Validación de la definición de síndrome de dificultad respiratoria aguda en pacientes pediátricos

Validación de la definición de síndrome de dificultad respiratoria aguda en pacientes pediátricos

Introduction. It is important to account with a definition of ARDS validated in pediatric patients to allow to assess its true incidence, factors associated with its development, and to indentify patients at early stages of their clinical course to begin supportive treatment.
Objective. The objective of the present study was to validate the ARDS definition of American Thoracic Society and the European Society of Intensive Care Medicine (ATS-ESICM) in pediatric patients.
Design. A study of a diagnostic test.
Material and methods. All the patients that die out in the Pediatric Intensive Care Unit in Santa Clara Hospital among January of 1996 and December of 2002, and was carried out them autopsy, were included in the study. It was registered for each patient the minor Pa0/FiO, value, presence of bilateral pulmonary infiltrates, evidence of left heart failure, and autopsy reports of the patients as a “gold standard” for the detection of the presence ARDS. We compared the clinical diagnosis or ARDS with the autopsy diagnosis or ARDS by three criteria: sensitivity; specificity, with their 95% confidence intervals; and likelihood ratio (LR). The best cutoff threshold value of PaO/FiO, to discriminate patients with and without diagnosis or ARDS, was indentified by the value giving the best combination of sensitivity and specificity and the best area under the ROC curve.
Results. We included 34 patients in the study. The definition of SDRA of the ATS-ESCIM had a sensibility of 80.7%, a specificity of 71.4%, a positive predictive value of 91.3%, a negative predictive value of 50% and a likelihood ratio of 2.82 comparing it with the pathological diagnosis of SDRA. The best cutoff point of PaO2/FiO2 value to discriminate patients with and without ARDS was 150.
Conclusions. The definition of ARDS of the ATS-ESCIM seems to discriminate in an acceptable way pediatric patients with and without SDRA although, the direction bias of our study suggests that the definition has an acceptable capacity to detect patients with SDRA. Subsequent studies are required in pediatric patients to confirm our finds.