcome (weight of secretions), questioning whether the study was adequately powered to achieve pre-planned aims. To the best of our knowledge, our study is the first to comprehensively assess the effects of MI-E and CPT, while avoiding potential biases. We found that when MI-E was used in conjunction with CPT, mucus clearance improved. A potential explanation for our positive results is the resulting PEF during MI-E (i.e. 96.9 [20.6] l/min). Previous studies in intubated and ventilated patients have implied that PEF higher than 160 l/min is required to mobilize secretions (27). Yet, a more recent review concluded that a cut-off point between 64–126 l/min could promote benefits in efficient clearance and successful weaning (28). Mechanistic in-vitro studies by Guerin et al. concluded that MI-E pressures should be increased up to 50 cmH2O to achieve the aforementioned PEF values (29) when endotracheal or tracheostomy tubes are used. In our study, patients were fully sedated, hence unable to perform expiratory efforts, while the resistance imposed by the ETT decreased the expiratory pressure by 34%; irrespective, we achieved efficient PEF figures as reported above.
In our settings, the MI-E device was set at + 40/-40 cmH2O pressure with middle inspiratory flow. We used these pressures because in previous clinical studies (14,19), patients better tolerated the inspiratory and expiratory efforts. Moreover, as recently demonstrated in a bench study (30), a reduction in flow rate during MI-E inspiratory phase augments the expiratory flow bias and enhance mucus displacement. Indeed, Volpe et al. achieved a higher PEF:PIF ratio, by setting MI-E to low inspiratory flow with an expiratory pressure higher than inspiratory pressure (i.e.: +40/-60). Conversely, in our settings the inspiratory and expiratory pressures were equivalent and we decided for middle flow, which may explain our slight inverse relationship between mucus clearance and PEF, while we failed to find an association with PEF:PIF ratio.
Ferreira de Camillis et al. observed short-term improvement of lung compliance following application of MI-E, in comparison with CPT (17), but long-term follow up was overlooked. Recently, Nunes et al. investigated in 16 intubated patients the effects of different MI-E pressure combinations vs. standard TS (18). In this randomized cross-over trial, when inspiratory/expiratory pressures of 50 cmH2O were applied, lung compliance improved immediately after the intervention, and 10 minutes thereafter. Despite these encouraging previous results, other publications have consistently failed to find benefits (15,16,18). In our study, compliance of the respiratory system increased immediately after CPT + MI-E intervention. Potentially, these positive variations in compliance are related to a higher number of performed MI-E cycles (30 and 20, respectively) in comparison with other negative studies that used lower number of cycles. Nevertheless, we found that the improvement in lung compliance was not sustained one-hour post-intervention and was not significant between groups. One possible explanation for this short effect is that MI-E acts as a recruitment maneuver, but if positive end-expiratory pressure is not adjusted following MI-E, derecruitment can still occur in the follow-up period.
Nunes et al. observed an improvement in SpO2 after either endotracheal suctioning or MI-E (18), while other publications failed to corroborate these results (15,16). Sanchez-Garcia et al. demonstrated an improvement in PaO2 following MI-E (15). Yet, it should be emphasized that in the study by Sanchez-Garcia continuous flow of oxygen at 8 l/min was administered at the filter port, adjacent to the MI-E device. Conversely, in our study, supplementary oxygen was not administered, but still an improvement in PaO2 and SaO2 was observed. In this context, it should also be taken into account the effects of MI-E on pulmonary perfusion. Indeed, during insufflation, MI-E creates a high transpulmonary pressure, which may displace blood toward collapsed alveolar regions, resulting in increased shunt and worse saturation (31).
In our study, a few episodes of respiratory and hemodynamic changes occurred during MI-E. However, these events did not differ significantly from the control group, lasted for a very brief period, and early protocol interruption was never necessary by attending clinicians. Indeed, both the MI-E and control group mainly experienced a slight increase in blood pressure and heart rate during interventions. Importantly, occurrence of these events was registered at the end of each intervention. This highlights that MI-E, CPT and TS all affects hemodynamic parameters, which should be taken into account in patients at risk of cardiac complications. In addition, as previously reported MI-E at times has been associated with hypoxemia, derecruitment, and pneumothorax (14,32); thus indication in patients with underlying acute or chronic pulmonary diseases should be carefully pondered. Finally, during both procedures cuff pressure decreased substantially to 22.9 (4.90) cmH2O in CPT + MI-E group and 24.32 (3.61) cmH2O during CPT alone. Main complication of this deflation may be ventilator acquired pneumonia (33, 34) which is directly related with an increase of burden during recovery (8). Despite these results, clinical guidelines recommend to maintain cuff pressure between 20 to 30 cmH2O to avoid microleaks and ventilator acquired pneumonia (35), being in consonance with our results.
Some limitations in this study merit consideration. First, volume of secretions was applied as a surrogate endpoint of mucus clearance, which could ultimately decrease accuracy of our results. However, the crossover design was chosen specifically to offset these limitations and prevent significant differences among patients. Second, respiratory physiotherapists could not be blinded to treatments allocation. Finally, the small sample size should be acknowledged. Yet, we carried out a comprehensive sample size analysis, based on a previous pilot study including 15 patients and in line with previous studies in this field of investigation.