© 2003 American Thoracic Society
A Rose by Any Other Name Is Yet a RoseAcute Respiratory Failure in ChildrenDivision of Pediatric Pulmonary and Critical Care Medicine University of Minnesota Medical School Minneapolis, Minnesota The etiology, natural history, and optimal treatment of respiratory failure have been the subject of active investigation for over 100 years. In the 1925 edition of Osler's text, respiratory failure was described as "frothy pulmonary edema that resembles serum, not the sanguinous transudative edema fluid seen in dropsy or congestive heart failure" (1). Despite that trenchant description of respiratory failure, a formal description of acute respiratory distress syndrome was not forthcoming until the publication of Ashbaugh and coworkers' work (2) in 1967, in which they described the key components of respiratory failure. Despite a remarkably insightful and almost presciently complete description of the pathophysiology of respiratory failure, consensus on rigorous diagnostic criteria was not achieved until 1994 (3). Wide acceptance of a definition of acute respiratory distress syndrome has enabled the completion of clinical trials that have resulted in commonly shared treatment principles that promise to enhance outcomes. Concomitant with establishing diagnostic criteria, the consensus conference report endorsed changing the name from adult to acute respiratory distress syndrome (ARDS) (3). The name change explicitly acknowledged that whereas the clinical manifestations of ARDS may differ between age groups, the fundamental alteration in alveolar–capillary permeability is present in both children and adults in ARDS, thereby legitimizing the application of lessons learned in trials performed in adults to the management of children. However, significant developmental differences between children and adults, and even between infants and children, transcend a simple change in nomenclature and mitigate against the wholesale acceptance of data derived in studies of adults. Thus, ongoing and effective clinical trials in children with ARDS remain essential. The information provided by Randolph and coworkers in the May 15 issue of AJRCCM (4) should prove tremendously beneficial in the design and performance of clinical trials in children with ARDS. Among the important observations is that, in the absence of preexisting illness, acute respiratory failure is relatively rare in children. The manuscript provides previously unavailable and essential data regarding the incidence and prevalence of mechanically ventilated children in pediatric intensive care units in North America. These normative data represent a boon to investigators conducting research in a patient population that is notoriously difficult to study. While carefully outlining the challenges entailed by performing a trial in children, the investigators argue compellingly for the need to perform such studies. Although they carefully outline the limitations of the data, there are aspects that merit close scrutiny. The authors use the term "acute respiratory failure" throughout their article—a term that introduces the very vagueness that they aim to eliminate by performing the present trial. In some ways, a trial that provides data, but lacks a rigorous definition for respiratory failure, harkens to a time when no such definition was available. A diagnosis of ARDS was provided for only 23 of the 303 patients, despite the report of a PaO2:FIO2 ratio of less than 200 in 52% of the patients (4). The lack of specificity regarding the use of diagnostic criteria weakens the message. Implicit in the objective of the trial is a goal of determining the normative incidence and natural history of respiratory failure in children. The very design of trial, however, may prevent determination of the real incidence and epidemiology of respiratory failure. In excluding all children with preexisting illnesses from the trial, the authors may have introduced the exact sample bias the exclusion criteria sought to circumvent. Use of exclusion criteria is legitimate only if the phenomenon being studied—in the present case, respiratory failure—is independent of the exclusion criteria. Intuitively, the incidence of respiratory failure is higher in children with illnesses such as pulmonary hypoplasia, bone marrow transplantation, and abnormal pulmonary vascular tone or in children in whom further care is futile and a "do not resuscitate" order is in place. It is likely that respiratory failure may have been the primary reason for deeming further care futile for some of the patients. Thus, the authors may have excluded the very patients most likely to experience the condition under study—acute respiratory failure. The highly circumscribed nature of the patient population may have excluded the patients the authors hoped to study. The study design thereby severely compromises the central objective of the investigation. The trial includes empiric evidence that further buttresses the notion that the trial may not be representative. Respiratory failure frequently complicates the condition of children who have sustained traumatic injury. No trauma patients were included in the trial. The mortality rate for patients with ARDS was 4.6% (4), a rate consistent with that observed in a trial of 21 patients cited by the authors (5). In the three largest trials of ARDS in children, however, the mortality rate was approximately 45% (6–8). At best, the study provides data regarding a significantly more circumscribed patient population than intended by the authors. The implications must be considered with great caution. Consequently, the significance of the study is diminished. At worst, the study design might lead to a conclusion with markedly less justification than warranted by the data. Thus, the data may lead to investigative nihilism relative to either the significance or feasibility of studying respiratory failure in children—a result that, in all likelihood, none of the authors would wish to endorse. Acknowledgments D.N.C. has no declared conflict of interest. I.Y.H. has no declared conflict of interest. REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
