Are oxygen indices effective for predicting pathological intrapulmonary shunt in mechanically ventilated children?

Background. This study was undertaken to evaluate whether oxygen indices ac curately predict pathological intrapulmonary shunt (Qsp/Qt), and to evaluat e the sensitivity and specificity of the FiO(2)-required formula to obtain a desired arterial oxygen tension (PaO2) in mechanically ventilated childre n. Methods. A prospective, hospital-based, comparative study was conducted on 50 mechanically ventilated children at the Intensive Care Units of the Nati onal Institute of Pediatrics (INP) in Mexico City. Blood gas data were pros pectively collected from 50 critically ill, mechanically ventilated childre n, 50 taken before and 40 after FiO(2) change. Assessment of Qsp/Qt, P(A-a) O-2, PaO2/FiO(2), PaO2/PAO(2), and P(A-a)O-2/PaO2 was carried out before an d after FiO(2) change. Results. In first blood gas data, 31 patients were hypoxemic (PaO2 <90 Torr ), 10 were normal, and 9 were hyperoxemic (PaO2 > 100 Torr). Forty patients required FiO(2) modifications that were carried out according to Maxwell's formula. Five children showed persistent oxygen disturbance after FiO(2) c hanges. P(A-a)O-2, PaO2/FiO(2), PaO2/PAO(2), and P(A-a)O-2/PaO2 had sensiti vities of 0.66, 0.71, 0.98, and 0.93, respectively, and specificities of 0. 79, 0.91, 0.29, and 0.64, respectively, to detect pathological Qsp/Qt. All oxygen indices changed significantly after FiO(2) modifications compared fr om initials; Qsp/Qt also showed significant change after FiO(2) change. Pea l-son product-moment showed lineal correlation between each index, and Qsp/ Qt demonstrated their significant correlation (p <0.01). Correlation of Qsp /Qt and PaO2/FiO(2) and PaO2/PAO(2) was significantly higher in younger chi ldren (<13 years) p <0.05. The FiO(2)-required formula to obtain a desired PaO2 had a sensitivity of 0.93 and a specificity of 0.75. Conclusions. The oxygen indices showed sufficient efficacy to detect pathol ogical intrapulmonary shunt, and to have a statistically significant lineal correlation that permits its use during the clinical evaluation of oxygen transport data in most mechanically ventilated children, which is consisten t with other reports on adult populations. However, one limitation for its use in clinical assessment, as reported in previous studies, would be that all indices in the present study are FiO(2)-dependent; therefore, when the FiO(2) varies, the use is misleading. The FiO(2)-required formula is effici ent for defining the appropriated FiO(2) for the obtaining of the desired P aO2, but will always be merely a guide that should be confirmed through blo od gas analysis. (C) 1999 IMSS. Published by Elsevier Science Inc.