Blowing Smoke: Examining the Benefits of High-Flow Nasal Cannula in Hypoxic Respiratory Failure: November 2015 Annals of Emergency Medicine Journal Club

Abstract

1.Frat et al1Frat J.P. Thille A.W. Mercat A. et al.High-flow oxygen through nasal cannula in acute hypoxemic respiratory failure.N Engl J Med. 2015; 372: 2185-2196Crossref PubMed Scopus (1310) Google Scholar randomized patients with noncardiogenic hypoxic respiratory failure to high-flow nasal cannula, noninvasive positive-pressure ventilation (NIPPV), or a nonrebreather face mask.A.Describe internal validity and how potential flaws in a trial’s internal validity may bias the results.B.Neither allocation concealment nor blinding was used by Frat et al1Frat J.P. Thille A.W. Mercat A. et al.High-flow oxygen through nasal cannula in acute hypoxemic respiratory failure.N Engl J Med. 2015; 372: 2185-2196Crossref PubMed Scopus (1310) Google Scholar in their trial design. What is the difference between blinding and allocation concealment? How might each bias a trial’s results? Did the authors take any measures to control for the lack of blinding and allocation concealment in their trial design?C.Patients randomized to NIPPV spent a mean duration of 8 hours per day receiving positive-pressure ventilation. When not receiving NIPPV, patients received high-flow nasal cannula. How might this influence the results?D.The authors allowed for patients in the high-flow nasal cannula and nonrebreather groups to receive NIPPV as rescue therapy if the treating physician deemed it clinically necessary. How many patients in the high-flow group and standard therapy group received rescue NIPPV therapy? How might this have influenced the results?2.A.Frat et al1Frat J.P. Thille A.W. Mercat A. et al.High-flow oxygen through nasal cannula in acute hypoxemic respiratory failure.N Engl J Med. 2015; 372: 2185-2196Crossref PubMed Scopus (1310) Google Scholar used a specific set of inclusion and exclusion criteria to select patients for enrollment. How might these criteria influence the generalizability (ie, external validity) of the trial results?B.The authors define a nonrebreather face mask as the general accepted “standard of care” in the management of hypoxic respiratory failure. Is this truly considered the standard therapy for patients with noncardiogenic hypoxic respiratory failure? How does the selection of such a standard affect the interpretation of the results?C.Examine the trial’s enrollment flow chart (Figure 1 in Frat et al1Frat J.P. Thille A.W. Mercat A. et al.High-flow oxygen through nasal cannula in acute hypoxemic respiratory failure.N Engl J Med. 2015; 372: 2185-2196Crossref PubMed Scopus (1310) Google Scholar). Of the 2,506 patients with hypoxic respiratory failure who met the study’s inclusion criteria, only 313 underwent randomization. How might such selective enrollment affect the validity of the trial? How might the exclusion criteria limit the applicability of these results to specific patient populations treated in the ED?3.Frat et al1Frat J.P. Thille A.W. Mercat A. et al.High-flow oxygen through nasal cannula in acute hypoxemic respiratory failure.N Engl J Med. 2015; 372: 2185-2196Crossref PubMed Scopus (1310) Google Scholar powered the trial to detect an absolute difference of 20% in the rate of intubation at 28 days.A.Define statistical power. How did the authors justify their power calculation?B.What is the purpose of a power calculation? Is a power calculation always necessary?C.Was this study underpowered? What criteria might suggest whether a study is underpowered?4.Frat et al1Frat J.P. Thille A.W. Mercat A. et al.High-flow oxygen through nasal cannula in acute hypoxemic respiratory failure.N Engl J Med. 2015; 372: 2185-2196Crossref PubMed Scopus (1310) Google Scholar found no statistically significant difference in the rate of intubation at 28 days between the control group and either high-flow nasal cannula or NIPPV.A.Define frequentist and Bayesian statistics. What are the primary differences between the 2 approaches to data interpretation and trial design?B.From a frequentist perspective, this trial was a negative-result study. How might the results have differed if a Bayesian approach had been used?C.How might this trial change practice?