Abstract

Fast breathing has been recommended as a predictor of childhood pneumonia. Children living at high altitude, however, may breathe faster in response to the lower oxygen partial pressure, which may change the accuracy of prediction of a high respiratory rate. To assess the usefulness of clinical manifestations in the diagnosis of radiological pneumonia or hypoxaemia, or both, at high altitude (2640 m above sea level), 200 children aged 7 days to 36 months presenting to an urban emergency room with cough lasting less than seven days were studied. Parents were interviewed and the children evaluated using standard forms. The results of chest radiographs and pulse oximetry obtained after clinical examination were interpreted blind. Radiological pneumonia and haemoglobin oxygen saturation < 88% were used as 'gold standards'. One hundred and thirty (65%) and 125 (63%) children had radiological pneumonia and hypoxaemia respectively. Crepitations and decreased breath sounds were statistically associated with pneumonia, and rapid breathing as perceived by the child's mother, chest retractions, nasal flaring, and crepitations with hypoxaemia. The best single predictor of the presence of pneumonia is a high respiratory rate, although the results are not as good as those reported by other studies. A respiratory rate > or = 50/minute had good sensitivity (76%) and specificity (71%) for hypoxaemia in infants. Hypoxaemia had a good sensitivity and specificity for pneumonia mainly in infants (83% and 73%, respectively). Logistic regression analysis showed that decreased or increased respiratory sounds and crepitations were associated with pneumonia, and that hypoxaemia is the best predictor when auscultatory findings are excluded. These results suggest that some clinical predictors appear to be less accurate in Bogota than in places at lower altitude, and that pulse oximetry can be used for predicting pneumonia.