Weak Cough Strength and Secretion Retention in Mechanically Ventilated Patients: Is There a Role for Cough-Assist Devices?

The article from Camillis et al¹ is one of the few studies on mechanically assisted cough for ventilated patients. Considering that the evidence on airway-clearance techniques in this setting is scarce and this device has proven efficacy for out-patients,² we think that this is a relevant contribution. However, the article deserves further comment.

In a controlled, randomized study, the authors observed that the use of the mechanical insufflation-exsufflation device (MI-E) in mechanically ventilated subjects resulted in greater secretion removal than a conventional chest physiotherapy protocol. The authors also found a significant increase in static lung compliance in the MI-E group compared to the conventional chest physiotherapy group (control). Work of breathing and airway-resistance variations were not different between the groups.

Respiratory physiotherapy is not just a matter of removing as much secretion as possible; it also involves being able to move peripheral secretions to more central regions so that they can be eliminated through coughing or suctioning.³ In this way, because mechanically ventilated patients are prone to secretion retention⁴ and present with reduced cough peak flow,⁵ cough-assist devices might bring some benefit. Because MI-E mimics a cough, it is likely that the secretions collected after this intervention in the study of Camillis et al¹ were predominantly from central airways, while in the control group, in which manual hyperinflation was used, may have had a higher effect on the lung periphery. This raises methodological concerns and some questions regarding the interpretation of the results. What were the effects of each intervention on secretion removal from large and small airways? Considering this, if the main outcome measure was recorded 30 min or 60 min later, would the results be different? If the subjects had been suctioned immediately before the interventions, would the results be different? What patient populations benefit from MI-E? Far from being simply mechanistic elucubrations, we believe that these questions should be addressed in future studies because the answers may influence the criteria for prescription and application of MI-E in patients on mechanical ventilation. It is also notable that the characteristics of the outcome measures (accuracy, reliability, and sensitivity)^6,7 and the sample heterogeneity increase the risk of bias in studies on airway-clearance techniques in mechanically ventilated patients. Direct measurements of mucus clearance, like those provided by radiopaque markers,⁶ are difficult to apply in the ICU setting. Thus, the commonly used marker of secretion clearance is the volume or weight of secretions collected after any intervention. For intubated patients, contamination with saliva or swallowing of mobilized secretions are not a concern, but the sputum volume/weight is not sensitive to small differences.⁶ Moreover, as mentioned before, lack of expectoration during physiotherapy treatments does not mean that mucus movement is not happening.⁷

Regarding the risk of selection bias, the amount of secretion and mucociliary clearance may vary among mechanically ventilated patients according to several factors, such as the use of drugs, fluid status, muscle weakness, humidification of respiratory gases, preexisting respiratory conditions, pulmonary infection, or ventilatory settings^1,6,7 Therefore, most studies on airway-clearance techniques have used a randomized crossover design, where the subjects serve as their own control.⁸ Conversely, in this study, the authors used a 2-arm design. In this case, the randomization may have minimized the between-group differences, but the use of any classification of secretion production at baseline (eg, scarce, if requiring endotracheal suctioning every 4 h or more; moderate, if requiring suctioning every 2–4 h; or abundant, if requiring suctioning every 2 h or less⁹) would be valuable to select the study participants or as a control variable.

The respiratory mechanics results of this study are difficult to interpret. There is no information on how static lung compliance, airways resistance, and work of breathing were recorded. The hyperinflation provided by the mechanical insufflation-exsufflation device (peak pressure = 40 cm H₂O) may have been responsible for the increase in static compliance. However, although manual hyperinflation was applied in the control group, no improvements in lung compliance and airway resistance were observed. Perhaps the volume delivered during manual hyperventilation was not enough to improve the respiratory mechanics in these subjects.

Finally, in the face of the results of this study and the rationale behind cough-assist modalities for intubated patients, we believe that the mechanically assisted cough has therapeutic potential for ventilated patients. It is likely that MI-E is useful for a selected group of patients, such as those with muscle weakness and without airway instability. Nevertheless, before its routine use, more studies are needed to investigate the safety and utility of MI-E for secretion removal from central and peripheral airways, as well as the impact on clinically relevant outcomes, including the incidence of ventilator-associated pneumonia, stay, weaning outcome, and mortality. In addition, because of the need to disconnect the patient from the mechanical ventilator, MI-E may be deleterious for patients at risk of lung collapse (high PEEP levels) or severe hypoxemia.¹⁰

• Copyright © 2018 by Daedalus Enterprises