Trends in use and benefits of non-invasive ventilation as first-line therapy in acute respiratory failure.

Abstract

Dear Editor, In a recent paper examining trends in the use and outcome of non-invasive mechanical ventilation (NIV) in acute respiratory failure (ARF) over a 15-year period (1997–2011) in 14 French intensive care units (ICUs), Schnell and colleagues [1] report a steady increase in the use of NIV as first-line ventilatory support, reaching 42 % in 2011. We recently reported the rates of intubation in hypercapnic [2] or hypoxemic patients [3] receiving NIV as first-line therapy in our ICU over a 3-year period: 430 patients, i.e., more than 140 patients per year (after excluding 35 patients with a ‘‘do not resuscitate’’ order) received NIV as first-line ventilatory support for respiratory failure of any cause (Fig. 1). In comparison, the study by Schnell et al. included 974 patients over a 15-year period, i.e., \5–10 patients per center and per year, bringing into question the representativeness of their cohort. Schnell and colleagues [1] report intubation rates of 21 and 18 % in patients with acute-on-chronic respiratory failure and in those with cardiogenic pulmonary edema, respectively. In our cohort, the rates of intubation were markedly lower both in patients with acute-onchronic respiratory failure (15 %) and in those with cardiogenic pulmonary edema (only 6 %), possibly due to a high NIV-case volume. By contrast, the overall intubation rate was 51 % in our patients with de novo ARF (Fig. 1), which is markedly higher than the 34 % rate reported by Schnell et al., suggesting that NIV was probably initiated in more severe patients in our center. Indeed, 58 % of our patients with ARF received first-line NIV. Similarly, a 46 % intubation rate in patients with de novo ARF receiving first-line NIV was recently reported from other highly skilled centers [4, 5]. Therefore, an approximately 45–50 % intubation rate could be the reference in the most severe hypoxemic patients, even in experienced units. Carrillo and colleagues reported that delayed intubation was associated with a poor prognosis [4]. However, we did not confirm these results when using predefined intubation criteria and a shorter time to intubation [3]. As all of the above studies found an