[Validation of a clinical prediction tool to evaluate severity in children with wheezing]

Acute lower respiratory infection in children usually causes Bronchial Obstructive Syndrome, which could include hipoxemia. Although pulse oximetry (SaO2) is the gold standard to evaluate hipoxemia, it is usually estimated from a clinical score not yet validated. We aimed to validate the respiratory distress score used in Argentina and to compare its performance with the one used in Chile.

We included 200 children aged under 2 years, with Bronchial Obstructive Syndrome. On admission SaO2 and, Argentinean and Chilean scores components (respiratory rate, heart rate, wheezing, chest indrawing, cyanosis) were assessed. We evaluated the score components ability to predict hipoxemia (SaO2 < or = 95 and SaO2 < or = 91) by logistic regression. Correlation between Argentinean score and SaO2 was estimated. The best threshold of both scores to predict hipoxemia was calculated by ROC curve. Sensitivity, specificity, predictive values and likelihood ratios of both scores to predict hipoxemia were calculated.

Chest indrawing was an independent predictor of hipoxemia (SaO2 < or = 95 and SaO2 < or = 91) (OR: 3.1 IC95%:1.6-5.9 and OR: 13.8 IC95%:1.8-105.4, respectively). The Argentinean score showed acceptable correlation with SaO2 (Spearman: -0,492; p< 0,0001). On SaO2 < or = 91 the Argentinean score showed the best diagnostic performance (auc= 0.904). An Argentinean score > or = 5 was the best threshold to predict hipoxemia (Sensitivity= 100%, specificity= 54.3%). The Chilean score was also evaluated, showing a performance slightly worst than the Argentinean.

An Argentinean score > or = 5 points was sensitive enough to predict hipoxemia (SaO2 < or = 91). This score only allowed identifying children who does not benefit from supplementary oxygen.