To the Editor We read with interest the study by Gupta et al1in the recent issue of JAMA Pediatrics. The value of high-frequency oscillatory ventilation (HFOV) in the management of severe acute respiratory failure is a recurrent question for pediatric intensivists.
In this retrospective observational study of a large database, Gupta et al1 compared the outcome of children with acute respiratory failure ventilated with either HFOV or by conventional mechanical ventilation. The authors observed HFOV was associated with worse outcomes and concluded that this result was similar to recently published studies in adults (OSCAR [OlmeSartan and Calcium Antagonists Randomized] and OSCILLATE [Oscillation for Acute Respiratory Distress Syndrome Treated Early]), comparing these 2 ventilator modalities but with one of the inclusion criteria based on the severity of the hypoxemia.2,3 We were intrigued by the authors’ conclusions. Using a retrospective method makes it difficult for their analysis to support their conclusion, as children with the most severe respiratory disease could have had HFOV. In children, hypoxemia is also a major marker of mortality in acute hypoxemic respiratory failure.4,5 The method used by Gupta et al1 to match the children with HFOV and conventional mechanical ventilation was based on a propensity score that did not evaluate acute hypoxemia. Among those included in the logistic regression model to set the propensity score, the only factors that partially reflected hypoxemia were the Pediatric Index of Mortality 2 score, the Pediatric Risk of Mortality 3 score, and the need for extracorporeal membrane oxygenation. The Pediatric Index of Mortality 2 score included the fraction of inspired oxygen alveolar-arterial difference in partial pressure of oxygen ratio and the Pediatric Risk of Mortality 3 score included alveolar-arterial difference in partial pressure of oxygen and partial pressure of carbon dioxide values; these scores reflected other organ dysfunctions and all missing values were considered to be normal. In addition to the matching process, the baseline severity of hypoxemia in the 2 groups could not be compared. As acknowledged in the Discussion, this key adjustment of hypoxemia was missing; therefore, it is possible that the difference in outcome was simply explained by a difference in the severity of the hypoxemia at baseline. The Conclusions should reflect that uncertainty.
Essouri S, Emeriaud G, Jouvet P. It Is Too Early to Declare Early or Late Rescue High-Frequency Oscillatory Ventilation Dead. JAMA Pediatr. 2014;168(9):861-862. doi:10.1001/jamapediatrics.2014.937