Noninvasive ventilation usage time and survival rate in patients with acute respiratory failure: some key insights

From the authors:

We would like to thank Y.M. Madney and colleagues for their considered comments. The data in our recent study are predominantly related to the hypercapnic group reflecting the demand for noninvasive ventilation (NIV) [1]. We agree that the hypercapnic and hypoxaemic respiratory failure (HRF) cohorts were not matched, which is not unexpected when considering that the pathogensis of these two entities are very different. We have previously demonstrated that there are poorer outcomes with hypoxaemic than hypercapnic respiratory failure in a respiratory high dependency unit [2], and there is an increasing proportion of patients with HRF managed in high dependency units. We believe that our recent study indicates that the two entities have different dose–response relationships for NIV but agree that a larger HRF cohort is required to determine the dose characteristics of this group.

Y.M. Madney and colleagues also discussed the potential impact of the methodology on the outcomes, specifically the exclusion criteria for acute NIV, as well as the potential confounding factor of escalation to therapy and the possible impact of delayed intubation. In our institution, direct intensive care unit (ICU) admission for respiratory failure only rarely occurs; for example, where early intubation is imminently required or in the presence of multi-system disease where other system support, including inotropic, is required. The main absolute contraindication for acute NIV is patient preference (where end-of-life planning precludes NIV) but NIV would not be offered in patients where this therapy is considered futile (for example, progressive end-stage interstitial lung disease without an acute exacerbation). We do not believe that escalation of care, including intubation or delay of intubation, affected our outcomes. First, only 12 (<2%) of the whole cohort of 654 patients were transferred to ICU following NIV in the high dependency unit. Secondly, of the 37 patients who ceased NIV out-of-protocol (before clinically stable), only one was transferred to the ICU. This patient did not receive invasive ventilation and survived to discharge. Finally, NIV was the ceiling of care in 211 patients (32% of the cohort) due to preferences or severity of comorbidities. These patients were not considered for escalation of care beyond acute NIV. Therefore, we believe that our findings were not biased by these methodological issues.

• Received May 5, 2020.
• Accepted May 5, 2020.

• Copyright ©ERS 2020