The importance of clinical symptoms and signs in the diagnosis of community-acquired pneumonia.

A sample of 153 children was drawn from a teaching hospital in Säo Paulo, Brazil. It comprised 51 pneumonia cases and equal number of non-respiratory and healthy controls matched by age and sex. Age ranged from 1 month to 7 years. They were all submitted to a standard protocol to investigate clinical symptoms and signs, and diagnosis of pneumonia was supported by X-ray images. Univariate data analysis contrasting pneumonia and non-pneumonia subjects suggested that the best pneumonia indicators would be chest auscultation, history of breathlessness, history of cough, chest in-drawing and fast respiratory rate, in descending order. A multivariate approach including also data from X-ray investigation was then tried with the application of multiple discriminant analysis to study the separation of pneumonia cases, non-respiratory patients and healthy children. It revealed that when many items of information are considered the performance of individual symptoms and signs change. The best predictors of pneumonia were then identified as chest in-drawing, chest auscultation, X-ray, history of breathlessness and toxaemia. Clinical symptoms taken all together contribute more than signs and equal X-ray in importance. Accordingly, it is concluded that any attention to X-ray should be secondary to clinical investigation, and it is suggested that the WHO's guidelines could profit with the inclusion of at least one clinical symptom, namely history of breathlessness which was found more useful than the WHO recommended breath count.