Noninvasive Ventilation for De Novo Respiratory Failure: Impact of Ventilator Setting Adjustments

To the Editor:

One of the main aims of partial ventilatory support is to adequately unload the respiratory muscles. On one hand, overassistance may cause diaphragmatic dysfunction (1) and asynchrony (2); on the other hand, excessive inspiratory effort associated with insufficient amounts of assistance may lead to respiratory distress and expose the patient to the risk of self-inflicted lung injury (3). During noninvasive ventilation (NIV) for acute respiratory failure, pressure support level adjustment is therefore likely to influence the outcome. Especially, it is expected that an amount of assistance insufficient to reverse respiratory distress would lead to NIV failure and precipitate intubation. However, adjusting the amount of assistance in clinical practice remains a challenge. In fact, measuring the amount of respiratory effort requires the use of an esophageal catheter (4, 5) that cannot be routinely used during NIV.

We read with great interest the report of Tonelli and colleagues on a series of 30 patients during a 24-hour NIV trial for de novo acute hypoxemic respiratory failure in whom an esophageal catheter had been inserted (6). All patients exhibited an excessive inspiratory effort upon NIV initiation, with a median esophageal pressure swing (ΔPes) as high as 33 (interquartile range [IQR], 24–39) cm H₂O in the NIV success group and 38 (IQR, 32–42) cm H₂O in the NIV failure group (P = 0.1). After 2 hours of NIV, the patients who succeeded the NIV trial had dramatically decreased their respiratory effort. Thus, their ΔPes became significantly lower than those of patients who failed the NIV attempt (11 [IQR, 8–15] vs. 31.5 [IQR, 30–36] cm H₂O; P < 0.001). The authors concluded that the magnitude of respiratory effort relief within the first 2 hours of NIV was therefore an accurate predictor of NIV outcome.

As the respiratory effort is significantly influenced by the amount of assistance, however, such a conclusion may rely on an interpretation bias, and further details about the ventilator’s setting adjustments are needed to support it. In fact, for a comparable respiratory effort at baseline, the positive end-expiratory pressure (PEEP) and pressure support levels did not significantly differ between the NIV success and failure groups (8 [IQR, 6–10] vs. 8 [IQR, 7.5–10] cm H₂O and 11 [IQR, 10–14] vs. 11 [IQR, 10–12] cm H₂O, respectively; P > 0.0.5 for both comparisons), suggesting a comparable ventilatory demand. The authors did not report subsequent changes in respiratory support. After 2 hours of NIV, however, the magnitude of the difference in dynamic transpulmonary pressure, when compared with that of ΔPes, suggests different ventilator setting adjustments between patients who succeeded or failed the NIV attempt (30.5 [IQR, 28–43.5] vs. 39.5 [IQR, 37.5–42.3] cm H₂O; P = 0.04). If so, the following interpretation could rather be found: in patients with de novo acute hypoxemic respiratory failure, NIV may avoid early intubation when the amount of assistance is properly adjusted in a manner that allows a significant decrease in respiratory effort. Could the authors report the PEEP and pressure support levels after 2 hours of NIV to further assess this alternative conclusion?